PakIEM - Model Design Report

Pakistan Integrated Energy Model
(Pak-IEM)
ADB TA-4982 PAK
Final Report Volume I

Model Design Report
Prepared by
International Resources Group
for
Asian Development Bank
and
Ministry of Planning and Development

Government of Pakistan
August 2011
Pakistan Integrated Energy Model ADB TA No. 4982-PAK
Knowledge Summary
The Asian Development Bank (ADB) has supported the Planning Commission of the
Government of Pakistan to develop an integrated, full-sector, energy system planning model
using the TIMES least-cost optimization framework1. The model, called the Pakistan
Integrated Energy Model (Pak-IEM), encompasses the entire energy system including
resource supplies, refineries and power plants, transmission and distributions systems for
fuels and electricity and the end-use devices that deliver the energy services the economy
demands. Pak-IEM can be used to assess the impacts of various options and strategies for
meeting the country's future energy needs in an optimal manner. The model integrates
planning factors pertaining to investments in technologies – both supply-side and demand-
side, plant operating life and O&M costs, fuel costs, including system infrastructure costs,
limitations in indigenous and imported resource supplies, environmental emissions,
technological improvements, energy efficiency and conservation to assess the costs and
benefits of policies that will shape the country for the coming decades. The model depicts
development of a Reference scenario out to 2030, and allows policy, resource supply,
technology and other options to be modified to create alternative scenarios. ADB also
supported training and capacity building of a multi-institutional planning team that includes
sector-based energy system experts and universities to ensure complete energy system
coverage and facilitate communication and buy-in from the various energy system
stakeholders in the country.
The initial analyses utilizing Pak-IEM examined options for achieving the evolution of the
Pakistan energy system over the next 20 years that is needed to achieve the government’s
economic growth projection, which corresponds to 5.6% average GDP between now and
2030. The Pak-IEM Reference scenario calls for:
Four-fold increase in electricity generation – 94,000 GWh to 410,000 GWh
82,000 MW of new power generation capacity additions
Three-fold increase in consumption of high value petroleum products – 6.2 Mtoe to
18 Mtoe
It also points out a looming Energy Security crisis where, by 2030:
Under current practices and policies proven conventional natural gas reserves will be
depleted.
Energy imports jump from 27% to over 45% of total supply.
Delays in finding the investment for critical energy projects will extend the current
load-shedding situation.
Alternative scenarios, analyzed using Pak-IEM, show that significant annual savings can be
achieved from Smart Policies (best practices):
Eliminating load shedding avoids Rs. 524 billion in economic losses.
Reducing electricity transmission and distribution losses by 7% saves Rs. 7.3 billion
(gross).
Improving end-use energy efficiency saves Rs. 41 billion (net).
Successful exploration to deliver 20% more gas saves an additional Rs. 37 billion
(gross).
Collectively, these Smart Policies delay and dampen the increased dependency on foreign
energy sources by over 20 Mtoe a year beginning 2030. Exploitation of non-hydro
1
The Integrated MARKAL/EFOM System (TIMES) model generator, the successor to the MARKAL
modeling framework, was been conceived, developed, and is continually supported by International
Energy Agency – Energy Technology Systems Analysis Programme (IEA-ETSAP),
http://www.etsap.org.
Pak-IEM Final Report Volume I – Model Design Page ii

renewables can further enhance energy security by reducing total imports to 38% of total
energy in 2030.
This integrated energy modeling capability now provides the GoP with a framework for
examining priority energy policy issues facing Pakistan, ranging from closing the current
supply-demand gap to improving energy security by fully promoting energy efficiency and
exploiting indigenous resources.
Pak-IEM Final Report Volume I – Model Design Page iii

TABLE OF CONTENTS
I. INTRODUCTION ......................................................................................................... 1
A. Background.................................................................................................................. 1
B. MARKAL/TIMES Modeling Framework ....................................................................... 2
C. A Framework to Institutionalize Energy Modeling Capability....................................... 4
1. Lessons Learned from Previous Pakistan Modeling Efforts ..................................... 4
2. Experiences Using MARKAL/TIMES in Other Countries.......................................... 4
3. Challenges to Sustaining Integrated Energy Analysis .............................................. 6
4. Conceptual Institutional Structure............................................................................. 7
D. Model Design Overview............................................................................................... 9
E. Scope and Structure of the Report ............................................................................ 10
II. OVERVIEW OF Pak-IEM DATA REQUIREMENTS .................................................. 12
A. Data Collection Process ............................................................................................ 12
B. Model Structure ......................................................................................................... 13
C. Cost Accounting......................................................................................................... 19
D. Energy Balance ......................................................................................................... 19
E. Sector Specific Data Needs and Design Issues ........................................................ 20
III. RESOURCE SUPPLY ............................................................................................... 22
A. Sector Overview ........................................................................................................ 22
B. Current Fossil Fuel Resources .................................................................................. 22
C. Future Supply Options ............................................................................................... 25
D. Renewable resources ................................................................................................ 28
E. Data Sources ............................................................................................................. 30
IV. REFINING AND OIL & GAS DISTRIBUTION............................................................ 32
A. Oil Refining Sector..................................................................................................... 32
B. Distribution of oil and gas .......................................................................................... 35
C. Data Sources ............................................................................................................. 35
V. POWER SECTOR ..................................................................................................... 36
A. Sector overview ......................................................................................................... 36
B. Approach to modeling the power sector generation .................................................. 37
C. Planned new build ..................................................................................................... 40
D. New technology options............................................................................................. 41
Pak-IEM Final Report Volume I – Model Design Page iv

E. Transmission and Distribution System ...................................................................... 45

F. Load shedding ........................................................................................................... 48
G. Summary of Main Power Sector Data Sources ......................................................... 49
VI. DEMAND SECTORS................................................................................................. 51
A. Modeling approaches ................................................................................................ 52
B. Agricultural Sector ..................................................................................................... 55
C. Industry Sector........................................................................................................... 58
D. Residential Sector...................................................................................................... 73
E. Commercial Sector .................................................................................................... 80
F. Transportation Sector ................................................................................................ 81
VII. DEMAND PROJECTIONS......................................................................................... 93
A. Overview.................................................................................................................... 93
B. Primary Demand Drivers ........................................................................................... 94
C. Demand Projections by Sector .................................................................................. 97
5. Agriculture .............................................................................................................. 97
6. Commercial ............................................................................................................ 98
7. Industry................................................................................................................. 100
8. Residential............................................................................................................ 103
9. Transport .............................................................................................................. 107
VIII. Base Year Calibration.............................................................................................. 111
IX. AREAS FOR FURTHER MODEL DEVELOPMENT................................................ 113
X. REFERENCES ........................................................................................................ 116
XI. APPENDICES.......................................................................................................... 119
Appendix 1. Model Naming Conventions ........................................................................ 119
1. Demand sectors ................................................................................................... 119
2. Commodity Names and Descriptions ................................................................... 119
3. Technology Names and Descriptions................................................................... 121
Appendix 2. Model Templates......................................................................................... 122
Appendix 3. List of Unpublished Data Sources ............................................................... 124
Appendix 4. Examples of Energy Planning Applications Using MARKAL/TIMES........... 129
Appendix 5: Justification for Selection of the MARKAL/TIMES Modeling Platform......... 132
Appendix 6. Job description for role of Lead Energy Modeler and Analyst..................... 136
LIST OF FIGURES
Pak-IEM Final Report Volume I – Model Design Page v

Figure 1: TA Organizational Structure .................................................................................... 2

Figure 2: Institutional Composition of the Planning Team ...................................................... 2
Figure 3: MARKAL/TIMES Model Overview ........................................................................... 3
Figure 4. Conceptual Institutional Structure for the Pak-IEM Modeling Capability ................. 9
Figure 5: Structure of the Pak-IEM Data Inputs .................................................................... 10
Figure 6: Data Development Process ................................................................................... 12
Figure 7: Primary Energy Production by Fuel type and Final Energy Consumption by Sector
for 2006-07 (Source: HDIP 2008 (Pakistan Energy Year Book 2007)) ................................. 19
Figure 8: Energy Resource Supplies in 2006-07, excluding biomass (Source: HDIP 2008) 22
Figure 9: Current Structure of Oil and Gas Resource Supply and Upstream Processing
Sectors.................................................................................................................................. 32
Figure 10: Product Shares for Refineries in 2006/07............................................................ 33
Figure 11: 2006/07 (a) Power Sector Fossil Energy Use (PJ) .............................................. 37
Figure 12: Planned New Generation Capacity, MW ............................................................. 41
Figure 13: Build rates (GW) per year (2015/2030) by Generation Type ............................... 45
Figure 14: Distribution of Electricity from Centralized and Decentralized Power Generation
and Imports ........................................................................................................................... 47
Figure 15: Load shedding on the PEPCO system, July 06-June 09 ..................................... 48
Figure 16: Pak-IEM approach to modeling load shedding .................................................... 49
Figure 17: Final Energy Consumption by Sector, 2006-07 ................................................... 51
Figure 18: Comparison of Final Energy Consumption estimates between Energy Year Book
2007 / Pak-IEM ..................................................................................................................... 52
Figure 19: Bottom-up and Top-down Approaches for Demand Sector Modeling ................. 53
Figure 20: Approach to Bottom-up Estimation of Agriculture Sector Energy Use................. 55
Figure 21: Agriculture Sector RES Diagram ......................................................................... 56
Figure 22: Industrial Sector Energy Consumption (PJ), 2006/07 (HDIP 2008)..................... 59
Figure 23: Industrial Sector Natural Gas Consumption (% PJ), 2006/07 (Source: SNGPL /
SSGC)................................................................................................................................... 60
Figure 24: General Approach to Industry Sector Modeling ................................................... 61
Figure 25: Reference Energy System – Iron and Steel (Independent Operators) ................ 63
Figure 26: Reference Energy System – Iron and Steel (Pakistan Steel) .............................. 64
Figure 27: Reference Energy System – Textiles .................................................................. 65
Figure 28: Reference Energy System – Cement .................................................................. 66
Figure 29: Reference Energy System – Sugar ..................................................................... 68
Figure 30: Information Needed to Characterize Residential Sector Energy Use .................. 74
Figure 31: Approach to modeling the residential sector........................................................ 75
Pak-IEM Final Report Volume I – Model Design Page vi

Figure 32: Shares of final energy consumption by type in the urban residential sector........ 76
Figure 33: Shares of final energy consumption by type in the rural residential sector.......... 76
Figure 34: Residential sector RES Diagram ......................................................................... 78
Figure 35: Season fractions for electricity demand in residential sector across different end-
use applications .................................................................................................................... 79
Figure 36: Timing of Demand Profiles in Residential Sector by space heating and cooling . 79
Figure 37: Shares of final energy consumption by type in the commercial sector ................ 80
Figure 38: Fuel Consumption (PJ) in the Transport Sector, 2006/07 (HDIP 2008)............... 82
Figure 39: Approach to Bottom-up Estimation of Energy Use in Road Transport Sector ..... 83
Figure 40: Pak-IEM Transport Sector RES Diagram – Passenger Vehicles......................... 84
Figure 41: Pak-IEM Transport Sector RES Diagram – Freight Vehicles............................... 84
Figure 42: Official GoP GDP / Sectoral Growth Targets ....................................................... 94
Figure 43: Reference, Medium and Low GDP Growth ......................................................... 95
Figure 44: Overall GDP Growth Rate Projections................................................................. 96
Figure 45: GDP Sectoral Contributions for Reference, Medium and Low Projections.......... 96
Figure 46: Projections of Popuation and Household Size..................................................... 97
Figure 47: Agricultural Sector Useful Energy Demands ....................................................... 98
Figure 48: Commercial Sector Useful Energy Demands .................................................... 100
Figure 49: Industry Sector Relative Demand Growth ......................................................... 101
Figure 50: Number of Urban and Rural Electricity Consumers ........................................... 104
Figure 51: Comparison of Energy Use per GDP for Pakistan and Other Countries ........... 105
Figure 52: Diagram of Useful Energy Demand Development ............................................. 106
Figure 53: Urban Residential Sector Useful Energy Demands ........................................... 107
Figure 54: Rural Residential Sector Useful Energy Demands ............................................ 107
Figure 55: Passenger Transport Demand Projections........................................................ 109
Figure 56: Freight Transport Demand Projections.............................................................. 110
Figure 57: Reported load curve values versus model outputs ............................................ 112
LIST OF TABLES
Table 1: Summary Sector Definitions ................................................................................... 13
Table 2: Reference Case Period Lengths............................................................................. 15
Table 3: Season / Day Time Slices....................................................................................... 15
Table 4: Typical Technology Characterization Parameters .................................................. 20
Table 5: Crude Oil and Natural Gas Reserves, and Years Remaining Resource ................ 23
Pak-IEM Final Report Volume I – Model Design Page vii

Table 6: Prices of Energy Commodities in 2006/07 .............................................................. 25

Table 7: Biomass energy use in Pakistan Rural / Urban Households (PJ), 2006/07 ............ 28
Table 8: Renewable potentials for Pakistan (Source: AEDB) ............................................... 30
Table 9: Summary of Data Sources for the Resource Supply Sector ................................... 30
Table 10: Key Operating Characteristics of Pakistan Refineries .......................................... 32
Table 11: Crude Input Shares for Pakistan Refineries.......................................................... 33
Table 12: Planned New Refineries ....................................................................................... 34
Table 13: Summary of Data Sources for Refining and Gas / Oil Distribution Sector ............ 35
Table 14: Data Requirements for Power Plant Characterization .......................................... 38
Table 15: Model Assumptions for PEPCO Thermal Power Plants ....................................... 38
Table 16: Model Assumptions for IPP Thermal Power Plants .............................................. 39
Table 17: Model Assumptions for KESC Thermal Plants ..................................................... 39
Table 18: Model Assumptions for WAPDA Hydropower Plants ............................................ 39
Table 19: New technology options in the Power Sector ....................................................... 42
Table 20: Potential hydro sites in Pakistan ........................................................................... 44
Table 21: Summary of Data Sources for the Power Sector .................................................. 49
Table 22: Technical Parameters for Characterization of End-Use Technologies ................. 54
Table 23: Stock and Operating Assumptions for Agricultural Machinery .............................. 57
Table 24: Season / Day Demand Shares for Agriculture Sector Electricity .......................... 57
Table 25: Summary of Data Sources for the Agriculture Sector ........................................... 58
Table 26: Sources of Data for Estimating Sub-sector Energy Consumption ........................ 62
Table 27: Key Data Assumptions for Steel Melting and Re-rolling Sectors .......................... 63
Table 28: Cement sector production and coal consumption, Million tonnes ......................... 66
Table 29: Key assumptions used for sugar sector in Pak-IEM model .................................. 68
Table 30: Brick Kiln Efficiencies (Source: ADB 2009b)......................................................... 69
Table 31: Industrial Captive Generation Plants (Source: UDI WEPP, Platts) ....................... 70
Table 32: Captive generation capacity by Industry Sub-sector............................................. 71
Table 33: Summary of Data Sources for the Industry Sector ............................................... 72
Table 34: List of Sector Associations.................................................................................... 73
Table 35: Data sources to characterize residential sector technologies ............................... 77
Table 36: Summary of Data Sources for the Residential Sector .......................................... 80
Table 37: Transportation Sub-sectors in Pak-IEM ................................................................ 81
Table 38: HSD Energy Balance (PJ) for 2006/07, and Allocation in Pak-IEM ...................... 82
Table 39: Parameters Required for Characterization of Transport Sector Technologies ..... 85
Pak-IEM Final Report Volume I – Model Design Page viii

Table 40: Vehicle Stock Statistics, 2006/07 (NTRC 2008) and Fuel Split Assumptions ....... 85
Table 41: Efficiency and Activity Assumptions for Different Vehicle Types, ENERCON 2003;
Lifetime and Occupancy Values Sourced from NTRC (JICA 2006b).................................... 87
Table 42: Summary of Data Sources for the Transport Sector ............................................. 92
Table 43: Energy Service Demand Drivers by Sector .......................................................... 93
Table 44: GDP Growth Rates Components for Medium and Low Projections ..................... 95
Table 45: Agricultural Sector Energy/GDP Elasticity Factors ............................................... 98
Table 46: Commercial Sector Energy /GDP Elasticity Factors ............................................. 99
Table 47: Industry Sub-sector Elasticity Factors................................................................. 101
Table 48: Textile Industry Component Growth Rates ......................................................... 102
Table 49: Cement Industry Energy / GDP Elasticity ........................................................... 103
Table 50: Average Household Energy Use in 2007............................................................ 105
Table 51: Passenger Transport Energy /GDP Elasticity Factors ........................................ 108
Table 52: Freight transport Energy /GDP Elasticity Factors ............................................... 109
Table 53: Core Names of Energy Commodities ................................................................. 119
Table 54: Names of DEM Commodities.............................................................................. 120
Table 55: Naming Conventions for Technologies ............................................................... 121
Table 56: Pak-IEM Reference Scenario Workbooks .......................................................... 122
Pak-IEM Final Report Volume I – Model Design Page ix

LIST OF ACRONYMS
AC Air conditioning
ADB Asian Development Bank
AEDB Alternative Energy Development Board
BAU Business As Usual (scenario)
CASAREM Central Asia – South Asia Regional Electricity Market
CCS Carbon Capture and Storage
CNG Compressed Natural Gas
DGPC Directorate General Petroleum Concessions
DISCO Distribution (Electricity) Company
EIA Energy Information Administration
ESMAP Energy Sector Management Assistance Program (World Bank)
ETSAP Energy Technology Systems Analysis Programme
EYB Energy Year Book
FBS Federal Bureau of Statistics
GCISC Global Change Impact Studies Centre
GDP Gross Domestic Product
GHG Greenhouse gas
GoP Government of Pakistan
GW Gigawatts
GWh Gigawatt hours
HDIP Hydrocarbon Development Institute of Pakistan
HESS Household Energy Strategy Study
HFO Heavy Fuel Oil (known as Furnace Oil in the EYB)
HIES Household Income Expenditure Survey
HP Horsepower
HSD High Speed Diesel
ICE Internal Combustion Engine
IEA International Energy Agency
IPP independent power producers
IPCC Intergovernmental Panel on Climate Change
IRG International Resources Group
ISGS Inter State Gas Systems
KESC Karachi Electric Supply Corporation
KWh Kilowatt hours
LDO Light Diesel Oil
LNG Liquefied Natural Gas
Pak-IEM Final Report Volume I – Model Design Page i

LPG Liquefied Petroleum Gas

LUCF Land Use Change and Forestry
MEC MEConsult
mpkm Million passenger-kilometers
mtkm Million ton-kilometers
MW Megawatts
NARC National Agricultural Research Centre
NTDC National Transmission & Despatch Company
NTRC National Transport Research Center
OCAC Oil Companies Advisory Committee
OGRA Oil and Gas Regulatory Authority
Pak-IEM Pakistan Integrated Energy Model
PEPCO Pakistan Electric Power Company
PIHS Pakistan Integrated Household Survey
PJ Petajoules
PM Particulate matter
PPEPCA Pakistan Petroleum Exploration and Production Companies Association
PSRMA Pakistan Steel Rolling Mills Association
RES Reference Energy System
SNGPL Sui Northern Gas Pipelines Limited
SSGC Sui Southern Gas Company
TA Technical Assistance
TERI The Energy and Resources Institute (India)
TOR Terms of Reference
UETT University of Engineering and Technology, Taxila
USAID United States Agency for International Development
VEDA VErsatile Data Analyst (FE-Front End / BE-Back End)
WAPDA Water and Power Development Authority
WEO World Energy Outlook (published by the IEA)
WEPP World Electric Power Plants Database (published by Platts)
Pak-IEM Final Report Volume I – Model Design Page ii

I. INTRODUCTION
A. Background
The Asian Development Bank (ADB) conducted a Technical Assistance (TA) with Pakistan’s
Planning Commission to assist the Government of Pakistan (GoP) in developing an
Integrated Energy Model (Pak-IEM) that enables a national energy planning group to assess
the impacts of various strategies for meeting future energy needs in an optimal manner. The
model integrates planning factors pertaining to financial investments, economic costs,
energy supply, national resources, energy use, environmental impacts, technology
improvement, energy efficiency, and conservation to assess the costs and benefits of
policies that will shape the country for the coming decades.
This allows national policymakers and stakeholders within the various ministries and
regulatory bodies, as well as the private sector, to collaborate and work through an open and
transparent process with a dedicated energy planning team capable of producing high
quality integrated analysis of energy planning strategies and national policy options. The TA
also established a multi-institutional capacity of local experts to steward Pak-IEM on an
ongoing basis.
This optimal integrated energy planning capability now provides the GoP with a framework
for examining priority energy policy issues facing Pakistan, ranging from closing the current
supply-demand gap to improving energy security by fully exploiting indigenous resources.
Building this capacity over an 18-month period, including conducting such assessments
during the mentoring phase, creates within the GoP a solid analytic foundation for designing
appropriate policies and planning the evolution of the energy system, including taking into
consideration strategies to account for future climate change pressures. The open nature of
the methodology and process employed allows these policies and options to be examined in
a clear and comparable manner to facilitate better communication and foster better
cooperation among the various stakeholders.
The organization of the TA is shown in Figure 1, with modeling capability centered at the
Energy Wing of the Planning Commission and supported by a network of key institutions in
Pakistan to comprise the Pak-IEM Planning Team as shown in Figure 2. The Consultant
team, which consisted of International Resources Group (IRG) with support from MEConsult
(MEC), designed the model, developed the model input data in cooperation with the
Planning Commission, other ministries, agencies, and organizations (with the assistance of
an Advisory Committee Task Force), and trained the Planning Team and Support Institutions
in the use and ongoing stewardship of the model. The model employs the MARKAL/TIMES
modeling framework.
Pak-IEM Final Report Volume I – Model Design Page 1

Pakistan Planning Advisory Committee

Ministries of
Asian Commission • Petroleum and Natural Resources
Development Define policy studies, integrate
• Water and Power
• Planning and Development
Bank into the planning process, and • Environment, Transport and
ensure ongoing f unding Communications
Pakistan Atomic Energy Commission
Regulatory Agencies, Development
Institutes, and Others
Planning Team Advisory Committee

IRG Consultant Team Task Force
Energy Wing staf f who understand
International Experts (IRG) Experts f rom the Ministries and
the model, and coordinate use with
National Experts (ME Consult) Industries to provide technical
the Host Institutions
support
Host Institutions University Research

Modeling experts who maintain Centers
the PIEM database and perf orm Prof essors and graduate students
policy analysis who use and improve the model
Figure 1: TA Organizational Structure
Planning Team Composition

Energy Wing ,Planning & Commission, Government of Pakistan
Global Change Impact Studies Center (GCISC)
Hydrocarbon Development Institute of Pakistan (HDIP)
National Transport Research Center (NTRC)
Pakistan Atomic Energy Commission (PAEC)
Pakistan Electric &Power Company (PEPCO)
Pakistan Institute of Engineering and Applied Sciences (PIEAS)
Pakistan Institute of Development Economics (PIDE)
University of Engineering & Technology, Lahore (UETL)
University of Engineering & Technology, Taxila (UETT)
Figure 2: Institutional Composition of the Planning Team
B. MARKAL/TIMES Modeling Framework

MARKAL/TIMES is the most widely used energy systems optimization model in the world
today, deployed in some 70 countries in over 200 institutions. It has a 30-year track record of
evolution and success. It was developed and is widely used and disseminated through the
International Energy Agency - Energy Technology Analysis Programme (IEA-ETSAP,
www.etsap.org). IRG team members have been involved in model development, capacity
building, and policy analysis involving MARKAL/TIMES during much of that time.
A MARKAL/TIMES model requires a core set of data to characterize the energy system that
the model represents. Once constructed, the model solves for the least-cost solution that
utilizes the available energy carriers (fuels, electricity, heat) and technologies (power plants,
refineries, end-use devices) to meet the energy service demands (e.g., space conditioning,

industrial process heat, lighting, passenger kilometers traveled), subject to physical

limitations (e.g., resource availability, hydro/solar/wind potential), targets (e.g., energy
independence and diversity in the terms of imports, renewables share of electricity
generation) and policies (e.g., greenhouse gas (GHG) limits, emission caps, nuclear), and
other constraints (e.g., rate of technology adoption) imposed on the system. A stylized
representation of the underlying energy system depicted in the model is illustrated in Figure
3.
PRIMARY ENERGY CONVERSION END-USE DEMAND FOR

SUPPLY TECHNOLOGIES TECHNOLOGIES ENERGY SERVICE
(PRIMARY ENERGY) (FINAL ENERGY) (USEFUL ENERGY)
Renewables, e.g. Fuel Processing Industry, e.g. Industry, e.g.

Biomass Plants, e.g. Steam boilers Process steam
Hydro Oil refineries Machinery Motive power
Hydrogen prod.
Mining, e.g. Ethanol prod. Services, e.g. Services, e.g.
Crude oil Air conditioners Cooling
Natural gas Power Plants, e.g. Light bulbs Lighting
Coal Conventional fossil
fueled Households, e.g. Households, e.g.
Imports, e.g. Solar Space heaters Space heat
Crude oil Wind Refrigerators Refrigeration
Oil products Nuclear
CCGT Agriculture, e.g. Agriculture, e.g.
Exports, e.g. Fuel cells Irrigation pumps Water supply
Oil products Combined heat and
Coal power Transport, e.g. Transport, e.g.
Gasoline car Person-km
Stock changes Fuel cell bus
Figure 3: MARKAL/TIMES Model Overview
The underlying components of a MARKAL/TIMES model include:

♦ A Reference Energy System (RES), which is a network describing the structure of an
energy system as interconnected flows of energy carriers (e.g., fuels, electricity,
heat) through energy-conversion (e.g., refineries, power plants) and energy-
consuming technologies (e.g., air conditioners, cars, furnaces)
♦ Resources and imports, both current and projected, which are generally described
through supply curves characterizing the amount of resource available at a particular
cost
♦ Energy service demands, specified at a level of sub-sector disaggregation that meets
the analytical requirements of the model user, and is supported by available or
estimated data
♦ Sets of technologies (organized by type and sector) with their characteristics, such as
investment cost, operating cost, efficiency, availability, emission rates, etc.
The term “technology or process” in MARKAL/TIMES is used as a general term for all
physical facilities and energy transformation devices, such as refineries, power plants, home
heating/cooling systems, light bulbs, and automobiles. The model “builds” physical devices
by increasing the technologies’ installed capacities and “operates” them as measured by
productive output. It does so by monitoring capital stock turnover to identify investment
opportunities and timing. Similarly, “commodity” is a general term that refers to the various
entities moving between processes, corresponding to energy, material, emissions, and
energy services. Energy carriers include primary resources, processed fuels, electricity,
heat, and renewable energy inputs – such as hydro, wind, solar, and geothermal energy.
Materials and emissions, similar to energy, are flows into and/or from the processes. The

demand for energy services (e.g., passenger transportation, lighting/heating/air conditioning

of rooms, process heat for iron and steel) are satisfied by end-use technologies and must be
met over the planning horizon, thereby serving to drive the Reference (Business-as-Usual or
BAU) scenario evolution of the energy system.
Energy service demands represent consumption by individuals and activities associated with
driving the economy. As such an initial projection is established according to relationships
between drivers (e.g. trends of Gross Domestic Product (GDP), population, urbanization,
commercial floor space, etc.) and individual sub-sectors of the energy economy; perhaps
adjusted internally according to own-price elasticities (not currently employed for Pak-IEM).
Then, using sensitivity analysis, the necessary configuration of the energy system to meet
future economic growth under varying assumptions and policies can be evaluated.
C. A Framework to Institutionalize Energy Modeling Capability

1. Lessons Learned from Previous Pakistan Modeling Efforts
Early in the TA, the project team gathered information from the national experts, the
Planning Commission and other stakeholders to assess lessons learned from previous
attempts to establish energy planning capacity for Pakistan. This was an important exercise
to understand what energy modeling capacity had previously existed and why it had not
been sustained. Such experiences have been used to develop a proposed framework to
institutionalize MARKAL/TIMES capacity in Pakistan, and ensure that it is sustainable in the
long term.
In mid-1980s USAID supported the Planning Commission with development of the RESPAK
model, which was a optimization planning framework model. The model was used for a few
years, but fell into disuse as the expertise established in the Commission was lost to other
assignments. In addition, as power sector privatization started in the early 1990s, least-cost
economic modeling was considered no longer necessary.
The RESPAK model was also criticized because of credibility issues arising from the large
number of assumptions and analyst judgments needed for the model, and its inability to
reflect real-world situations, especially in the near-term, such as the inability of developers to
implement near-term optimal power plants.
A key lesson learned was that the institutional set-up in the Planning Commission was not
sustainable, as the necessary high-caliber people could not be retained. They either got
promoted to other posts within the government administration or left for higher paid private
sector jobs. Maintenance of a reliable repository of data and the need for regular model
updating were other challenges which could not be sustained. Extensive data collection, its
diversity of sources, and respective analysis on an ongoing basis overwhelmed the limited
resources of the Energy Wing resulting in a gradual abandonment of the model.
2. Experiences Using MARKAL/TIMES in Other Countries
The IRG team has been involved in many energy system modeling projects in different
regions of the world, many of which are listed in Appendix 4 of this report. Many of these
countries, e.g., Malaysia, Indonesia, China and India are similarly situated to Pakistan in
terms of level of development, strong agricultural base to the economy, sharp divides
between urban and rural household energy consumption, etc. The IRG Team also has
experience with more developed countries regarding the establishment of MARKAL/TIMES
models and institutional capacity building. Of particular relevance to the situation in Pakistan,
see the Energy Policy and Systems Analysis Project in Appendix 4.
Finally, the IRG Team has provided energy system modeling support to their national
planning agencies (e.g., US Department of Energy and Environmental Protection Agency,
UK Department of Energy and Climate Change and UK Committee on Climate Change).
These experiences using energy system planning and optimization models in both

industrialized and developing countries provides important insights and lessons for the
Planning Commission, as it seeks to maintain, develop and use Pak-IEM. The insights and
lessons from other countries are synthesized below, and include the benefits of using such
models, but also the associated challenges faced by various national planning teams.
Benefits of Integrated Energy Analysis
The benefits of utilizing a modeling capability such as MARKAL/TIMES are described below.
By their nature, integrated energy models are large, complex and data intensive
mathematical tools that require highly trained specialists if they are to be used properly.
Expertise from all sectors on the energy system must be represented and a dedicated staff
is required to maintain and use these models if meaningful policy analysis is the desired
product. Without a dedicated core staff that is supported with sufficient resources and
regular requests to perform relevant policy analyses, trained staff will disperse and learned
modeling skills will be forgotten.
• Provision of a framework for integrated analysis. Energy / climate policy can be
assessed in an integrated way using MARKAL/TIMES. In many countries, policies
are developed for single sectors or specific policy areas, with limited understanding
of the impacts on other sectors or policies, and without insights concerning how
multiple policy objectives could be achieved more cost-effectively. Recent analysis
using MARKAL/TIMES models under the RESMD2 project for USAID has supported
the joint assessment of renewable energy, energy efficiency and low carbon
objectives, highlighting the strong synergies between different policy areas. In
Pakistan, this type of framework has been absent, with individual energy-related
policy assessed by sector, and independently of other policy objectives.
• Economics as an important policy driver. MARKAL/TIMES has provided insights
into the economic tradeoffs of various policies and goals. In the UK, the
understanding that system costs in the longer term are not prohibitive (based on
MARKAL analysis) has been critical to the acceptance amongst policy makers of a
low carbon, sustainable energy system. Given the challenge associated with raising
the necessary medium to longer investment, such insights are invaluable for
proactive planning in Pakistan.
• A long term planning perspective. As most major investments in the energy sector
(e.g., power plants and refineries) will be around for decades it is critical that near-
term decision making and policy formulation take into consideration the longer term
implications, which is a strength of MARKAL/TIMES. This has been particularly
critical for those countries or agencies who have been undertaking low carbon
analyses using MARKAL/TIMES models, which illustrate the significant shift in types
of technologies that need to be employed (e.g., IEA Energy Technology Perspective
(ETP), US-DOE Integrated Benefits Analysis (IBA), South Africa Long Term
Mitigation Strategies (LTMS)). With rapid economic growth forecast over the next 20
years in Pakistan, significant investment will be made in the energy system, and it is
critical that the medium to long term implications of such investments are understood.
• Usability of modeling tools. The diverse and extensive global community using
MARKAL/TIMES indicates that whilst complex, the tools for managing the model and
analysis process are able to be well understood and used effectively for a range of
policy and planning issues by diverse and dedicated experts around the world under
widely varying environments. The Pakistan Planning Team will become part of this
2
Regional Energy Security and Market Development project has built national integrated energy
planning models in 11 countries in Southeast Europe and Eurasia..

international MARKAL/TIMES community that can be tapped to provide modeling

support.
• Engagement with policy community. The least cost paradigm that
MARKAL/TIMES employs provides a critical perspective that is well understood and
accepted by policy makers, which is why it is so widely used across many different
countries. This fact should continue to help engage policy makers across
Government in Pakistan.
• Transparency. The modeling frameworks allows for transparency, both in terms of
assumptions and the model's mathematical foundation. This transparency addresses
an important prerequisite for stakeholder engagement and acceptance of results.
Across many countries, for example the UK, the ability for stakeholder’s to know
exactly what assumptions have been made for different sectors has ensured greater
engagement, and avoided criticism associated with other so called ‘black box’
approaches. It also means that stakeholders know what the current assumptions are,
and can be challenged to provide improved data where available. This was the
approach adopted during the development of Pak-IEM, where assumptions were
sent to different experts for scrutiny. However, it is also important to note that
ensuring transparency can also be challenging, with data assumptions needed to be
made publically available, for example, through extensive documentation.
3. Challenges to Sustaining Integrated Energy Analysis

Although MARKAL/TIMES is the most common integrated energy system optimization
model, other frameworks exist, such as LEAP, ENPEP and MESSAGE. None of these
modeling frameworks, with the exception of LEAP (which is essentially an accounting
framework), has the longevity or breadth of usage on an ongoing basis of MARKAL/TIMES.
Appendix 5 provides a summary of the characteristics and capabilities of each of these
modeling frameworks.
These challenges to building and sustaining local capacity for integrated energy system
modeling and analysis have been gathered from the experiences of all types of energy
system models over a period of more than 30 years. Much of the impetus for integrated
energy, environment and economic modeling grew out of the oil shocks of the 1970s. These
lessons learned come from industrialized and developing country experiences. They have
been gained thru successes and failures and although some model may be more user
friendly than others, these lessons fundamentally reflect the complex nature of the integrated
energy system that is being modeled.
• Maintaining team capacity. To become a competent modeler takes several weeks
of training followed by several months of continued work with the model under the
guidance of a skilled and experienced manager. Therefore, losing key personnel
from a team can have significant impact on the capacity to maintain its model and
undertake studies. For many of the successful modeling teams across the world a
critical mass of 3-5 people (as opposed to 1-2) is imperative to sustaining capacity.
These core energy modeling experts need to be committed full-time to continually
providing support and technical advice to policy makers. The institutional structure
described below safeguards the current expertise, by proposing a core team to
maintain the model, and ensuring a range of other institutions are involved with
similar expertise.
• Keeping models relevant. Models need to be maintained, improved and updated
regularly (based on the data available). It has been seen in many countries that
models that are not kept current leading to the loss of confidence in them as credible
tools for policy analyses. Country teams that are successful in maintaining models

either have strong research funding or are used frequently by Government for
different studies. For example, the UK MARKAL model has been systematically used
for all of the major energy and climate strategy reviews over the last eight years,
ensuring that it has the necessary funding and is kept current and thereby relevant.
In addition, for the past 5 years, it has had significant additional academic research
funding under the UK Energy Research Centre.3
• Developing data systems and a network of providers. Due to the data
requirements of MARKAL/TIMES models, robust data systems and procedures are
needed to manage data, while a network of providers is needed to ensure that the
best available data flows to the modeling team. Strong links to statistical agencies,
and to different sector experts, is important. This issue has been recognized in recent
capacity building work undertaken by IRG and colleagues in the USAID-funded
RESMD project. In a region where statistical information is not strong across specific
sectors, additional links with experts and data providers has been critical. Developing
a network of data providers has been an important part of the Pak-IEM TA project,
whereby relevant data providers have been identified and asked to provide data in
specific formats.
• Engagement with and understanding of modeling tools by different
stakeholders. Key stakeholders need to be engaged and able to understand the
relevance of MARKAL/TIMES models, and how they can provide important insights
on energy policy issues. Therefore the model should be made available to a wider
suite of qualified institutions, as well as encouraging requests to the Planning Team
to conduct analyses on their behalf of issues of particular interest to them. In the
USA, the release of a national MARKAL model by the US EPA has been extremely
successful, resulting in the engagement of numerous research and academic
institutions in further developing and improving the model, and running a range of
different analyses.
As reflected in the section below, an important part of the TA was developing an institutional
structure that ensures a wide range of model users as well as strong engagement from
interested stakeholders.
4. Conceptual Institutional Structure

The following recommendations for creating a sustainable institutional structure for the Pak-
IEM MARKAL/TIMES modeling capacity are based on consideration of the above history,
the observations in different countries and the experiences of the IRG team and the Planning
Commission. These recommendations resulted in the conceptual institutional structure for
the Planning Team, shown in Figure 4, which closely resembles how the Planning Team was
organized during the TA.
• Manage and coordinate modeling activities in Pakistan through a dedicated
Planning Unit, overseen by the Planning Commission. A central group needs to
have lead responsibility, retaining the master version of the model, and overseeing
any future development. A key responsibility of this leadership role is to ensure that
the model remains fully documented (to maintain transparency) and is quality
controlled (to maintain credibility). A dedicated unit can also be available to
undertake assignments at short notice for the Planning Commission, other
Government departments, and the private sector.
3
Further information on UKERC can be found at the website http://www.ukerc.ac.uk/

• Recruit high caliber individuals to the Planning Unit who have an engineering
or economics background. During the TA, the project team provided advice on the
type of qualifications and skills that would be needed by members of the modeling
team. One of the specifications written, for the role of Lead Energy Modeler and
Analyst, is provided in Appendix 6. While there is a need for several individuals with
diverse background and skills, one (or two) individuals need to be devoted full-time to
overseeing the modeling system and team.
• Host the model at several different institutions where it is likely to be used. The
Planning Team composition (see Figure 2) reflects this important view that having
a number of host organizations fosters wider use of the model, establishes a national
modeling community to promote joint working and learning, and expands the range of
sectoral and technical expertise of the team.
The types of institution are also important. While government planning needs to
be the first priority, private sector involvement should be encouraged with a focus on
enabling them to use the model to examine issues affecting their sectors. Universities
can deliver training courses on and research using MARKAL/TIMES, and seed the
next generation of energy planners.
• Produce annual or biannual energy outlook report. The regular publication of a
policy and perspective report based on Pak-IEM would ensure that the model is
consistently developed, used and improved.
• Establish at network of data providers from different sectors that can provide
information for model updates. The TA has identified and established the relevant
sources of data. Since data is needed on an ongoing basis, there is a need to
develop the process, procedures and tracking system that identifies data
requirements and providers, monitors and follows up on data requests, and quality
controls said data, which in turn should lead to database for storing, organizing and
disseminating the wealth of data provided.
• Maintain the Advisory Committee Task Force, which provided essential guidance
during the TA. This would ensure that continued stakeholder involvement can be
maintained, model development scrutinized, data requests satisfied and priority
analysis needs identified.

Pakistan Planning Pak-IEM Advisory

Commission Committee
Donor Agencies
Def ine policy studies, integrate
results into the planning process, and Ministries, Agencies, Regulators
ensure ongoing f unding & Other Stakeholders
Planning Unit Advisory Task Force

Technical Support for
Model Improvements Energy Wing staf f who maintain the
Experts f rom the Ministries and
Pak-IEM database, perf orm policy
and Studies Industries to provide technical
analysis, and coordinate studies and
support and guidance
model improvements
Research Centers &

Universities Study Teams
Analysts, prof essors and graduate Assembled by Planning Unit f or
students who use and improve the specif ic studies or tasks
model
Figure 4. Conceptual Institutional Structure for the Pak-IEM Modeling Capability
D. Model Design Overview

The Pak-IEM is utilizing the TIMES model generator, which is the successor to the MARKAL
modeling framework which has been conceived, developed and continually supported by
International Energy Agency – Energy Technology Systems Analysis Programme (IEA-
ETSAP). It employs the VErsatile Data Analyst (VEDA) data and model management
system to organize and handle the input data and process the results thereby overseeing all
aspects of working with the model. The development of the model input data is organized
into a set of Excel workbooks (templates) that are managed by VEDA-FE (Front End) to form
the complete model for calibration (for the year 2006/2007) and development of the
Reference scenario out to 2040, and beyond (see Figure 5). Any number of alternate
scenarios can then be introduced to the mix as future options and policies for analysis.

Energy Model Horizon & Demand

Balance System Settings Projections
Resource
Supply
Future
Pak‐IEM Technologies
Calibration &
Fuel Shares
Demand Sector Base‐Year
Supply Sector Templates
Base‐Year
Templates Agriculture
Upstream
Commercial Residential
Power Industry Transport
Figure 5: Structure of the Pak-IEM Data Inputs
E. Scope and Structure of the Report

The initial design of the Pak-IEM Reference Energy System was determined through
discussions with MEConsult (MEC), the Planning Team, and Advisory Committee Task
Force experts to determine the appropriate level of detail for the model based upon the
available data collected to date and the policy issues for which the model will be used.
The initial Model Design Report in July 2009 presented a near finalized overview of the base
year model templates and data assumptions across different sectors to the Planning Team.
A presentation to the Advisory Committee Task Force Group provided a briefer discussion of
the model design, and a more in-depth discussion on the ongoing critical data issues.
Feedback from these groups was critical in informing the model structure and design, and
assisting with critical data issues, most notably leading to a Task Force Focus Group
Mission in early August to again attempt to complete the known data collection process.
The Model Review Report, submitted in January 2010, was the second interim report
providing additional detail on assumptions across new technologies, an overview of the key
demand drivers and a preliminary version of the Reference scenario.
This report presents the final version of the model structure, and comprehensively describes
the assumption and data sources. It is meant to be the primary documentation resource. The
report is organized according to the major sectors of the energy system that are represented
in the model. Sections II and III present an overview of data collection activities, and key
model design and data issues. Sections IV, V, and VI describe model structure and data
assumptions across resource supply, upstream, and power generation sectors. Section VII
covers end-use demand sectors. The latest model version for which this report relates is
organized in a single input data folder named contained in PIEM_TMPL_101024(Final
Report).rar.

This report does not provide information on how to use the model. Pak-IEM Final Report –
Volume II – Policy Analysis describes the initial use of the model to assess general policy
and technical issues that serve to demonstrate how the model can be used and the nature of
the results it produces. Comprehensive guidance in the form of detailed instructions on Pak-
IEM under VEDA is provided in an associated document titled Pak-IEM Final Report:
Volume III – User’s Guide.

II. OVERVIEW OF Pak-IEM DATA REQUIREMENTS

A. Data Collection Process
The crucial first step in developing a credible integrated energy planning capability for
Pakistan is acquiring and assembling the detailed data needed for a comprehensive national
model. This begins with a thorough review of the nature of the existing economic, financial,
and technical data related to the Pakistan energy system from resources (domestic and
imports) through end-use demands (by sector) that is necessary for the model. The data
development components are illustrated in Figure 6.
Data collection activities to characterize the base year energy system (in 2006-07) were
largely completed at the end of the first year. Then the Planning Team was organized into
Sector Groups and charged with addressing outstanding data needs, validating the data
depicting the current energy system and proposing options for consideration for the future.
The data needs and sources used for the different sectors are listed in the sector-relevant
sections of this report which follow.
An index of primary sources for the information included in Pak-IEM is provided in the
associated sector sections of this report. Whenever possible, the source data are imbedded
directly into the Pak-IEM model templates. The actual reports and other data sources are all
located on the project specific FTP site, hosted and managed by IRG. It is intended that the
Energy Wing will hold all of these background data files. Unpublished sources are listed in
Appendix 3, while published sources can be found in the Reference section.
Input Data
Energy service demands

Resource cost & supply
Technology characterization
Infrastructure costs & options
Policy & physical constraints
Emission factors
Etc.
Energy
Service
Demands
End Use
Technologies
Industrial
Commercial Sectoral
Conversion Lighting Residential Activity
& Process Air Conditioning Agricultural
Buildings Transportation General
Technologies Factories Activity levels
Resources Vehicles Growth rates Parameters
Power plants Energy Intensity
CHP plants
Biomass GDP Growth
District heat plants
Coal Population
Refineries
Oil Urbanization
Fuel processors
Gas
Geothermal
Hydro
Solar
Wind
MARKAL/TIMES Model
Figure 6: Data Development Process
Once the initial year data was gathered and prepared, the data collection process shifted its
focus to the requirements for development of the Reference scenario which depicts the
anticipated evolution of the Pakistan energy system under a Business-As-Usual (BAU)
situation. This involves sourcing data on:

• Demand drivers and their relationship for different sectors

• Information on known new power plant, refinery and major energy infrastructure
builds and refurbishments
• Future resource potential and energy prices
• Technology options for future production / consumption of energy.
These data collection activities were initiated by means of a number of Advisory Committee
Task Force Focus Group meetings starting in August 2009 with repeated focused
discussions as needed, when the Consultant team met key stakeholders in a series of
sector-specific meetings. These meetings also addressed critical data needs for the base
year energy system characterization.
In addition to the sector-specific information discussed later, there is a range of issues that
are important for characterizing the underlying energy system, such as the number of
regions, sector categorization, and time horizon, along with other assumptions that cut
across all sectors such as discount rate and emission factors which are discussed next.
B. Model Structure
Regional issues. Pak-IEM covers the whole of Pakistan’s energy system, and can therefore
be used for national policy studies. This first version of the model is comprised of a single
region, which is most appropriate at this stage of model development, given the nature of
some of the available data. It is recognized that significant regional differences in energy
supply and demand do exist, and that political boundaries are also an important
consideration, therefore future versions of the Pak-IEM should move towards regionalization;
in view of this, data collection, where possible, captured any regional breakdowns.
Sector definitions. The model has been designed according to the following sectors as
summarized in Table 1.
Table 1: Summary Sector Definitions
Sector Description
Supply-side
Resource supply Domestic resources and imports
Processing and distribution
Processing (refining) and distribution of oil and gas
(Upstream)
Electricity generation Individual power plants and IPPs, along with self-generation in Industry
Demand-side
Tractors (for field and transport), water pumping and miscellaneous
Agriculture
demands
An aggregate (single) commercial sector broken out for the basic
Commercial / Public sector
energy service demands (e.g., cooling, lighting, cooking, etc.)
Six major industrial sectors (iron & steel, textiles, sugar, cement,
fertilizer, bricks) modeled at either the process level, or based upon
Industry generic energy service needs (e.g., process heat, motor drive,
miscellaneous). Other industry modeled using a generic energy service
approach.
Aggregate urban and rural residential sectors broken out for the basic
Residential
energy service demands (e.g., cooling, lighting, cooking, etc.)
Each basic transportation mode (e.g., passenger vehicle, trucks,
Transportation buses, rail, ship, air) demand is modeled along with various devices for
meeting each transport mode
Emission factors. Emissions being tracked in Pak-IEM include CO2, SO2, and NOx. It will
be possible to add additional GHGs and local pollutants in future versions of the model. The

focus in this version is CO2, due to fast developing policy issues on climate change (in
particular the ongoing deliberations for a post-2012 agreement on greenhouse gas
reductions). SO2 and NOx are being tracked as important air quality pollutants, and provide
the potential for undertaking assessment of policy co-benefits.
For CO2, and to some degree SO2, the actual emission levels are directly proportional to the
amount of fuel used and the carbon or sulfur content of the fuel (e.g. coal, crude oil,
petroleum products, natural gas, etc.). Some technologies have controls to remove a
proportion of the fuel-based emissions, and these will be specified according to their
abatement percentage. NOx and particulate matter (PM) emissions are technology-specific,
so generic device-specific data (e.g., power plants, industrial processes, commercial and
residential boilers and furnaces, and vehicles) is needed to fully account for these emissions.
It is understood that for CO2 the national inventory uses Intergovernmental Panel on Climate
Change (IPCC) default emission factors, based on 2006 guidance documents.4 This is
consistent with what is currently being used in Pak-IEM, using data from the GAINS-ASIA
model.5 For air quality pollutants, emission factor information from the GAINS-ASIA model is
being used. For the base year, weighted (as opposed to technology-specific) emission
factors have been used (by sector-fuel), as they include a range of different technologies
which have different emission factors.
It is important to note that biomass use in the model is considered to be carbon-neutral, as
per inventory guidelines.6 The question is whether the biomass used is sustainable; if not,
the emissions associated with unsustainable extraction are captured as removals under the
Land Use Change and Forestry (LUCF) sector. In an energy-only model, biomass could
therefore have carbon emissions associated; however, it would be important to know
whether and to what extent the biomass resource was sustainable (e.g. extraction is less
than (or equal to) rate of re-growth).
Model Horizon. As part of establishing the scope of the model a start period (base year)
and series of model milestone periods (of varying duration) must be established (thereby
determining the model horizon7). The start period is particularly critical as all the base year
and calibration data will be developed for this year. The length of the first period is a single
year, in order to calibrate the model to the energy balance for that year. The start year
chosen for the Pak-IEM model is 2006/07, running from beginning of July 2006 to the end of
June 2007. (Note that this is reported as 2007). This was chosen as the most recent year for
which data were generally available for all parts of the energy system (at the time the project
started). 2007/08 was subsequently considered as an alternative but it was decided that
some of the required data for certain sectors were not available.
The number of milestone periods and their respective duration for the planning horizon is
shown in Table 2. Earlier time periods tend to be shorter because more is known about the
4
Based on personal communication with HDIP
5
Data provided by GCISC from GAINS-ASIA model, developed by IIASA (see following website for
further information – http://www.iiasa.ac.at/rains/gains_asia/main/index.html?sb=1. The name of the
data files can be found in Appendix 3.
6
The Revised 1996 Inventory guidelines from the IPCC state that biomass fuels are included in the
national energy and emissions accounts for completeness. These emissions should not be included in
national CO2 emissions from fuel combustion. If energy use, or any other factor, is causing a long
term decline in the total carbon embodied in standing biomass (e.g. forests), this net release of
carbon should be evident in the calculation of CO2 emissions in the LUCF sector.
7
Note that TIMES supports extrapolation of all parameters permitting the modeling horizon to be
extended with minimal additional data required.

evolution of the system, in respect of energy policy, resource forecasts, and planned new
build; for later periods, less is known and therefore longer periods are more suitable. Right
now the primary focus is out to the 2040 planning horizon. However, since TIMES allows for
a separation of data years and model years, Pak-IEM can be run for any desired set of years
(periods).
Table 2: Reference Case Period Lengths8
Period Start Year Period Reporting Year Period End Year Period Length
2007 2007 2007 1
2008 2008 2009 2
2010 2010 2010 1
2011 2011 2011 1
2012 2012 2012 1
2013 2013 2013 1
2014 2014 2014 1
2015 2015 2015 1
2016 2016 2017 2
2018 2020 2022 5
2023 2025 2027 5
2028 2030 2032 5
2033 2035 2037 5
2038 2040 2042 5
Discount rate. A global discount rate is used in the model to discount all costs back to the
base year and calculate the total discounted system cost (used in the model’s objective
function). The model is currently using a default 7% rate. Note that sector or even
technology-specific discount rates may also be introduced if appropriate (e.g. to examine
impediments to the uptake of efficient (more expensive) appliances). Such discount rates
have been used in the electricity generation sector to account for the cost of capital.
Season and Day Time-Slices. Some demands for energy are typically non-uniform over
the year. In particular, electricity and natural gas have seasonal and time-of-day
consumption patterns that need to be properly modeled and tracked. Therefore, the setting
up of the daily and seasonal temporal resolution of the model is important. The timeslices in
the model representing this resolution are shown in Table 3.
Table 3: Season / Day Time Slices
Months per Hours
Season Months Day Slices Hours
Season per Day
April,
Intermediate (I) 3 October- Day (D) 11 0700-1800
November
May- 1800-1900,
Summer (S) 5 Shoulder (S) 4
September 2100-2400
November-
Winter (W) 4 Peak (P) 2 1900-2100
March
Night (N) 7 2400-0700
An explanation of how these timeslices were determined is provided in the Box below. Key
inputs in determining the timeslice structure are the typical 24-hour load dispatch curve from
8
This Periods Definition set in the model is labelled PIEMMSY.

Pakistan Electric Power Company (PEPCO) and the Karachi Electric Supply Corporation
(KESC), and monthly sectoral consumption patterns.9
Based on the timeslice structure, timing of demands have then to be constructed. These
have been developed on the basis of expert judgment, and are further detailed in each of the
sections describing demand sectors. The timing of demands, for electricity (hourly basis) and
gas (seasonal basis) need to be calibrated to match the supply profile in the base year. They
are also used to determine the timing of demand in future years.
Determining Timeslices for Pak-IEM
There are a number of decisions to make relating to the seasonal and diurnal splits that are needed
to capture differences in generation during different times of the day in response to demand,
particularly the peak demand. Time-slices need to be structured in a way that will enable the model
to provide a ‘reasonable’ representation of the 24 load curve across different seasons.
The situation in respect of Pakistan is of course complicated by the serious load shedding issues
that have been occurring over the last few years.
PEPCO and KESC data have been used to determine the time-slice structure, representing the
combined system across all of Pakistan. This includes seasonal electricity consumption by broad
sector (residential, commercial, agriculture, industry) and daily load profile for typical weekdays /
weekend day (Sunday) in each month.
Seasonal electricity consumption for specific sectors in PEPCO / KESC systems is shown below.
7000
6000
5000
4000 AGRICULTURE
GWh
COMMERCIAL
3000 RESIDENTIAL
INDUSTRY
2000
1000
0
Jul Aug Sept Oct Nov Dec Jan Feb Mar Apr May June
The first issue was how to split the seasons up, based on the level of demand in the different
months. The proposal was to have 3 seasons:
• Summer: May-Sep (5), accounting for higher demand driven by higher cooling demand.
• Winter: Dec-Mar (4), in which occurs the lowest seasonal demand, and
9
Data were provided directly by KESC / PEPCO to the project team for use in Pak-IEM. A full listing
of these files, and the data providers, is provided in Appendix 3. In summary, KESC monthly
consumption and load curve information were provided in files KESC data sheet.doc and
Hourly_Demand_KESC.xls. PEPCO monthly consumption and load curve information can be found in
PEPCO_Consumption_Patterns.xls and PEPCO 24 hour and seasonal load profile data_v03.xls

• Intermediate (INT): Apr, Oct-Nov (3), which are the months between the high-low demand
periods.10
Daily load data, provided by PEPCO and KESC, is shown for the three seasons proposed above. A
daily time-slice structure was proposed that fit all three seasons, and was relevant for both weekday
and weekend days. This was as follows:
• Night (0.00-07.00);
• Day (07.00–18.00);
• Peak (19.00-21.00), and
• Shoulder (18.00-19.00, 21.00-24.00).

In all three seasons, a 2-hour peak can be observed although occurs slightly later in summer and is
less pronounced and earlier in winter. To ensure that the model does not have an overly complex
representation, each season has the same daily timeslices. The graphs below represent the load
level over time for typical weekdays; no distinction has been made for weekends as the shape of the
curve is very similar to that observed below.
Summer
18000
16000
14000
12000 2006/07
2006/08
10000
2006/09
8000
2007/05
6000 2007/06
4000
2000
0
1 3 5 7 9 11 13 15 17 19 21 23
10
If seasonal storage was to be used in the model (e.g. pumped hydro storage) the intermediate
season would need to be changed to a 2 season depiction – Spring / Fall – to ensure storage and
dispatch can occur from one season to the next. However, this was not considered likely in the
Pakistan context so the single intermediate season is employed.

Winter
18000
16000
14000
12000
10000 2006/12
2007/01
8000
2007/02
6000
2007/03
4000
2000
0
1 3 5 7 9 11 13 15 17 19 21 23
Intermediate
For the intermediate season, it could be argued that November belongs to the Winter season due to
the much lower load over the 24 hour period. This pattern emerged recently (following the use of
year-specific data) from the combined load data from KESC / PEPCO. However, moving this month
to the winter group is unlikely to impact significantly on the model solution, and would require
significant base year recalibration.
18000
16000
14000
12000
10000 2007/04
8000 2006/10
6000 2006/11
4000
2000
0
1 3 5 7 9 11 13 15 17 19 21 23
An ongoing model development priority is to further assess whether the data used (2006/07)
capture the full peaking requirement for the system. There has been some discussion that

the peak should be more pronounced, and that load profile data from earlier years may be
more suitable, when load shedding was not a significant issue.11
Supply-demand equilibrium. It is important to recognize that MARKAL/TIMES models
always calculate a supply-demand equilibrium, whereby energy service demands are
satisfied in every period. This makes assessment of future supply-demand gaps difficult.
However, modeling techniques can be used to explore specific problems associated with
insufficient energy supply, as described in Section V.F in the case of power sector load
shedding.
C. Cost Accounting
All costs in the model (investment, operating and fuel) are expressed in 2007 US Dollars. All
cost data is first converted to US Dollars, with Pakistan conversion factors sourced from the
latest Economic Survey of Pakistan (GoP 2009). Deflators to adjust source data for the
model to get all costs to US Dollars on a 2007 basis are sourced from US Bureau of
Economic Analysis.12
D. Energy Balance
The national energy balance provides a comprehensive picture of the supply and
consumption of energy in the form of each fuel type used and the sectors (demand, power,
and refineries) where these fuels are consumed. The energy balance for 2006/07 for
Pakistan is sourced from the Energy Year Book (EYB) 2007 (HDIP 2008), and provides the
basis for characterizing the start or base year of the model. Figure 7 summarizes this data
showing total primary energy supply and final energy consumption by sector.
Nuclear Other Govt.

0.9% 2%
Hydro
12.6%
Oil Domestic
Coal 30.0% 21%
LPG Transport
7.3%
0.5% 27% Commercial
4%
Agriculture
Natural Gas 2%
48.6% Industrial
44%
Total Primary Energy Total Final Energy

2667 PJ / 60 Mtoe 1591 PJ / 36 Mtoe
Figure 7: Primary Energy Production by Fuel type and Final Energy Consumption by
Sector for 2006-07 (Source: HDIP 2008 (Pakistan Energy Year Book 2007))13
11
Alternative load profile data (for PEPCO system only) listed in Appendix 3, and labeled Hourly Load
Data [DATE].xls
12
Bureau of Economic Analysis, National Income and Product Accounts Table, Table 1.1.9 Implicit
Price Deflators for Gross Domestic Product, http://www.bea.gov
13
The 2008 Pakistan Year Book shows some small changes in estimates for 2006-07 on primary
energy production (as identified based on a review by PAEC); however, these do not fundamentally
change the base year picture so have not been incorporated to date.

A number of issues were identified with the Energy Balance, which resulted in amendments
to the overall Yearbook balance. These included the handling of high-speed diesel and non-
commercial biomass.
• Concerning high speed diesel (HSD), based on bottom-up estimates of consumption
across all end-use demand sectors, a shortfall was identified. This was attributed to a
significant amount of black market diesel not accounted for in the official energy
balance. Based on expert judgment, a 10% increase in HSD was assumed for
2006/07. This issue is discussed in detail in the transport section of this report.
• Non-commercial fuels, in particular various forms of biomass, are not present in the
official Energy Balance due to lack of statistics. Examples include wood use in the
residential sector and bagasse use in the sugar industry. Estimates of biomass use in
the sugar sector were made, as have non-industrial sector estimates. No commercial
biomass use is published in the Energy Balance either.
E. Sector Specific Data Needs and Design Issues

The following sections of this report focus on describing the structure of the different sectors
depicted in the model, and key data sources and assumptions used for both the base and
future years.
All sectors are linked to the centralized energy balance, ensuring that base year
consumption of energy matches what is presented in the Energy Yearbook (augmented as
discussed above). On a sector-by-sector basis, the number and types of sub-sectors are
then considered. The level of detail is dependent on the energy uses in a given sub-sector,
the availability of suitable data, the sub-sector energy demands to be modeled, and the role
of future demand drivers (e.g. GDP, population) in projecting the future demand for energy
services.
Each sub-sector was then characterized by the technologies that it uses in the base year
and the alternatives available in future years. The data required for technology
characterization is extensive, and Table 4 provides a summary of the typical parameters
used in MARKAL/TIMES models.
Table 4: Typical Technology Characterization Parameters
Technology
Description
Characteristics
Year in which the technology is initially available. For all Base Year technologies the
Start Year
default start year is the base year.
Existing capacity of each power plant, refinery, or technology type, which should be
Residual stock determined from statistical data where possible. Otherwise, for the demand devices
this data can be derived from the decomposition of sectoral and end-use shares.
Number of years the technology is typically expected to be in service; this value
determines how the existing capacity is retired, where it is assumed that for demand
Lifetime devices the available technology linearly decreases over the lifetime if no retirement
profile is provided (e.g., anticipated power plant and refinery retirement should be
explicitly provided).
Corresponding to the amount of output (energy or demand serviced) per unit of fuel
Efficiency
consumed.
Total capital cost for each new technology, including interest during construction. This
Investment is needed for all new technologies and any existing technologies for which future
investments (e.g., retrofits) are possible.
Fixed Operating and
Fixed cost associated with the capacity of the facility or device.
Maintenance
Variable Operating Variable operating cost associated with running the facility or device, other than fuel
and Maintenance costs.
Utilization Factor Annual utilization factor of device calculated from the amount of time that the fully

Technology
Description
Characteristics
rated capacity of a device is actually used during a year (for example, heating and
cooling devices have a relatively low annual utilization factor as they are only used part
of the year and even then only at full rating some of the time, for power plants there is
usually utilization data available (seasonal for hydro and some other renewables)).
Provided when there is more than one existing device using the same fuel to meet a
Base Year Share
sub-sector demand.
Share of each energy carrier (material) consumed per unit of output; the value is
Input Energy Carrier normally 1 (default) unless more than one commodity is consumed by the device (e.g.,
dual-fuel power plant, solar hot water with electric back-up).
Share of each product produced (e.g., gasoline/diesel/etc. from refinery, electricity and
heat from couple heat/power plants) or demand serviced by devices, normally 1
Output Energy
(default) unless more than one product is produced or demand is served (e.g., heat
Carrier
pump doing both heating and cooling, car used for city/inter-city travel – note that
these require different efficiencies and utilization factors).
Assumptions concerning future year evolution of the energy system are also critical. This
includes near-term known changes to the energy system, resulting from revised or new
policies, or planned / nearly completed projects. It also relates to the different (economic,
demographic, etc.) drivers across sectors that will result in growth of energy requirements in
future years, and the different technology options that will be needed to meet increased
demands.

III. RESOURCE SUPPLY

A. Sector Overview
The resource supply sector includes primary energy resources, supplied domestically
or via imports, as well as exports of energy commodities. The resource supply picture
is provided in
Electricity
Indigenous Production
Nuclear Imports
Hydro Exports
Coal
LPG
Av. Fuel
Naptha
Heavy Oil
Diesel
Gasoline
Crude
Natural Gas Total 29.3
‐5 0 5 10 15 20
Mtoes
Figure 8 below. Pakistan has significant indigenous resources of natural gas (though
declining) and hydropower, while for crude oil and refined oil products there is a high
reliance on imports. All of the naphtha produced by domestic refining and over 25% of the
aviation fuel is exported. Coal is mainly used in the industry sector; domestic coal tends to
be low quality (sub-bituminous or lignite) and used primarily for brick making; imported coal
tends to be used in industries such as cement and iron & steel.

Electricity
Nuclear Imports
Hydro Exports
Coal
LPG
Av. Fuel
Naptha
Heavy Oil
Diesel
Gasoline
Crude
‐5 0 5 10 15 20
Mtoes
Figure 8: Energy Resource Supplies in 2006-07, excluding biomass (Source: HDIP

2008)
B. Current Fossil Fuel Resources

The primary energy totals shown in
Electricity
Nuclear Imports
Hydro Exports
Coal
LPG
Av. Fuel
Naptha
Heavy Oil
Diesel
Gasoline
Crude
‐5 0 5 10 15 20
Mtoes
Figure 8 meet the energy needs of the economy (ignoring the issues of access to affordable
energy, load shedding, as well as the substantial use of biomass) in 2006/07. In future years,

the energy system will need to determine (within constraints imposed) where to source
energy resources. For domestic resources, this may be dependent on the costs of extraction
in different provinces, the available reserves in different fields, and of course the
requirements of power and demand sectors. The balance of energy needed will then have
come from imports.
Each domestic resource is characterized according to the following parameters:
• Maximum annual extraction rates, constraining the model’s access to resources on
an annual basis
• Proven reserves, constraining the model’s cumulative access to resources over the
model time horizon
• Cost of extraction (and delivery), to influence model choices about which resource is
most economic.
Extraction rates
Extraction rates in future years (out to 2030+) for oil and gas fields are capped at the
maximum extraction values in projections sourced from Directorate General Petroleum
Concessions (DGPC). That is the model may produce up to the level each year as long as
the proven reserves are not exhausted. Extraction rates for coal come from DG Minerals.14
To account for new oil and gas resource finds in future years from continued exploration, an
additional 5% uplift is applied to known reserves. Cumulative resource levels (see Table 5)
are capped using data on reserves provided in the Energy Year Book (HDIP 2008).
Indicative estimates on the number of years that resources will last are also provided, based
on i) the maximum permitted extraction rate in the model and ii) if extraction continued at the
rates observed in the base year.
Table 5: Crude Oil and Natural Gas Reserves, and Years Remaining Resource
Commodity type / Reserves Years remaining (Max Years remaining (Current
Region (Mtoes) Rate) Rate)
Crude oil – BALO 0.2 NA NA
Crude oil – NWFP 12.0 9.7 9.7
Crude oil – PUNJ 15.6 11.9 14.8
Crude oil – SIND 22.0 9.5 10.0
Crude oil – Total 49.8 10.2 11.1
Natural gas - BALO 190.2 23.1 23.1
Natural gas - NWFP 55.7 12.5 12.5
Natural gas - PUNJ 27.7 20.5 25.6
Natural gas - SIND 415.9 13.9 20.8
Natural gas – Total 689.5 15.6 20.4
These data indicate that domestic crude oil supplies will last approximately 10 years, while
natural gas will last between 15-20 years. Associated gas also exists, with production rates a
function of the oil extracted. The reserves of associated gas add a further 11.3 Mtoes of
resource to the system.
14
See following data files received from DG PC and DG Minerals – Long term Gas Projections-
DGPC.xls, Long term Oil Projections-DGPC.xls and Long term Coal Projections-DG Minerals.xls.

Costs
Primary energy resources are usually characterized by cost-supply curves, which
characterize the amount of each resource that can be made available at a given cost in each
period of the model planning horizon. This is the approach used for natural gas supplies,
where fields are grouped or “binned” together based on the price of gas from the field, with
the price variation representing differences in extraction costs. Gas prices are sourced from
Oil and Gas Regulatory Authority (OGRA).15 The rate of future increases in gas prices are
based on information from the IEA’s World Energy Outlook (WEO) 2009 (IEA 2009b) (in the
absence of Pakistan-specific data).
Coal prices are sourced from the Energy Year Book (HDIP 2008). These are province-
specific, and dependent on coal quality and demand in local markets. Price projections are
provided out to 2015 in the Energy Year Book; post-2015, prices are assumed to grow
based on growth rates used in European projections (in the absence of Pakistan specific
data). Imported coal prices are sourced from WEO 2009 (IEA 2009b).
For oil supplies, the situation is different. Prices are primarily determined by the regional or
global market prices for crude, and therefore provide no insights into variations in the costs
of domestic production. The import crude price is based on an average gulf prices (plus
average freight rate assessment), sourced from the EIA.16
The refinery sector sees the same crude price whether it is domestic or Arabian Light crude;
the Government however mandates that all available domestic crude needs to be used by
the refining sector. Therefore, the refining sector does not make decisions on the basis of
price, and is more concerned about quality in terms of the product mix that can be produced
from each crude type.
To ensure any domestic crude that is produced is used in the refining sector the domestic
crude price is set at 70% of the import price. This also ensures that the domestic crude price
does not interfere with the overall marginal price of the commodity, which should be tied to
international market prices. Future prices of imported and domestic crude, as well as
imported oil products, are tied to IEA projections for oil. The price of crude in 2007 was
around $60 per barrel. The latest estimates from IEA can be found in WEO 2009, with oil
rising (in real terms) to $100 per barrel in 2020, and $115 per barrel in 2030 (2008 prices).
Note that this is the long run price of oil, and does not therefore take account of short-term
volatility in the oil markets.
The price of oil products is based on information provided in the Energy Year book (HDIP
2008). The economic price is based on what is known as the ex-refinery price17 plus the
freight and distribution margin. Changes in prices in future years are tied to crude oil
projections. The import price is determined by applying a 20% increase whilst export prices
have a 40% decrease applied. This deflation factor is important to ensure that the model
does not start producing oil products simply for export.18
15
Well head gas prices can be found at http://www.ogra.org.pk/
16
Energy Information Administration (EIA) website,
http://tonto.eia.doe.gov/dnav/pet/pet_pri_wco_k_w.htm
17
The ex-refinery price of a product, which is paid to local refineries, equates to the landed cost of the
product. See OCAC website for further description - http://www.ocac.org.pk/ex_refinery5.asp
18
Allowing exports is particularly important in the model to ensure smooth refinery operation. Due to
the product slate of a refinery, and limited flexibility, a situation could be envisaged whereby operation
ceases and switches to imports because there is no demand for specific products produced.

Base year prices used in the model, which do not include sector taxes / subsidies are
provided in Table 6 below.

Table 6: Prices of Energy Commodities in 2006/07

Commodity Commodity type $2007 / GJ $2007 / toe $2007 / MMBtu
Crude Oil Domestic 7.04 311.1 7.43
Import 10.06 444.5 10.61
Oil Products Motor Spirit 15.01 663.4 15.83
Kerosene 15.50 685.1 16.35
High Speed Diesel 14.23 629.0 15.01
Light Diesel Oil 12.83 567.1 13.54
LPG 12.08 534.1 12.75
Heavy Fuel Oil 8.82 389.8 9.30
Aviation Fuel 14.71 650.3 15.52
Naptha 12.38 547.0 13.06
Natural Gas Domestic 1.58-2.48 69.88-109.68 1.67-2.62
Domestic (Assoc) 0.35 15.6 0.37
Imports 6.60 291.7 6.96
Coal Domestic 1.07-3.27 47.15-144.73 1.125-3.45
Imports 3.93 173.7 4.15
Alternatively, the model can be run with full prices (including taxes / subsidies) that are seen
in reality by different sectors. This information is important to help understand how the
system responds under market distortions, using a “financial” rather than “economic”
approach. These prices are sourced from the Energy Year book (HDIP 2008) for oil products
and Pakistan Economic Survey (GoP 2009) for gas and electricity.19
C. Future Supply Options

A number of future major supply options have been included in the model. They include:
• Thar Coal
• Tight gas reserves
• Liquefied natural gas (LNG) shipping and regasification process
• Iran-Pakistan and Central Asia gas pipelines
• Regional electricity transmission network.
All of the above options require large up-front investment to ensure access to new energy
resources. These investments have a large minimum threshold, although in the standard
model set-up can be built based on any capacity increments. (The model could treat these
as “lumpy or discrete investment” options, where the project can only be built to a specific
size and smaller incremental investments may not occur). The model can be run using
discrete investments if desired, which is further described in the accompanying User’s
Guide.
19
A scenario template labelled SectorPrices is used in the model to reflect sector price mark-ups.

Thar coal reserves

Whilst use of indigenous coal is relatively low, as shown in the energy balance, very large
reserves of lignite coal are available in the Thar coalfield. Total resources are estimated at
175 billion tonnes, accounting for 95% of Pakistan’s coal resource base. TCEB (2008)
suggests that Thar could be a potentially important fuel source for electricity generation,
providing 100,000 MW (for 30 years) based on an estimated consumption of 536 million
tonnes per year. Given the long term outlook of Pak-IEM, this resource is important to
characterize.
The reserves of this field are available from 2020 onwards (in the model), with cost of mining
derived on the basis of available feasibility studies (PPIB 2008).
Tight gas reserves
Pakistan Petroleum Exploration and Production Companies Association (PPEPCA) and
Schlumberger have recently estimated that tight gas reserves of between 24 and 40 TCF
exist in various basins in Pakistan.20 This reserve has been included in the model although
costs of the gas have been put at double the import price (until this option is better
understood and properly characterized based on additional information). The availability of
the resource is from 2020.
LNG imports
At the request of the Government, Sui Southern Gas Company (SSGC) is developing the
Pakistan Mashal LNG Project to establish an LNG import facility capable of taking 3.5 million
tons per annum (mtpa), equivalent to 500 million cubic feet daily (MMCFD) of gas. The LNG
import project with a re-gasification facility will be located in the vicinity of Karachi. It is
estimated that first imports will arrive in 2011.21
LNG regasification plant costs are sourced from the Energy Information Administration
(EIA).22 Current year and projected prices for LNG have been sourced from WEO 2009 (IEA
2009b). The amount of LNG imported can increase to twice the level in 2020 and three times
the level in 2030 compared to that observed under the Mashal project.
In addition to the above option which feeds into the general gas network, there is also the
option of a specific facility to provide gas for a coastal power plant. Currently, this option has
the same characteristics as the generic LNG option (although the potential capacity in future
years needs to be revisited).
Transnational Gas Pipelines
An Iran-Pakistan gas pipeline is being constructed (much of which is complete on the Iranian
side) which will see a 42-inch diameter pipeline importing 750 MMCFD of natural gas. The
Pakistan section of the pipeline is expected to cost US$ 1.3 billion and take four years to
complete. Gas is expected to be available from this option by 2015, initially at 263 MMCFD,
rising to 750 MMCFD by 2017.23
20
Information provided by DG PC – see Response of DGPC 7-10-2010.pdf in Appendix 3.
21
Information sourced from SSGC website, http://www.ssgc.com.pk/ssgc/projects/lng/
22
See EIA website, http://www.eia.doe.gov/oiaf/analysispaper/global/lngindustry.html
23
Data is sourced from direct correspondence (see file ISGS's Response 28-9-10.pdf) and the ISGS
website, http://www.isgs.pk

Turkmenistan is another major source of gas, and discussions have been ongoing since
2002 on a pipeline between Turkmenistan and India, via Afghanistan and Pakistan (known
as TAPI). Agreement has been made on levels of gas provision, with Pakistan getting 960
MMCFD in Year 1, rising to 1325 MMCFD by Year 3. The Asian Development Bank (ADB) is
acting as the lead developer and coordinator for the project.
The following pipeline options have therefore been included in the model:
• New natural gas pipeline from Iran (available from 2015)
• New natural gas pipeline from Iran (Phase 2), adding capacity equivalent to the first
phase (available from 2020)
• New natural gas pipeline from Central Asia (available from 2025), although this is
currently excluded as an option from the Reference case.
Technical facts about Iran-Pakistan Pipeline

The raw natural gas from the South Pars gas field in the Persian Gulf is brought onshore to Iran’s
port city of Assaluyeh where it is refined and processed to pipeline quality. The Iranian segment of
the Iran/Pakistan/India (IPI) gas pipeline would comprise about 1,150 km long pipeline from
Assaluyeh to the Iran-Pakistan border, travelling along the southern coastal area of Iran. Iran has
already constructed almost 80 percent of the 900 km 56-inch diameter pipeline from Assaluyeh to
Iranshehr the capital of its Sistan/Baluchistan province. The pipeline has the capacity to carry 3.2
Bcfd of gas, out of which 2.1 Bcfd is meant for export. The remaining about 250 km pipeline section
from Iranshehr to Iran-Pakistan border is in the planning and tendering stage, which is expected to
be completed in 2010. Pakistan has planned to build a 42-inch diameter pipeline with design
capacity of 30 MMcmd (1.06 Bcfd) along the coastal line. The approximate length of the Pakistan
segment of the IPI gas pipeline is 800 km from Iran-Pak border to the Pakistan gas off-take point
near Nawabshah where it will be connected with the existing gas distribution network.
Source: Inter State Gas System (ISGS), http://www.isgs.pk
It has not been possible to source the price of gas imports from ISGS. It was recommended
by the oil and gas expert group that such prices be based on 80% of crude price (Japanese
Crude cocktail) as typically reported in press.24
Regional electricity transmission network
Afghanistan, the Kyrgyz Republic, Pakistan, and Tajikistan have been pursuing the
development of electricity trading arrangements and the establishment of a Central Asia –
South Asia Regional Electricity Market (CASAREM). The initial plan is to export power in the
range of 1,000 to 1,300 MW from the Kyrgyz Republic and Tajikistan to Pakistan and
Afghanistan. It is envisaged that the major share of the export will be used by Pakistan, while
a relatively small quantity of power (100 to 300 MW) will be imported by Afghanistan.
Pakistan has also expressed interest in increasing imports over the medium to long term
beyond the initial power requirements of 1,000 MW (ADB 2009).
The total length of the transmission line is 750 km approximately, out of which 16% passes
through Tajikistan, 75% through Afghanistan, and 9% through Pakistan. Transmission line
costs are estimated at US$ 275,000/km. Total costs of the project are put at $US 543
million, with US$ 143 million associated with the Pakistan section, mainly on converter
station / substations. The costs of linking Kyrgyz Republic is a further US$ 172 million. IDC
and contingency takes the total project costs from US$ 715 to 931 million.
24
Pak-IEM oil and gas expert meeting at Serena Hotel (2nd September 2010)

The main uncertainty concerning this option is the price of imported electricity, and the timing
of the actual construction of the grid. The current assumption in the model is that the
CASAREM system would be available from 2015, and that similar additional capacity could
be envisaged by 2020. However, the option is rarely used in the model due to the high price
of electricity assumed in the absence of any additional information.
D. Renewable resources
With increasing pressure on traditional fossil fuel resources in future years, renewable
options are important to consider. Pakistan is, however, already heavily dependent on
renewables, with a very high percentage of the population using biomass energy, particularly
in rural areas. Memon et al (2008) state that the rural population meets more than 95% of
their domestic energy needs by burning biofuels (citing Sial (2002)).
Current biomass consumption
Biomass is not included in the Energy Year Book of Pakistan, as it is often considered a non-
commercial fuel and is also very difficult to estimate. However, it is clear that it is a very
important energy source for many households, particularly in rural areas for cooking and
heating, and therefore needs to be integrated into Pak-IEM.
The approach to estimating biomass totals is as follows.
• Energy Sector Management Assistance Program (ESMAP) (2006) provides data on
the percentage of households in rural and urban areas that use biomass fuels –
wood fuel, dung cake, and crop residues. This shows for example that 78% of rural
households use wood fuel whilst for urban households, it is 27%. This does not
indicate what other fuels are being used, or if indeed this is the main fuel consumed.
• Average annual consumption levels for different fuels (by household) have been
sourced from the Household Energy Strategy Study (HESS) (1991). These levels
and the information on numbers of households using biomass determine the
aggregate level of biomass used.
The HESS data is also used to estimate how biomass types are used, for the provision of
energy services e.g. cooking, water heating etc. Table 7 shows the estimates by energy
service type and biomass type. Biomass is primarily used for cooking.
Table 7: Biomass energy use in Pakistan Rural / Urban Households (toes), 2006/07
Wood Dung Agr. Residues Total

Urban
Cooking 1041 271 136 1448
Space Heating 68 0 0 68
Water Heating 204 0 0 204
Other 23 23 0 45
Total Urban 1335 294 136 1765
Rural 0 0 0 0
Cooking 7692 2036 1561 11290
Space Heating 724 0 0 724
Water Heating 837 0 0 837

Other 68 136 0 181

Total Rural 9299 2149 1561 13009
Total 10633 2443 1697 14774
The HESS survey features heavily in the above estimates, in the absence of any other data.
However, for biomass, it is probable that there have been limited changes in the quantity
used in those households for which it is the primary energy source. There may however,
have been improvements in the efficiency of use, depending on whether improved stoves in
particular have been introduced.
The estimated total of biomass is approximately 650 PJ (--- Mtoe). The share of biomass in
the base year of Pak-IEM as a percentage of household energy consumption is 66% overall,
89% for rural households, and 22% for urban households.
The aggregate estimate of biomass use from the International Energy Agency (IEA) Non-
OECD Energy Balances (2009), which provides estimates of biomass consumption in 2006
and 2007, is 27.2 Mtoes (or 1140 PJ). This is significantly higher than the bottom-up
estimate used in Pak-IEM. The IEA estimate is sourced from HESS and therefore does not
represent new information that should necessarily be considered. It has, however, been
used to derive an estimate of fuel wood used by industry.
Another important factor that needs to be reflected in the modelling is free vs. purchased
biomass. This is important for understanding costs of energy service provision in the future,
and the likelihood of households switching to more modern fuels (due to lack of incentive if
indeed energy is already free). In the model, free biomass is not free in the sense that there
are likely to be costs associated with growing / gathering it, particularly in respect of time.
Therefore, a low cost of 20% of purchased biomass (based on HESS data) is assumed,
important for incentivising switching to more efficient stove technologies. Data from the
Pakistan Integrated Household Survey (PIHS), cited in World Bank (2003), suggests that
60% of rural biomass is free, while for urban areas the value is nearer 5%. These values
have been applied to all biomass types.
Future biomass resource potential and patterns of use
Currently, the model is restricted to base year consumption levels of biomass. The rationale
is that the levels of fuel wood and other biomass consumption are near to their available
potential. An important issue that requires further work is the future rates and timing of
switching away from biomass to more modern fuels.
Both the base year consumption levels and biomass potential estimates have been subject
to significant discussion with experts from the National Agricultural Research Centre
(NARC).25
Potential for other renewable energy
The main other type of renewable energy used in Pakistan is hydro electricity generation.
Other renewable energy types are not so commonplace, although a limited number of
household / community level biogas, micro-wind, micro-hydro, and solar projects have been
implemented in various locations across Pakistan. The installation of a larger-scale wind
power project at Jhimpir, Sindh began in 2009. Initially, this farm will have a 6 MW capacity,
rising to at least 50 MW over the next few years.
25
Biomass potential estimates provided by NARC can be found in document labelled Biomass
resource-Munir.jpg (listed in Appendix 3).

The Alternative Energy Development Board (AEDB)26 is the main expert body on
opportunities for renewable energy development in Pakistan. It has produced a range of
estimates on potentials of different energy types; these provide the basis for estimates of
renewable potential in Pak-IEM.
Table 8: Renewable potentials for Pakistan (Source: AEDB)
Energy Type Estimate (MW)
Wind 340,000
Solar 2.9 million
Hydro (Mini) 2000
Hydro (small) 3000
Hydro (large) 50,000
Bagasse Cogeneration 1800
Waste-to-power 500
Geothermal 550
More work is required in the future to explore how technically / economically feasible these
potentials are, and whether or not they can be represented as resource supply curves.
Note that other than biomass most renewable resources are not characterized in the
resource supply sector but rather in power generation (renewable electricity generation) or
end-use demand sectors (for non-electricity sources or remote generation based
applications). This is because such resources are often dependent on technology
specifications, e.g., type of hydro-turbine technology, size of wind turbine, and the possible
variability in the supply of the resource itself depending upon the technology and local
circumstances.
E. Data Sources
Most of the data for resource supply sector is included in the Pakistan Energy Yearbook; the
exception is information on costs or prices by resource and projected extraction rates of
different resources, as well as the biomass as discussed in the previous section. Table 9
provides a summary of the data sources currently being used.
Table 9: Summary of Data Sources for the Resource Supply Sector
Data Type Data Source Notes and Comments
Coal
Reserves by Field Pakistan Energy Yearbook 2007 Table 4.1
(HDIP 2008)
Current Production Levels and Pakistan Energy Yearbook 2007 Production in Table 4.2; Costs in
Costs by Field (HDIP 2008) Table 4.5
DG Minerals (PDF file Long Term Projections to 2015
Future Extraction Rates and Costs
Coal Projections-DG Mineral
by Field th
provided on 8 April 09
Table 4.5 for indigenous coal;
Pakistan Energy Yearbook 2007
Current Import Levels and Prices Default PET (European)
(HDIP 2008)
assumption for imports.
Thar coal cost estimates PPIB (2008)
26
Alternative Energy Development Board (AEDB) website, http://www.aedb.org/Main.htm


Oil & Petroleum Products
Reserves by Field Pakistan Energy Yearbook 2007 Table 2.1
(HDIP 2008)
Pakistan Energy Yearbook 2007 Table 2.2
Current Production Levels by Field
(HDIP 2008)
Projections to 2035, split into
DG PC data (File Long term Oil
Future Extraction Rates by Field committed and anticipated
Projection-DGPC.xls)
supplies
Pakistan Energy Yearbook 2007 EIA used for imported crude
Current Import Levels and Prices (HDIP 2008), Energy Information prices
Administration (EIA)
Provides some additional level of
Oil Sector Data on Primary Pakistan Oil Report 2006-07
detail to what is found in the
Resources and Refining (OCAC 2008)
Energy Yearbook
IEA (sourced World Energy
Future Crude Oil Prices
Outlook 2009)
Natural Gas
Reserves by Field Pakistan Energy Yearbook 2007 Table 3.1 (Table 3.1.1 for
(HDIP 2008) Associated Gas)
Pakistan Energy Yearbook 2007 Table 3.4 (Table 3.4.1 for
Current Production Levels by Field
(HDIP 2008) Associated Gas)
Projections to 2035, split into
DG PC data (File Long term Gas
Future Extraction Rates by Field committed and anticipated
Projection-DGPC.xls)
supplies
Data (capacity / cost) provided by
Infrastructure Characterization distribution companies (letter from
th
DG Gas dated 26 May 09)
Oil and Gas Regulatory Authority
Current Prices by Field
(OGRA)
Future LNG / Transnational pipeline
ISGS / SSGC websites
options
Electricity (Imports)
Proposed regional interconnection CASAREM project (ADB 2009)
Biomass
Indigenous Supply by Type ESMAP (2006), HESS (1991)
Renewables
AEDB website
Potential estimates
(http://www.aedb.org)
Some of the above sources were not publicly available, and were provided directly to the
project team for use in Pak-IEM. A full listing of these files, and the data providers, is
provided in Appendix 3. The Energy Wing holds all of the data files listed.

IV. REFINING AND OIL & GAS DISTRIBUTION

The upstream sector characterizes the processes for transforming primary energy resources
into petroleum oil products (from crude) and natural gas (from raw extracted gas), and
transportation to end-use sectors. Therefore, there are two main sub-sectors – refining and
oil / gas distribution.
The petroleum and refined products RES in
Figure 9 shows how this sector has been structured in the model. Currently, gas processing
is combined with gas pipeline / network process. Limited account is taken of oil distribution
infrastructure in the model although this is an area that may need to be reconsidered in
future model versions, particularly in view of industry concerns about port constraints, and
differences between the transportation of various products e.g. HFO is only transported by
road.27
In the oil refining sector, each refinery is modeled individually, taking account of plant crude
mix, efficiency, and output product slates.
27
Useful information on distribution and transportation can be found in OGRA (2008)

Figure 9: Current Structure of Oil and Gas Resource Supply and Upstream Processing
Sectors
A. Oil Refining Sector

There are seven refineries in Pakistan. They are listed in Table 10, which includes their key
operating characteristics for the base year 2006/07.
Table 10: Key Operating Characteristics of Pakistan Refineries
Refinery Name 2006/07 Crude 2006/07 Product Efficiency (%) Refining Capacity
Input (PJ) Output (PJ) (Mtoes)
Attock 81.16 76.96 94.8% 1.97
Bosicor 32.54 30.24 92.9% 1.45
Dhodak 4.46 4.75 100.0%** 0.12
ENAR Petrotech 4.19 4.33 100.0%** 0.12
National 126.92 101.81 80.2% 2.87
PakArab 169.58 162.68 95.9% 4.50
Pakistan 90.29 88.33 97.8% 2.10
Source: HDIP 2008
* Capacity value for National Refinery (119.78 PJ) is assumed to be equal to crude input values.
** Efficiencies for Dhodak and ENAR have been set to 100%; data currently puts the efficiency at 107% and
103% respectively.
An important assumption in the model concerns the type of crude that domestic refineries
use. Attock and Dhodak refineries primarily use domestic crude oil in refining operations.
Other refineries are limited to using between 15-20% domestic crude, as most are set up for
using higher quality Arabian crude. This is reflected in the assumptions concerning shares of
crude used in future years (see Table 11).
Table 11: Crude Input Shares for Pakistan Refineries
Refinery Name Domestic Crude Imported Crude Max. Future Max. Future
Share Share Domestic Crude Imported Crude
Share Share
Attock 1.00 0.00 1.00 0.40
Bosicor 0.00 1.00 0.20 1.00
Dhodak 1.00 0.00 1.00 0.40

ENAR Petrotech 0.00 1.00 0.20 1.00

National 0.16 0.84 0.20 1.00
PakArab 0.10 0.90 0.20 1.00
Pakistan 0.11 0.89 0.20 1.00
Source: HDIP 2008 (for current shares); Communication with Refinery stakeholders for future shares.
The product output shares from each refinery in 2006/07 are shown in Figure 10 below.
100%
90%
80%
70% Naptha
Lt Diesel
60%
LPG
50% Av. Fuel
Kerosene
40%
Heavy Oil
30%
Diesel
20% Gasoline
10%
0%
Attock Bosicor Dhodak ENAR National PakArab Pakistan
Figure 10: Product Shares for Refineries in 2006/07

Source: HDIP (2008)
In future years, existing refineries are allowed overall flexibility of 10% across their product
slate (an assumption cross-checked with refinery experts), except for gasoline, where a 4%
relaxation factor has been assumed. A rule has also been introduced that states if the share
of a product is less than 3% in the base year, it will be permitted to produce up to 3% of this
product in future years. Information provided by Attock and Pak Arab refineries resulted in
being able to calculate the product slate flexibility, at a similar level in aggregate terms to the
assumption on flexibility for other refineries.
The existing refinery stock (in crude input capacity) has been held constant to 2040, based
on the assumption that refineries tend to be retrofitted on an ongoing basis and therefore will
remain in the energy system. The costs of such retrofitting activities are not known across
the whole stock, although some information has recently been provided by selected
refineries, and should be considered in future model updates (see Model Development
Priorities, section VIII for further information).
In the development of the model, there was significant discussion with experts concerning
refinery margin. A significant element of this is the operating cost. For existing refineries we
are using $1.25 / bbl although some experts have suggested this should be potentially
higher. To date data has not received from individual refineries to differentiate this cost
estimate.
Planned new refinery builds, referenced in OCAC (2009) and listed in Table 12, are “forced”
into the model solution in the year in which they are expected to be commissioned. For

Bosicor, investment costs have been provided for a second-hand refinery build. These have
been used for Transasia and Indus, although with a 20% multiplier due to relative simple
configuration of Bosicor. An operating cost of $1/bbl has been assumed.
Table 12: Planned New Refineries
Refining capacity Investment cost, $ Commissioning
Planned Refinery
(BPD) /BPD Date
Bosicor Oil Pakistan Limited* 115,000 5000 2010
Transasia Refinery Limited 100,000 5215 2013
Indus Refinery 250,000 5215 2014
* Investment cost information provided for Bosicor based on 2nd Hand Refinery costs. Equivalent to ~$5000/bbl.
All refineries assumed to primarily use imported crude.
The product slate for the planned refineries is based on information in OCAC (2009). Khalifa
Coastal refinery was dropped as a planned refinery due to the significant uncertainty around
whether it will be built.28
There is a single new refinery technology in the model, only available in the model from 2017
after the planned refineries have been built. This technology is relatively flexible, with a
product slate that is typical of the product slates observed for the planned build refineries.
Investment costs are based on the proposed Khalifa refinery, reported to cost $5 billion for a
250,000 bbl / day capacity. This is estimated at over $8 million per PJ. UK / US model
sources put new refinery costs at around $5 million per PJ. Based on discussions with the
refinery expert group, it was decided to use an average investment cost of around $6.75
million per PJ.
From 2017 to 2040, additional new capacity is capped at one new refinery with a capacity of
100,000 bbl per day every three years. This results in new capacity of over 1500 PJ by 2040,
which is at the optimistic end of what could happen, based on discussion with the expert
group.
B. Distribution of oil and gas

In Pak-IEM, the distribution infrastructure (capacity or costs) for crude oil / petroleum
products are not modeled explicitly. Oil and gas transportation costs are taken into account,
based on the additional costs seen by refineries and end-use sectors.
For gas, the pipeline network is critical as the primary means of transportation. Capacity of
the system is based on comparing the flow of gas in the base year (2006-07) and utilization
of the system for that year. There are two primary gas networks, Sui Southern (SSGC) and
Sui Northern (SNGPL), reflecting both transit pipelines and the capacity of the associated
distribution network. For both networks, utilization is near 100% across most parts of the
network; therefore, the capacity is based on the 2006-07 flow of gas, with an assumed 5%
spare capacity. In future years, exceeding 2006-07 levels plus the 5% extra capacity will
require additional investment in the system. Costs for incremental capacity are being sought
28
All of the assumptions around the refinery sector assumption were discussed in detail at a Pak-IEM
expert meeting on September 2, 2010.

but are not yet finalized in the model.29 Furthermore, a nationally integrated gas network is
assumed owing to the fact that SSGC / SNGPL are well connected.
C. Data Sources
The key sources of information for the upstream sector are listed in Table 13. Information is
primarily sourced from the Energy Yearbook, supplemented by the Pakistan Oil Report
(OCAC 2008).
Table 13: Summary of Data Sources for Refining and Gas / Oil Distribution Sector
Refinery capacities Table 2.3
(HDIP 2008)
Crude oil processed Table 2.4
(HDIP 2008)
Refinery specific data Tables 2.5 and 2.5.1-7
(HDIP 2008)
Bosicor also provided cost
Data provided by two refineries –
Refinery flexibility / operating costs estimates regarding plant
Attock, Bosicor and National
capacity upgrade
Planned new build Pakistan Oil Market Report 2007-08
Data provided directly (see file
Letter & Data from DG Gas.pdf,
Gas network capacity / utilization SSGC / SNGPL
and Response of DG Gas 4-10-
2010.pdf)
provided in Appendix 3. The Energy Wing holds all of these data files.
29
Costs of system expansion have been provided by SSGC and most recently by DG PC, although
these are in terms of pipeline length rather than energy delivered to the system, which is what is
needed. Further work is needed to assess how to use such information. See file Response of DG Gas
4-10-2010.pdf, listed in Appendix 3.

V. POWER SECTOR
The Pakistan power generation and distribution sector in 2006/07 has been modeled in
detail, with individual power plants characterized based on their historic operating
characteristics. Problems with electricity generation in Pakistan, in particular load shedding
due to lack of capacity at high demand and peak periods, means that detailed modeling of
this sector is important for stakeholders who will use the model and / or its outputs to think
through planning options in the medium and long term.
A. Sector overview
Coal, 0.01
Coal, 0.07 Nuclear, Imports,
0.19 0.01
Oil, 2.28
Hydro,
Furnace
2.60
Oil, 6.52
Gas, 8.64
Gas, 2.92
Diesel Oil,
0.04
(a) (b)
Figure 11(a) shows that fossil energy requirements for the power sector are
dominated by natural gas and furnace oil.
Coal, 0.01
0.19 0.01
Oil, 2.28
Hydro,
Furnace
2.60
Oil, 6.52
Gas, 8.64
Gas, 2.92
Diesel Oil,
0.04
(a) (b)
Figure 11(b) shows that over 90% of all electricity generation is provided by these oil and
gas plants along with hydropower plants, which provide about 33% of all electricity.

Coal, 0.49
8.24 0.62
Oil,
100.88
Furnace Hydro,
Oil, 115.02
288.25
Gas,
381.89
Gas,
128.91
Diesel Oil,
1.95
(a) (b)
Coal, 0.01
0.19 0.01
Oil, 2.28
Hydro,
Furnace
2.60
Oil, 6.52
Gas, 8.64
Gas, 2.92
Diesel Oil,
0.04
(a) (b)
Figure 11: 2006/07 (a) Power Sector Fossil Energy Use (Mtoes)
and (b) Electricity Generation by Plant Type (Mtoes)
All large centralized hydropower generation is overseen by Water and Power Development
Authority (WAPDA). Thermal generation in the model can be categorized into three groups:
1) public sector thermal plants operated by PEPCO, 2) independent power producers (IPPs),
and 3) plants operated by Karachi Electricity Supply Company (KESC).
B. Approach to modeling the power sector generation

The approach to modeling the above groups of generation plant differs, primarily due to the
type of plant and availability of data. However, the types of data required to characterize
power plants in Pak-IEM apply across all of these groups are listed in Table 14.

Table 14: Data Requirements for Power Plant Characterization

Characteristic Unit Comment
Name
Fuel(s) used
Capacity MW Electric rating only
Remaining life Years TIMES can also use the start year for the plant
Efficiency (by fuel) OUT/IN
Fixed O&M Cost $M/MW
Variable O&M Cost $M/PJ
Availability factor Annual or seasonal Hydro, wind and solar plants have seasonal availability
factors.
Fuel Input Share(s) Fraction Needed for multi-fuel plants.
Output to Power Ratio Fraction Needed when non-electric outputs are also produced,
such as heat (co-generation plant), fuels or water.
Investment cost for new $M/MW
options
Table 15, Table 16, Table 17, and Table 18 provide an overview of the modeling approaches
taken for the PEPCO, IPP, KESC, and WAPDA power plants respectively. Generation from
decentralized wind and solar generation are also included in the model, although they are
not detailed here due to their relative low generation output (2006/07 total of 600 MWh).
Table 15: Model Assumptions for PEPCO Thermal Power Plants
Parameter Description of approach Source
Existing Installed Established at the “block” level Data provided directly by PEPCO (see file
Capacity (aggregation of similar units of same PEPCO THERMAL POWER STATIONS1.xls)
type and fuel(s) used) using the
installed de-rated capacity
Input Fuel Available by unit, aggregated by block t Data provided directly by PEPCO (see file
Consumption for consistency Planning Commission 20-03-09.xls)
Fuel Shares for Evaluated based on block fuel Data provided directly by PEPCO (see file
Duel Fuel Plant consumption Planning Commission 20-03-09.xls)
Efficiency Evaluated as the ratio between block Data provided directly by PEPCO (see file
production and consumption. For multi- Planning Commission 20-03-09.xls)
fuels plants we assumed the same
efficiency for both fuels
Utilization Factor Evaluated on the basis of the installed Data provided directly by PEPCO (see file
de-rated capacity and block production Planning Commission 20-03-09.xls)
in 2007
Availability factor As per base year value A constraint has been put in the model that
(for future years) ensure operation of plant at least at 2006/07
levels (see template Scen_REF-ELCvXX.xls)
Remaining lifetime Retirement schedule provided by Data provided directly by PEPCO ( see file
of plant PEPCO Response to Remaining Power Sector Issues
3-2-10.pdf)
Operating costs The O&M costs are calculated from the Cost calculations undertaken by project team
various tariff decisions of NEPRA
related to PEPCO thermal power plants
and IPPs
Potential for Rather than invest in new plant, an Three plants are being considered for retrofit;

retrofit important alternative option might be other existing plant do not have this option
the retrofit of existing plant
Table 16: Model Assumptions for IPP Thermal Power Plants

Existing Installed Plant level capacity (de-rated) Data provided directly by PEPCO (see file
Capacity Data of Existing IPP.xls)
Efficiency Plant level efficiency adjusted based Data provided directly by PEPCO (see file
on calibration to get correct fuel Data of Existing IPP.xls)
consumption totals
Utilization Factor Calculated using energy use by plant Data provided directly by PEPCO (see file
from Electricity Market Data Data of Existing IPP.xls) plus Electricity
publication (and capacity as provided Market Data report
above) because IPP data is for 2005-
06
Availability Factor As per base year value A constraint has been put in the model that
levels
Remaining Lifetime Based on generic technical lifetime of Planning Team advised on lifetime level
of Plant technology type, accounting for start
year of plant.
Operating Costs The O&M costs are calculated from Cost calculations undertaken by project team
the various tariff decisions of NEPRA
related to PEPCO thermal power
plants and IPPs
Table 17: Model Assumptions for KESC Thermal Plants

Existing Installed Plant level capacity Data provided directly by KESC (see file PIEM
Capacity KESC data.pdf)
Efficiency Plant level efficiency Data provided directly by KESC (see file PIEM
KESC data.pdf)
Utilization Factor Plant utilization in base year Data provided directly by KESC (see file PIEM
KESC data.pdf)
Availability Factor As per base year value A constraint has been put in the model that
levels
of Plant technology type, not accounting for
start year of plant.
Operating Costs The O&M costs are calculated from Cost calculations undertaken by project team
the various tariff decisions of NEPRA
related to PEPCO thermal power
plants and IPPs
Table 18: Model Assumptions for WAPDA Hydropower Plants

Existing Installed Existing installed generation capacity Electricity Market Data report
Capacity
Utilization Factor Seasonal availability factors provided for See file Power Sector Load Shedding &
2006-07 (and 2 subsequent years) Captive Power Data.pdf
Availability Factor Seasonal availability factors provided for See file Power Sector Load Shedding &


(for future years) 2006-07 (and 2 subsequent years)* Captive Power Data.pdf
of Plant technology type, accounting for start year
of plant; assessing need to increase
typical lifetimes so Pakistan-relevant
Operating Costs Derived from tariff set by WAPDA where. Cost calculations undertaken by project
one composite tariff is awarded by team (but could be subject to update to
NEPRA for all hydro power plants, make plant specific)
therefore same value applied to all plants
Many of the existing thermal plant would not continue to operate due to the very low
efficiencies observed, unless forced to do so. In the current model Reference case, many of
these plants are forced to operate at least at the level that was observed in the base year,
2006/07, which seems reasonable, based on the realities on the ground.
C. Planned new build

The Pakistan Government has a plan to install significant levels of new generation capacity
over the next 10 years. The approach taken in this model is to ensure any thermal plant
proposed for commissioning by June 2010 are “forced” into the model. Planned hydro plants
are also included in the model even where commissioning dates are post-2010, if the
completion rate of construction is greater than 50%.
A total of 7.9GW of planned capacity is included in the model Reference case, as shown in
Figure 12 below. Of this 1.3GW of capacity is short-term (3-5 years), based on rental
contracts. Data has been provided directly by PEPCO and KESC.30
Cost data for planned new builds are based on the assumptions provided in the next section,
as new technology options. For rental plants, costs assumptions are from a recent review by
the ADB (2010).
Nuclear, 325
Hydro, 1421
IPPs, 2690
KESC, 960
GENCO, 1175
Rentals, 1332
30
PEPCO information in letter Response to Remaining Power Sector Issues 3-2-10.pdf. KESC
information in KESC data sheet.doc.

Figure 12: Planned New Generation Capacity, MW
D. New technology options

In the medium to longer term, a full range of power generation options are included in the
model. The assumptions for these different options are listed in Table 19.
Estimates of the cost and performance of future power generation technology options are
based on a range of sources, including US data from the Assumptions to the Annual Energy
Outlook 2010 (EIA 2010), European data from the IEA Projected Cost of Electricity 2010
Edition (IEA 2010), and Pakistan-specific estimates from AEDB and others. Data from these
sources were compared for each technology option, and the data most appropriate to
Pakistan was selected based on review with the Planning Team. In many cases, the local
and international data sources are in good agreement. In a few cases, Pakistan specific
factors resulted in selection of the Pakistan data or modification to the international values.

Table 19: New technology options in the Power Sector

Fuel Start Investment Fixed Variable Construction Discount Annual
Technology Efficiency Lifetime
Type Year costs O&M O&M time rate AF**
% Years $/kW $/kW $/GJ Years % %
New Gas Turbine Open Cycle - Gas or Fuel Oil NGA 2011 32.7% 30 709 11.79 0.96 2.00 10% 70%
HFO 2020 38.8% 674
2030 45.0% 638
New Heavy Oil Reciprocating engines with
HFO 2010 48.0% 30 1200 16.92 1.72 3.00 10% 70%
Combined cycle
New Heavy Oil Steam turbine with Reheat Cycle 1 HFO 2010 40.0% 40 1000 17.52 0.67 3.00 10% 70%
New Coal Integrated Gasification Combined Cycle COA 2016 45.8% 40 2898 37.63 0.79 4.00 10% 75%
2030 2608
New Gas turbine Combined Cycle - Natural gas
NGA 2010 47.4% 30 1063 12.15 0.56 3.00 10% 70%
and diesel
HFO 2020 53.7% 1010
2030 60.0% 957
New Gas turbine Combined Cycle with Reheat -
NGA 2011 54.0% 30 1200 15.84 0.83 3.00 10% 70%
Natural gas and diesel
HFO
New Gas turbine Combined Cycle - LNG
LNGP 2010 47.4% 40 1169 26.79 1.24 3.00 10% 70%
dedicated
2020 1111
2030 1052
New Nuclear Power Plant NUCD 2011 34.1% 40 4501 87.60 0.13 6.00 10% 85%
2020 4051
2030 3601
New Coal Steam Turbine Plant (using imported
COAP 2010 37.1% 40 2000 26.79 1.24 3.00 10% 75%
coal)
2020 1900
2030 1800

New Coal Supercritical Steam turbine Power Plant

COA 2010 37.1% 40 2560 26.79 1.24 3.00 10% 75%
(Thar)
2020 43.5% 2432
2030 50.0% 2304
New Hydro Power HYD 2016 34.1% 50 1911 53.19 1.43 4.00 10%
New Solar Photovoltaic Systems SOL 2011 34.1% 20 6390 11.37 0.00 10%
2020 5432
2030 4473
New Solar Thermal Power Systems SOL 2011 34.1% 30 3500 19.08 0.67 3.00 10%
2020 3255
2030 2975
New On Shore Wind Turbines-Classification 4 WIN 2012 34.1% 20 2500 63.24 0.11 2.00 10%
New Off Shore Wind Turbines WIN 2012 34.1% 50 3700 87.84 0.19 2.00 10%
2020 3441
2030 3145
New Coal IGCC with Carbon Capture &
COA 2017 41.1% 40 4260 44.88 1.20 3.00 10% 75%
Sequestration
2030 48.0% 3408
New Adv Gas turbine CC with Carbon Capture &
NGA 2017 39.6% 30 2088 19.37 0.80 3.00 10% 70%
Sequestration
2030 56.0% 1775
New Municipal Solid Waste MSW 2010 25.0% 30 2809 111.18 0.00 3.00 10% 70%
* Weighted average. The costs of new hydro are project specific but are not listed in detail in this table.
** Renewable generation options have seasonal or diurnal based AFs, not shown in this table.

Hydro generation potential in the model is shown by site in Table 20 below. The total
potential is around 35 GW, with costs differing significantly depending on site-specific factors
and the latest feasibility assessment. Seasonal availability factors are the same across all
sites, sourced from the power sector experts on the Planning Team.
Table 20: Potential hydro sites in Pakistan
Capacity, Investment cost,
Hydro site Year available
MW $M/GW
Diamer Basha 2020 4500 1730
Golen Gol 2012 106 1382
Kurram Tangi 2013 83 7582
Munda 2017 740 1145
Kohala 2017 1100 1844
Keyal Khwar 2016 122 1123
Phandar 2016 80 824
Basho 2015 28 1176
Lawi 2016 70 1188
Dasu 2021 4320 1699
Bunji 2022 7000 919
Akhori 2024 600 5176
Lower Spaigah 2021 567 1019
Palas Valley 2022 621 1011
Tarbel 4th extension 2014 960 691
Harpo 2015 33 1283
Pattan 2025 2800 2017
Thakot 2025 2800 2017
Dudhnial 2026 800 2118
Yulbo 2028 3000 2118
Tungas 2030 2200 1797
Skardu 2030 1600 4853
Yugo 2030 520 5430
Source: WAPDA (see file Response to Remaining Power Sector Issues 3-2-10.pdf)
Availability factor for other renewables such as wind and solar have been provided for all
timeslices, taking account not only of seasonal differences but also variation across the day.
Wind technologies have been categorized into three onshore types and a single offshore
technology. These represent the differences in the wind resource (as reflected in the
different capacity factors used), sourced from recent work by USAID / NREL.31
Carbon Capture and Storage (CCS) options are also included in the technology dataset, as
shown in Table 19. The structure for modeling the CO2 storage approach is described in the
Users' Guide. Information on the costs and potential for storage still need to be developed.
The options in Table 19 only relate to public supply of electricity. Captive generation options
in industry are described in the industry section of this report, as are microgeneration options
in residential / commercial sectors.
31
Wind Resource Assessment and Mapping for Afghanistan and Pakistan, Undertaken by National
Renewable Energy Laboratory (NREL) under the SARI-Energy project (funded by USAID).)

The maximum level of investment in new generation technologies is controlled in the model
to reflect barriers faced in the real world, such as availability of capital, technical capacity for
construction, and supply chain constraints. Annual build rates have therefore been
introduced in the model’s Reference scenario (see Figure 13). This figure compares annual
build rates in 2015 with those in 2035, which are generally higher. These rates considered
historical patterns of new build, and were developed in discussion with the Planning Team.
3
2.5
2
GW / yr
1.5
0.5
0
Nuclear Oil Gas Coal Hydro Solar Wind (total)
Figure 13: Build rates (GW) per year (2015/2030) by Generation Type
E. Transmission and Distribution System

The modeling of the power sector in Pak-IEM, including electricity transmission and
distribution to end-use sectors, is shown in Figure 14 . Electricity generated from large-scale
plants operated by PEPCO, KESC, WAPDA and IPPs, produce centralized electricity
(ELCC). This then feeds into the transmission grid (labeled ELCT in the model) and then a
general distribution system (becoming ELCD). ELCT and ELCD are subject to losses
associated with the grids, and incur costs associated with maintaining and operating the
network. In addition, they are subject to capacity constraints that can only be overcome by
additional investment in new capacity.
Combined losses on both the transmission and distribution system were estimated to be
24% in 2006/07. Transmission losses are estimated to be 6%. Distribution losses include
technical and non-technical losses (theft and non-payment), and while the split between
such losses are uncertain, it is thought that non-technical losses account for over half of the
distribution losses. In the Reference scenario, the level of these losses is held constant in
future years, with no options currently modeled for reducing such losses.32 Current grid
capacity is assumed to be a function of the total installed generation capacity, with each unit
of additional generation capacity incurring investment costs to build extra transmission
32
A scenario labeled Pol-T&D is available to model a reduction in T&D losses over time.

capacity. Utilization of the existing system is set at 95%, whilst investment and operating
costs for additional capacity above the existing stock are based on information provided in
ADB (2008).
Distributed electricity (ELCD) feeds into the different sectors, and becomes <sect>ELC,
which is prefixed by the sector abbreviation, e.g. RSDELC for electricity to the residential
sector. Sector-specific differences in distribution of electricity can be captured at this level.
Thus non-technical losses to sectors can be modeled in the future, if corresponding data can
be provided. The same mechanism can be used to reflect differential electricity tariffs for
different sectors.
Electricity produced by small-scale decentralized power plants (including some renewables),
imports,33 and back-up generation are labeled ELCR (or remote generation). They do not go
through the main transmission and distribution system because they are connected in close
proximity to the customers being served. Relative shares, currently under review, are used
to apportion this type of generation to the different end use sectors. Figure 14 depicts the
stylized RES for the power sector (excluding industry cogeneration), reflecting the energy
forms, power plant types, and sector distribution nodes.
As discussed in Section III, an important option in future years may be imported electricity
via a regional transmission grid.
33
Imports from Iran are distributed onto a remote grid, which is not connected to the main transmission and
distribution network in Pakistan and meets the needs of specific communities in Balochistan.

Figure 14: Distribution of Electricity from Centralized and Decentralized Power Generation and Imports
F. Load shedding
Load shedding has become a real problem over the last three years, as shown in Figure 15
below. Estimated demand far exceeded what the system provided (shown by the blue
shaded area), particularly during the summer months. This has also been a significant
problem for the KESC system.34
14000
Demand
12000 System
10000
8000
GWh
6000
4000
2000
0
Jul 06 Oct 06 Jan 07 Apr 07 Jul 07 Oct 07 Jan 08 Apr 08 Jul 08 Oct 08 Jan 09 Apr 09
Figure 15: Load shedding on the PEPCO system, July 06-June 09
Given the significant economic problems arising from unreliable supply, it is important that
this near term issue is properly modeled. The approach taken is illustrated in Figure 16.
Projection of the demand for energy services has been developed from basic drivers that
factor in the recent load shedding. This is necessary so as not to bias the projections by
using the current suppressed levels of consumption. Thus, the top line in Figure 16
represents the projected demand, while the blue area (to 2009) represents recent data on
actual consumption. The deficit (unmet demand) between 2007-2009 is shown, identified by
the white arrows.
34
Data provided directly to modelling team by PEPCO / KESC. See file Power Sector Load Shedding
& Captive Power Data.pdf (PEPCO) and KESC data sheet.doc (KESC).

140000
Deficit
130000 Actual supply
120000
GWh 110000
100000
90000
80000
70000
2007 2008 2009 2010 2011 2012 2013 2014 2015
Figure 16: Pak-IEM approach to modeling load shedding
The amount of unmet demand in future years is a model input, where the analyst can control
the rate at which the unmet demand must be satisfied. In the example above (which is the
model Reference scenario), the deficit is set to be eliminated by 2015. The analyst can
shorten or lengthen this period or change the rate of removal or amount of unmet demand in
a period as sensitivities to the Reference case.
To ensure that the investment cost of removing the deficit in electricity supply is captured,
the deficit is modeled as a “dummy” commodity, and forced into the base year demands.
The commodity has to be forced using lower bounds, as the unmet demand is costed at a
high level. The cost of load shedding was $0.44 /kWh from the 1989 paper “Power Outages
in the Industrial Sector of Pakistan.” This has been adjusted to $0.6 /kWh to put it on a
$2006 basis.
This commodity is shared across demand sectors based on their share of electricity
consumption in the base year. As this commodity is removed (i.e. the problem starts to be
addressed), the system has to respond to fill the energy gap, therefore requiring investment
either in new generation capacity or other energy types that could meet the energy service
demands.
G. Summary of Main Power Sector Data Sources

Table 21 lists the main data sources for the power sector.
Table 21: Summary of Data Sources for the Power Sector
Thermal / nuclear plant
Public plant – PEPCO See Table 15
Private plant – KESC See Table 17
Private operated plant –
See Table 16
IPPs
Hydropower plant
Plant characterization See Table 18


Non-hydro renewable
generation
Pakistan Energy Year Book 2007 Resource information for future years
Capacity and generation
(HDIP 2008) sourced from AEDB website
Transmission and
distribution
Pakistan Energy Year Book 2007
Losses
(HDIP 2008)
Based upon current installed power
Capacity
plant capacity
Operating / investment
Data from ADB (2008)
costs
See file Power Sector Load Shedding &
Data provided directly by PEPCO /
Load shedding Captive Power Data.pdf (PEPCO) and
KESC
KESC data sheet.doc (KESC)
Future options
See file Response to Remaining Power
List of future generation projects
Planned new build Sector Issues 3-2-10.pdf (PEPCO) and
provided directly by PEPCO / KESC
KESC data sheet.doc (KESC)
Based on review of number of International data sources included IEA
New technology options international sources, and Pakistan (2005b), AEO (2010), ETSAP information
specific data and data from European NEEDS model35
provided in Appendix 3. The EW holds all of these data files referred to.
35
ETSAP (Energy Technology Systems Analysis Program) information can be found at ETSAP
website – http://www.etsap.org/index.asp

VI. DEMAND SECTORS

Demand sectors characterize the consumption of energy services across the economy, for
example, for the provision of cooling and heating, transportation, and industrial production. In
Pak-IEM there are five demand sectors – agriculture, commercial, industry, residential, and
transportation. Demand sector energy consumption in Pakistan in 2006/07 is in the main
based on estimates from the Energy Year Book (HDIP 2008). However, there are some
significant differences between Energy Year Book sector estimates and those used in the
model, due to the inclusion of biomass in residential and industry sectors, adjustments for
black market diesel, and the re-allocation of diesel from transport to the agriculture sector.
The proportion of energy use across the different demand sectors, as presented in the
Energy Yearbook, is shown in Figure 17(a), while the Figure 17(b) shows the allocation in
Pak-IEM.
Other Govt. (a) Energy Yearbook sector shares

2%
Domestic
21%
Transport
27% Commercial
4%
Agriculture
2%
Industrial
44%
Total Final Energy

1591 PJ / 36 Mtoe
36
(b) Pak-IEM sector shares
Other Govt,
1.3%
Transport,
14.2%
Agriculture,
6.3%
Domestic,
41.0%
Industrial,
34.6%
Commercial,
Total Final Energy
2.5%
2411 PJ / 54 Mtoe
Figure 17: Final Energy Consumption by Sector, 2006-07
36
Total energy in Pak-IEM is higher due to inclusion of estimates of biomass use in industry and black market
diesel.

In total, as shown in
60
Final energy consumption, Mtoes 50
Other Govt
40
Transport
Agriculture
30
Industrial
Commercial
20
Domestic
10
0
EYB‐07 Pak‐IEM
Figure 18, final energy is 50% higher than that observed in the Energy Year Book. This is
primarily due to inclusion of biomass in the residential (+650 PJ) and industry sectors (+75
PJ wood and 50 PJ bagasse) plus an increase of 10% of diesel (HSD) consumption due to
black market sales.

60
Final energy consumption, Mtoes 50
Other Govt
40
Transport
Agriculture
30
Industrial
Commercial
20
Domestic
10
0
EYB‐07 Pak‐IEM
Figure 18: Comparison of Final Energy Consumption estimates

between Energy Year Book 2007 / Pak-IEM
A. Modeling approaches
The sector final energy consumption values from the Energy Yearbook provide the basis for
calibration of the model’s start year of 2006/7. However, this sector-level energy
consumption data has to be developed for the individual sub-sectors, and perhaps sub-
groups, according to the nature of the different energy service activities in each demand
sector, such as pumping and plowing (agriculture), lighting and space conditioning
(commercial and residential), process heat and machine drive (industry), and passenger and
freight transport (transportation). These energy services need to be developed as sub-
sector demands, as each serves a unique (non-substitutable) demand and requires a
different set of technologies (end-use devices) to provide those services.
It is important to distinguish that MARKAL/TIMES, unlike many other models, uses the
concept of energy service demand, not final energy demand. This enables the model to
evaluate fuel mix and device choices reflecting industry and consumer choices for meeting
future service level.
Energy service demands can be specified in whatever units are convenient (e.g., vehicle-
miles traveled for transportation, lumens for lighting, BTUs for space cooling, etc.). In these
cases, the end-use devices must be defined in terms that convert final energy carriers (in
PJ) into these energy service units.
There are two general approaches used to develop the sub-sector level breakout for each of
the demand sectors: top-down decomposition and bottom-up aggregation. Both approaches
involve a number of steps, which are illustrated in Figure 19. The approach chosen for a

particular demand sector depends on the data availability, and is discussed under each
sector.
For most sectors, a top-down decomposition approach is necessary because the existing
stock of end-use devices and the amount of fuel each consumes is not known. In this
approach, decomposition shares are applied to the sector energy consumption totals to
determine sub-sector totals for each fuel type, which are then further broken down into the
individual end-use services, perhaps with some intermediate splits as well (e.g., residential
into rural/urban, particular industries).
These shares are a straightforward mechanism employed to disaggregate the sector level
fuel use to each sub-sector demand, though they often need to be determined by surveying
or expert judgment. Data on sub-sector final energy consumption can be estimated from
basic data such as number of households and final energy consumption per household. This
data may also be estimated from general information on energy consumption patterns (e.g.,
per capita consumption) in similar countries in the absence of country-specific data.
For other sectors, shares are estimated using a bottom-up approach, based on technology
stock and use/performance information to aggregate up to the sector energy consumption
values. In this case the device details are aggregated and crosschecked against the sector
energy balance. For Pak-IEM, the bottom-up approach has been employed for agriculture
and transportation sectors.
Subsector demand modeling for

base year 2006/07
Top‐down approach
Decomposition Sector energy

shares ‐ subsector consumption
Calibration check
Proxy data
Decomposition Subsector energy Bottom‐up approach

shares ‐ technology consumption
Stock‐based
estimates
Technology stock
by type
Technology type
energy consumption Efficiency
Decomposition
Stock‐based Activity level
Figure 19: Bottom-up and Top-down Approaches for Demand Sector Modeling
For each demand sector the number of sub-sectors is based on the need to adequately
model the energy system at the appropriate level of detail, when supported by data. An
example of possible sub-sector detail might be commercial building types (which can be
classified according to purpose (e.g., government, hospital, school, hotel, restaurant, and
other commercial services) or size (e.g., large, small), or residential location (e.g., urban or
rural) or household income class (e.g., low, medium, high). These additional breakdowns are
important to consider and represent when the energy use patterns and/or policies to be
examined warrant separation into these finer groupings. The goal is not simply to add detail,
but rather to develop the level of detail that is supported by the available data and

appropriate for the intended uses of the model. At the current time the commercial sectors
are modeled only at the aggregate sector level and residential simply split between
urban/rural, with appropriate demand services (e.g., cooling/heating, lighting, refrigeration,
cooking), until a more solid understanding of a finer breakdown is supported by the data, or
developed by means of surveying. In particular, investigation is underway looking to further
breaking down the residential sector according to income class, as the nature and expected
growth of the demand for energy (and devices deployed) varies substantively according to
income level.
For each of the demand services, information on the demand timing is required to properly
distinguish the time-of-use of the demand for commodities such as electricity. The demand
timing splits determine the amount of each demand service that must be satisfied in each
season and time-of-day time-slice. In order to properly represent these patterns, data on
daily and seasonal electricity consumption (and seasonally for gas) by end-use category is
needed or needs to be derived / estimated. In particular, utility load curves by sector and by
sub-sector would be ideal. Data on the effects of climate on daily and seasonal demands
can be useful (that is, degree day information for the calibration and typical year). Data on
industrial practices (e.g., shifts per day, seasonal activities) on daily or annual demands for
electricity is also useful. However, to date only limited information on these fronts has been
secured, and thus consideration needs to be given to surveying, particularly in the residential
sector.
For certain sub-sectors where the demand is uniform throughout the year, which is often the
case for industrial facilities that have continuous production, it is assumed that the demand
follows the overall fractional split for season and time-of-day. However, if a demand typically
does not follow those splits (e.g., residential lighting is mostly at night, residential and
commercial space cooling mainly in the summer), then demand service timing splits need to
be specified. An approach has been set up by the modeling team that allows for the
calibration of demand fraction shares to accurately represent the daily and seasonal load
curves in the model, as was discussed in Section II.C.
End-use technologies to meet the demands for energy services need to be fully
characterized. Table 22 lists the main parameters that are needed for each device. These
are considered in greater detail in each of the sector-specific sections.
Table 22: Technical Parameters for Characterization of End-Use Technologies
Technology Name
Input Energy carrier
Capacity PJ/a
Remaining life (existing) Years TIMES can also use the start year for existing devices
Lifetime (new) Years New devices
Efficiency OUT/IN
Investment cost $M/PJa New devices only
Fixed O&M Cost $M/PJ
Variable O&M Cost $M/PJa
Utilization factor Annual
Input share Fraction For duel-fuelled technologies
Output share Fraction For technologies providing more than one energy
service e.g., heat pump providing heating and cooling,
car doing short/long-distance driving

B. Agricultural Sector
Agriculture is one of the largest sectors of the economy in Pakistan, contributing 21.8% of
GDP, and an even larger share of employment, providing an income for 44.7% of the total
labor force (Pakistan Economic Survey 2008-09, GoP 2009). Due to the size of the sector,
energy use is significant (over 6% of final energy consumed), even if agricultural production
is relatively non-energy intensive. The main demands for energy include:
• Water pumping for irrigation
• Motive power provided by tractors and self-propelled field machinery for tilling,
cultivating, and harvesting
• Tractor-based transport of goods
The model depicts the above demand services, plus a general “other” energy use category,
primarily to capture unallocated electricity use (e.g., lighting, heating and cooling for storage
buildings, etc.).
Energy use in agriculture
The sectoral energy use in the base year (2006/07) totals 34 PJ (based on EYB 2007),
excluding any high speed diesel use, which is included in the Energy Yearbook transport
sector estimates. Bottom-up estimates for Pak-IEM, based on technology stock data (ACO
2005), put agricultural HSD use at approximately 118 PJ, significantly increasing the energy
consumption attributed to the agriculture sector (from 2% to over 6% of final energy).
Agriculture sector calibration approach
A bottom-up approach has been used to estimate energy use across the different agriculture
sub-sectors, based on stock data of agricultural machinery, and typical annual usage. The
basic approach is shown in Figure 20.
Stock Based on Machinery Census

(no. of tractors / pumps) (adjusted to 2006/07)
X
Adjusted to account for
Activity (hours/yr) unused machinery / downtime
due to disrepair
X
Consumption (l/hr or kWh) ENERCON / manufacturer’s
estimates
Lack of independent AGR

Energy use by technology HSD estimates makes
validation difficult
Useful energy delivered

Efficiency by technology
Figure 20: Approach to Bottom-up Estimation of Agriculture Sector Energy Use
The resulting estimates have been cross-checked against the Energy Balance values to
ensure base year calibration. This calibration process has been undertaken in conjunction
with the transportation sector, because the HSD used in agriculture is allocated to that sector
in the Energy Yearbook. Useful demand estimates are then developed based on typical
efficiencies and annual use of technologies.

The structure of the agriculture sector in Pak-IEM is shown in Figure 21, illustrating the use
of fuels by different device types for the provision of agriculture energy services.
Sector fuels Technology type Energy Service demand
Tractor haulage
Tractors (small) (Useful energy, PJ)
High Speed
Diesel (HSD) Tractors (large) Tractor farm
operations (Useful
Field Machinery energy, PJ)
HSD Tubewells
Tubewell / lift pump
water pumping
LDO Tubewells (Useful energy, PJ)
Light Diesel Oil
(LDO)
LDO Other
Electricity ELC Tubewells

(ELC) Other AGR energy
use (Useful energy,
ELC Other PJ)
Figure 21: Agriculture Sector RES Diagram37
Sector technology characterization

The structure for this sector requires the following types of data:
• Technology stock data – tractors / irrigation pumps / other self propelled field
machinery
• Stock activity levels / fuel consumption
• Characterization of existing end-use technologies
Technology stock data and utilization rates are presented in Table 23 and are sourced from
the 2004 Machinery Census (ACO 2005). Fuel consumption in tractors is taken from
ENERCON (1988), and has been validated based on personal communication with an
industry expert.38 Tubewell water pump consumption is based on Qureshi et al (2003), Bom
and van Steenbergen (1997), and TERI (2006). Again, this information has been validated
by industry experts.
37
Based on discussions with experts at NARC, the use of tractors for water pumping has not been included. It is
considered an old practice that is no longer common since the introduction of high-speed diesel pump sets.
38
Personal Communication with sector expert, dated 08.03.09 (See file Independent Expert TRACTORS 30-03-
09.pdf)

Table 23: Stock and Operating Assumptions for Agricultural Machinery

Hourly fuel
Stock
End-use technology type consumption Utilization
(000s)1
(l/hr)
Small-medium tractors (<55 HP) 384 4.7 12%
Large tractors (>55 HP) 83 5.7 12%
Combines 5 10 8%
Tubewells / Lift pumps (using HSD) 668 2.4 6%
Tubewells / Lift pumps (using Light Diesel
165 1.5 6%
Oil)
Tubewells / Lift pumps (using electricity) 97 11%
1
Adjusted to 2006/07 basis
Tractor estimates have been validated by industry sales figures, and stock estimates. These
have been used to make the adjustment of stock to a 2006/07 basis.39 NTRC (2008) has
also published “On the Road” estimates for tractors, which have not been used. They
estimate around 878,000 tractors, which is significantly higher than the other sources of
information used. NTRC are aware of these differences, and agree that Machinery Census
estimates are the best to use.
Base year stock of combines was provided by experts at one of the August 2009 stakeholder
meetings.40
Characterization of new technologies is based on the following sources:
• USAID41 project to install energy efficient tubewells in South Punjab, cross checked
against estimates in TERI (2006)
• Tractor investment costs provided by NARC42
Another important assumption for this sector is timing of demand for energy services that
use electricity. Electricity consumption in this sector is primarily for tubewell operations, the
timing of which is based on the overall seasonal pattern of consumption, as provided by
PEPCO, and expert judgment in regard to the daily profile. The shares allocated to each
season and day time slice are shown in Table 24.
Table 24: Season / Day Demand Shares for Agriculture Sector Electricity
Season Season share Day / Night Day share*
Summer (S) 0.47 Day (D) 0.10
Intermediate (I) 0.25 Shoulder (S) 0.19
Winter (W) 0.28 Peak (P) 0.0
Night (N) 0.71
* Day / Night shares are assumed to remain the same across all seasons
39
Personal Communication with sector expert, dated 08.03.09 (See file Independent Expert TRACTORS
30-03-09.pdf).
40
Data provided at Sector Task Force meeting in August 2009 (See notes from meeting in Agricultural
& Transport_subgroup_Summary_v03.doc).
41
USAID Energy Efficiency & Capacity Project, Islamabad, Pakistan.
42
See file labeled NARC Correspondance_Tractor&StoveCosts.doc.

These shares can be used to calculate demand fractions for every time slice period. The
summer night is the period when most electricity for pumping is required (35% of the annual
total), as this is the hottest season and the time of day when water tends to be pumped.
Summary of agriculture data sources
Table 25 provides a summary of the key data sources being used in the agriculture sector.
Table 25: Summary of Data Sources for the Agriculture Sector
Pakistan Machinery Census 2004 (ACO Tractor stock validated by
Stock data
2005) independent sector expert
Pakistan Machinery Census 2004 (ACO
Activity levels / utilization
2005)
ENERCON (1988) for tractors (validated by
NARC (Personal Communication) independent sector expert)
Pakistan Journal of Water Resources (2003) for irrigation pumps (validated by
Fuel consumption independent sector expert)
Energy for Sustainable Development (1997)
Applied Engineering in Agriculture (1993)
TERI (2006)
Efficiency ENERCON (1988)
Time slice demand PEPCO seasonal electricity consumption Expert judgement for day / night
shares data shares
Characterization of NARC; National Energy Map for India (TERI
future technology options 2006); USAID tubewell project
C. Industry Sector
Energy use in industry
Industrial manufacturing constitutes the second largest sector in Pakistan’s economy,
contributing around 18% to national GDP (GoP 2009). It is also the largest consumer of
commercial energy, and accounted for 35% of final energy consumed in 2006/07.
Natural gas accounts for almost half of all energy consumed, while coal consumption
accounts for approximately 25% (

Heavy Oil, 0.66 Lt Diesel,

0.05 Kerosene, 0.08
Diesel,
Electricity, 1.78
0.81
Bagasse,
1.11
Wood (Purch.),
1.77
Domestic Coal, 1.56
Natural Gas, 8.30
Imported Coal, 2.80
Figure 22). Cement and brick-making industries are the largest consumers of coal. The
consumption values below include energy used for captive generation, which is not identified
separately in the energy balance.

Heavy Oil, 0.66 Lt Diesel,

0.05 Kerosene, 0.08
Diesel,
Electricity, 1.78
0.81
Bagasse,
1.11
Wood (Purch.),
1.77
Domestic Coal, 1.56
Natural Gas, 8.30
Imported Coal, 2.80
Figure 22: Industrial Sector Energy Consumption (Mtoes), 2006/07 (HDIP 2008)
The share of natural gas consumption across selected industries is shown in Figure 23. The
textiles sector is by far the largest individual consumer, followed by iron and steel and
fertilizer industries.
Iron and
Steel
10%
Other
Industry
32%
Textiles
44%
Fertilizer
9%
Brick kilns
0% Cement
4% Sugar
1%

Figure 23: Industrial Sector Natural Gas Consumption (%), 2006/07

(Source: SNGPL / SSGC)43
Identification of the main industrial energy consumers provides a basis for deciding which
sub-sectors to model individually, and thereby with process-step level detail. On this basis,
the following industry sub-sectors were agreed for inclusion in Pak-IEM:
• Textiles
• Sugar
• Iron and Steel
• Cement
• Fertilizer
• Brick-making44
Other factors in determining the above sub-sector categories include availability of:
• Output statistics to relate energy use and production
• Sectoral future growth information
• Data to understand industrial processes within each sub-sector
All remaining sub-sectors are included in an “Other Industry” category, which accounts for
the largest share of final energy consumption, at 34%. A review was undertaken of the paper
and pulp, and glass industries but based on their aggregate energy consumption, it was
decided not to model them as separate sub-sectors.
Approach to industry sector modeling
Figure 24 illustrates the basic approach to industrial sector modeling. Overall industrial
energy consumption can be determined from the energy balance, while sub-sector
consumption is estimated based on a range of sources, including data from the gas and
electricity distribution companies (Table 26).
43
Data from SSGC and SNGPL was provided directly for use in the project. The relevant files, listed
in Appendix 3, are Categ-det_Jul'06-Jun'07.xls (SSGC) and SNGPL Consumer Data 2007 and
2008.xls (SNGPL).
44
Brick-making is modelled very simply, and as a separate sub-sector to track coal consumption.

Energyuse
Energy industry use
in Industry ENERGY BALANCE
DISCO electricity data
Industry subsector splits Gas billing data

ENERCON MIS data
ENERCON audit data

Generic energy Process based
services modelling Industry-specific data
modelling
Energy services linked Production statistics

to production
Figure 24: General Approach to Industry Sector Modeling

Table 26: Sources of Data for Estimating Sub-sector Energy Consumption

Fuel type Source / comments
Electricity (public PEPCO administered DISCOs / KESCO. Billing data available by sub-sector. Also
supply) provides seasonal load profile
Electricity (self- New survey: Sectors / equipment manufacturers surveyed by MEC, to assess installed
generation) capacity in-country
UDI: Estimates based on capacity (UDI database), typical use patterns and efficiencies.
Year of installation also from UDI
Census of Electricity Establishments (CEE) 2005-06
Natural gas Gas distribution companies. Billing data from SSGC and SGNPL
Heavy oil ENERCON MIS. Based on plant-specific data from ENERCON database
HSD ENERCON MIS. Based on plant-specific data from ENERCON database
Coal Pakistan Energy Yearbook 07. Disaggregated in the Energy Balance
Bagasse Pakistan Sugar Mills Association, ENERCON MIS
Once sub-sector energy consumption has been estimated, there is a need to know more
specifically how energy is used to provide services and in different industrial process steps.
There are two approaches that can be taken.
• Generic energy service modeling – Energy use is allocated to broad energy services
(e.g., motive power and process heat) that are required in the manufacture of
products.
• Process based modeling – A more detailed approach that requires good process-
specific data. Individual technologies within the manufacturing process are
characterized in terms of their efficiencies (unit output / unit inputs) and operating
costs. For example, instead of generic technologies providing process heat in the
iron and steel sector, specific process steps involving different technologies (e.g.,
blast furnace, re-heating, induction) are individually specified.
The Pak-IEM model uses a combination of the approaches, depending on the sub-sector
and data availability. Finally, the energy services are linked to production; for every unit of
production (e.g. tonne of sugar), “x” petajoules of energy services (e.g. process heat) will be
required.
Approach to sub-sector modeling – Iron and Steel
The iron and steel sector in Pak-IEM is split into two sub-sectors – private sector
independent producers who produce steel through melting and steel products in rolling mills,
and the part public-owned large integrated steel works near Karachi, Pakistan Steel.
In the independent sector, steel melting is carried out using induction furnaces (approx. 140
plants) to produce ingots / billets for production of steel-based products. The main
technology to produce required process heat is induction furnaces (with limited use of
electric arc furnaces in Pakistan) using electricity. There is limited self-generation in this

sector.45 Another important independent sector is steel re-rolling. The main technology to
produce required process heat is a re-heating furnace using fuel oil / natural gas.
Two key sources have been used to characterize the sector in the base year. The first is a
presentation by Magna Steel whilst the second is data provided directly by Pakistan Steel
Rolling Mills Association (PSRMA).46
Table 27: Key Data Assumptions for Steel Melting and Re-rolling Sectors
Sector Steel Melting Steel Re-rolling
Registered Units 140 200
Total Units 350-400 300-500
Base Year Production (Mt) 2.4 4
Electricity consumption (kWh)/ tonne
output 800 130
Gas consumption (m3 gas) / tonne output 100
The structure of the independent steel sector in the model is shown below in Figure 25.
There are two key technologies in the sub-sector – induction and re-heating furnaces. A third
technology generically represents “other” energy consumption in unregistered units. Captive
generation levels are considered to be extremely low in this sector, and therefore not
modeled.
Fuels Processes Demand
Induction
Electricity Furnace Billets (Mt)
(Melting)
N Gas Re‐heating Construction steel

Furnace (Mt)
Other Energy Useful energy for

Use unregistered units
(PJ)
Figure 25: Reference Energy System – Iron and Steel (Independent Operators)
New technology options include new plant or energy efficiency opportunities across existing
units. Energy efficiency potential is taken from recent analysis by Hagler Bailly (ADB 2009b).
It is modeled through the use of a retrofit measure to ensure that only existing stock is
included.47
45
Pakistan Steel Melters Association website – http://steelmelters.com/
46
Data sourced from files Steel Sector Presentation Final on 10-8-09.ppt (from Magna Steel) and
Pakistan Steel re-Rolling Mills Association.pdf (see Appendix 3).
47
Retrofit measures require the use of sophisticated user constraints, which can be found in model
scenario file Scen_REF-DMD.xls.

The other sub-sector is the large integrated works Pakistan Steel Mill. It produces a range of
iron and steel products through an integrated process, as illustrated in the RES diagram
below.48 Some aspects of the process have been simplified; for example, cold and hot rolling
of products, and galvanizing are grouped into “Other” processes.
As discussed, this detailed process-based modeling requires detailed data, at the minimum
production and energy consumption at each of the process stages, with associated
operating costs. These data have been provided by Pakistan Steel,49 and provide the basis
for detailed modeling (see Figure 26 below).
Fuels / Materials Processes Demand
Coal
Coke Ovens COG
COKE
Sintering IRON
plant ORE
NGA
COKE SINT.
ORE Pig Iron (Mt)
Blast BFG
Furnace
ELC
HOT
METAL Cast billets
Steel (Mt)
production
COG RAW
Cogen Steel products
Plant STEEL
BFG (Mt)
Re‐heating
+
Other Other process

processes useful energy
LTH (PJ)
Figure 26: Reference Energy System – Iron and Steel (Pakistan Steel)
While the base year characterization is known, understanding what technical options can be
considered in future years is more difficult, as such options should be site-specific. As Pak-
IEM is further developed, it will be important to revisit this issue with the company, to better
understand energy efficiency potential. Default new technology options have been included
in the model for blast furnace (ETSAP technology factsheet50) and basic oxygen furnaces
(TERI 2006), with other process steps costed relative to these core processes. However, in
reality, it is likely that the plant will retrofit rather than re-built over time; therefore, further
discussion with the industry is important.
Specific data for the above plant have not been published here to ensure confidentiality for
this operator.
48
Details of the integrated plant processes can be found on the Pakistan Steel website –
www.paksteel.com.pk/
49
Production and fuel consumption data provided in file Pak Steel Data 12-2-2010.pdf
50
This can be found at ETSAP website - http://www.etsap.org/index.asp

Approach to sub-sector modeling – Textiles

The textile sector is one of the largest energy-using sectors in Pakistan, accounting for over
20% of industry final energy consumption. It is also a sector of significant economic
importance, with Pakistan the fourth largest producer of cotton. In 2006-07, it accounted for
over 60% of total exports ($6.6 billion), 46% of manufacturing output, and 38% of
employment in the economy (GoP 2008).
The sub-sector can be divided into spinning, weaving, and finishing. Dyeing and printing,
which are large energy users within finishing, could also be separated out. However, this
would be heavily reliant on data availability. Spinning and weaving are large users of
electricity primarily for motive power as well as for heating and ventilation.
There is a large non-mill textile sector in Pakistan; it has been decided that both this sector
and the mill sector will be included in Pak-IEM because their respective energy use cannot
be distinguished. No distinction has been made between mills undertaking specific
processes and integrated mills.
The structure of the current sub-sector, shown in Figure 27, is generic, with general sub-
sector energy services modeled rather than specific processes. Most of the assumptions
concerning how much energy is used for production of yarn versus cloth, and at what
efficiency, is based on expert judgment primarily from MEC. Detailed data was not
sufficiently sourced during the project to allow for more detailed modeling, and this remains a
priority for future development.
Electricity
Motive power
(spinning) Mill Yarn Production
(Mt)
Electricity Space heat /
Ventilation
Process heat
(Weaving)
Mill Cloth Production
Motive power (Metres Sq.)
(Weaving)
HFO
N Gas
Process heat Useful energy for

Electricity
Finishing (PJ)
Other electric
Figure 27: Reference Energy System – Textiles

Estimates of captive generation capacity are sourced from a survey of distributors under this
study (for gas) and from the Platts UDI database (for oil-based generation). These sources
are described in more detail in the later section. The capacity for the textile sector is
estimated to be significant, at about 1.8 GW, the majority being gas.
Retrofit options are available from the recent Hagler Bailly energy efficiency analysis for ADB
(2009b). These options retrofit existing stock, and allow for modeling of energy efficiency
improvement. New technology options are less well characterized but take account of the
relative cost of retrofit options.

Approach to sub-sector modeling – Cement

Pakistan currently has 29 cement plants with a combined production capacity of 37 million
tons of cement, mainly Portland cement. In 2006/07, 22.8 million tons were produced. Due to
their excess capacity, cement companies are exporting to Middle East, Africa, Afghanistan,
and India. In the first nine months of FY08, exports increased by 114% to 5.1 million tons
(expected to be 6.6 million tons for the full year) (GoP 2008).
Energy consumption in the cement sector is dominated by coal use in rotary kilns for clinker
production. Most plants use a dry rather than wet process, a distinction which is important for
understanding energy needs per unit of production.
The structure of the sector in Pak-IEM is relatively simple, and would probably be classed as
generic energy services modeling, with no specific characterization of the different
production processes – raw material preparation, clinker production, and cement grinding.
Gas
Self‐
Oil generation
Electricity (Captive)
Electricity
Cement Kiln ‐ Dry
Single Stage
Cement Kiln ‐ Dry

Multi‐Stage Cement (Mt)
Coal
Cement Kiln ‐ Wet
Figure 28: Reference Energy System – Cement
As shown in the above RES diagram, there are four main technologies; sector-based
(captive) generation and three types of cement production – wet, dry single stage and dry
multi-stage. Coal is the primary fuel used in the cement kilns; electricity constitutes a much
smaller proportion of energy consumption in this sector. The production levels and coal
consumption shown in Table 28 are from ADB (2009b), and provide the basis for splitting
production and fuel consumption in the base year.
Table 28: Cement sector production and coal consumption, Million tonnes
Production (Mt of Coal use per t of
Kiln Type Coal Use (Mt)
cement) cement (t)
Wet 4.08 1.26 0.309
Dry Single 5.18 1.37 0.264
Dry Multi 16.61 3.08 0.185
There are three new technology options in the cement sector.

• Two retrofit options, upgrade of wet process cement plants and dry single process
cement plants to dry multi-stage. Data are sourced from the recent Hagler Bailly
energy efficiency analysis for ADB (2009b).
• New build option, to install dry multi-stage plant, based on data from TERI (2006).
A significant amount of the electricity used in the cement sector is generated from captive
generation. It has been estimated at around 300 MW based on fuel consumption (close to
the aggregate capacity level from the UDI database of 270 MW). Some uncertainty
surrounds the estimate of gas plant capacity, which is much higher than the MEC or UDI
estimate). This is driven by the natural gas value for the cement sector from the Energy Year
Book which is allocated entirely to electricity generation in the sector.
Three new technology options exist for captive generation. These include:
• Gas generation plant
• Oil generation plant
• Generation plant using waste heat recovered from cement kilns, and providing the
opportunity for sale of electricity to the grid. [The data for the characterization of this
plant is from a NEPRA determination document, for a particular plant considering this
option.]
For captive generation new technology assumptions, see the relevant section below.
Approach to sub-sector modeling – Sugar

Sugar production energy requirements are largely met by the use of bagasse in
cogeneration plants to produce electricity and steam. Steam drives much of the plant
machinery, including some of the motive power requirements. The production is seasonal,
occurring only during the summer cane harvesting period. The majority of the energy needs
of this sub-sector are provided by the bagasse, which is the residue of the sugar cane after it
has been crushed and its sugar extracted.
Modeling is based on the generic energy services approach, with process heat and
electricity requirements coming exclusively from the cogeneration plant. During the off-
season, other energy is needed for maintenance so most plants remain grid-connected.
The model has also been set up to allow for electricity supply to the grid. It also has a
distilling process included, for the production of molasses, some of which is used for the
production of ethanol, including fuel grade ethanol that could be used in the transport sector.
In the base year, all of this fuel grade ethanol is exported although could potentially be used
in the domestic transport sector. The structure of the different processes in the model is
shown in Figure 29.

Co‐generation
Bagasse
To grid
LP steam
Motive power
Electricity (electric) / Other
electric
Sugar (Mt)
Process heat
(Motive power
steam)
Natural gas
Other processes
HF Oil incl. distilling
Molasses
Ethanol Ethanol
production
Figure 29: Reference Energy System – Sugar

Key assumptions and sources used for the base year characterization are provided in Table
29.
Table 29: Key assumptions used for sugar sector in Pak-IEM model
Data type Million tonnes Source
Sugar cane (SC)
production 54.9 PSMA (2007)
Supply of SC to mills 40.5 PSMA (2007)
ENERCON MIS
Bagasse for energy 8.6* database
Sugar production 3.53 FBS (in GoP 2008)
Molasses production 1.91 PSMA (2007)
Ethanol production Dewan mushtaq
capacity 0.4 Group**
Ethanol export 0.274 PSMA (2009)
USDA Foreign
Proportion of exported Agricultural Service
ethanol of fuel grade 13% (2007)
* Cross checked against typical cane to bagasse ratios of 25% e.g. WADE (2004).
** Presentation on ‘Overview of Sugar Manufacturing Sector of Pakistan’ By Mian Kausar Hameed, COO -
Sugar Operations, Dewan Mushtaq Group
The key new technology for the sugar sector is the retrofitting of existing cogeneration
plants, to increase steam pressure and increase potential electricity generation for sale to
the grid. Total additional power that existing plants can produce has been estimated by MEC
at 655 MW. Costs and characteristics of new cogeneration plants are based on a NEPRA
tariff determination document for a specific sugar mill looking at the sale of surplus electricity
to the grid.

Approach to sub-sector modeling – Brick production

The brick production sector is relatively informal compared to other sectors, with numerous
kilns around the country, proving bricks for local construction activities. The primary fuel
used in these kilns is coal, the total of which is estimated in the Energy Year Book.
Data on technologies is sourced from the recent Hagler Bailly energy efficiency analysis for
ADB (2009b). The majority of existing brick production uses Bull’s Trench kilns, with a small
percentage (5%) of Clamp kilns. There is significant scope for efficiency improvement, as
illustrated below in Table 30.
Table 30: Brick Kiln Efficiencies (Source: ADB 2009b)
Kiln Type kg/brick Efficiency*
Vertical 0.1 1.47
Bull’s Trench 0.19 1.00
Clamp 0.23 0.79
* Normalised against Bull’s Trench kiln efficiency
New technology options are either new vertical kilns, or Bull’s Trench and Clamp kiln being
retrofitted with a vertical kiln.
Approach to sub-sector modeling – Fertilizer
The fertilizer sector is important to model as a separate subsector due to the high level of
gas consumption, both as a fuel and a feedstock. Forty-six percent of industry gas
consumption is used by the fertilizer industry, nearly 80% for feedstock. Gas for feedstock
and fuel are separately modeled, with feedstock gas not being processed like gas used as a
fuel.
This separation is also important to reflect different pricing of gas. The Government provides
an indirect subsidy to fertilizer manufacturers by selling feedstock gas (80% of the raw
material cost) at approximately 50% lower rates compared to the price for commercial users
(GoP 2008). Data on differentiated sector prices is primarily sourced from the Energy Year
Book (HDIP 2008).
The model also has a fertilizer import option, which comes into play when the price of gas
rises significantly. The price of the imports is based on assumptions provided by the Pak-
IEM Planning Team.
Timing of electricity demand by industry sub-sector
Based on PEPCO data, the demand for electricity by season is relatively flat, with limited
seasonal variation. All sectors have a 24 hour shift pattern with the exception of sugar and
the general Other industry category, which have 16-hour day shifts.
Captive power generation capacity in industry
An important aspect of the industry sector modeling is the captive power generation, which
provides electricity for self-consumption across different industrial sub-sectors. Results of a
survey of manufacturers / distributors of equipment undertaken by the project team suggest
captive generation capacity of approximately 2.5 GW, almost 1.8 GW of which is in the
textile sector (see Table 32 below).
The survey only covers gas fuel equipment as diesel and Heavy Fuel Oil (HFO) equipment
installations are few, and run intermittently (as back-up) due to high fuel costs. Estimates

from the UDI WEPP database51 published by Platts are therefore used for the oil-based
plants. Gas engines have a high investment costs; industry therefore strives for high
utilization to ensure return on investment. The feasibility of installation only works if they
perform base load duties; otherwise they are unlikely to be purchased.
The survey estimates far exceeds estimates from the previous source of information, the
UDI WEPP database. The WEPP database puts total captive generation in Pakistan industry
at 1.2 GW (250+ units), including the capacity of 0.7 GW for the sub-sectors shown in Table
31.
Table 31: Industrial Captive Generation Plants (Source: UDI WEPP, Platts)
Sector Input fuel Unit type* Capacity (MW) Unit number
Sugar Mill BAG ST 133.3 49
ST/S 10.4 4
OIL IC 1.4 2
UNK ST 2 1
ST/S 1.5 1
Sugar Subtotal 148.6 57
Cement GAS IC 49.5 3
OIL IC 226.0 38
Cement Subtotal 275.5 41
Pakistan Steel** COG ST/S 165 3
Textiles GAS GT/S 7.4 2
IC 38.1 15
IC/H 3.0 3
OIL IC 72.2 35
IC/H 15.5 16
ST/S 7.2 1
Textile Subtotal 143.4 72
Sector Total 732.5 173
* Unit type definition: ST-Steam turbine; ST/S-Steam turbine with steam sendout (cogen); GT/S-Gas turbine
with steam sendout (cogen); IC-Internal combustion (reciprocating engine or diesel engine); IC/H-Internal
combustion engine with heat recovery (cogen)
** Integrated works only; no independent sector operators included
51
UDI World Electric Power Plants Database (WEPP), Published by Platts (www.platts.com/)

Table 32: Captive generation capacity by Industry Sub-sector

(Results of MEC Ltd Survey)
INDUSTRIAL SEGMENTS AND INSTALLED POWER IN MW
NO MANUFACTURER/ TEXTILE PETRO‐ COMM‐ HOSPITAL & STEEL & CHEM‐ FOOD WOOD, CEMENT & GLASS IPP SUGAR, ROW
DISTRIBUTOR LEUM ERCIAL PHARMA ENGIN. ICAL PAPER& CERAMICS FERT. RICE, TOTAL
BOARD OTHER (MW)
1 GE JENBACHER 683 4 23 11 38 31 24 11 24 5 216 18 1087
2 CATERPILLER 741 31 47 6 11 1 2 0 25 8 16 9 896
3 WAUKESHA 228 105 2 1 14 2 9 3 NIL 11 NIL 4 378
4 DUETZ MWM
5 GAUSCOR 61 0 2 3 4 3 17 2 0 0 0 0 92
6 KAWASAWI GAS 3 3 2 0 28 0 0 3 0 0 0 0 38
TURBINE
7 TURBOMAC GAS 77 0 0 2 0 11 0 0 0 0 0 8 98
TURBINE
A TOTAL CAPACITY 1793 142 75 24 94 48 51 18 49 24 232 39 2497

(MW)
B DAILY OPERATING 24 24 12‐14 14‐16 18‐20 24 16 24 24 24 24 24

HOURS
UTILIZATION
C 0.87 0.87 0.65 0.70 0.75 0.80 0.75 0.70 0.75 0.75 0.87 0.75
FACTOR

Additional information on captive generation has been sourced from the 2005-06 Census of
Electricity Establishments (FBS 2007) and used for cross-checking model estimates.
Summary of industry sector data sources
Table 33 provides a summary of the current industry sector data sources.
Table 33: Summary of Data Sources for the Industry Sector
Sub-sector energy
consumption
Note that this is used to derive shares
Electricity consumption PEPCO DISCO billing data
between sub-sectors
Sui Northern Consumption data See file SNGPL Consumer Data 2007
Gas consumption and 2008.xls
Sui Southern Consumption data See file Categ-det_Jul'06-Jun'07.xls
Other energy consumption ENERCON MIS database
Cement sector totals sourced directly
Cement sector EYB 2007 (HDIP 2008) from EYB. Coal information provided
directly by HDIP.
Captive generation
Project-based survey Published by Platts. An industry-
Capacity / technology type complimented by UDI WEPP based review will compliment this.
database
Capacity by sector Based on project team survey
Iron and Steel sector
Production and sales data for Independent sector -see Steel Sector
independent sector in Magna Presentation Final on 10-8-09.ppt;
Production data
steel presentation. Pakistan Pakistan steel – see Pak Steel Data
Steel data provided directly. 12-2-2010
Energy consumption (process As per above
As per above
level)
Technology characteristics As per above As per above
Textile sector
All Pakistan Textile Mills http://www.aptma.org.pk/Statistics.asp
Production data
Association (APTMA) website
Energy consumption (process ENERCON Energy Audit
level) dataset
ENERCON Energy Audit
Technology characteristics
dataset
Sugar sector
Pakistan Economic Survey PSMA providing data on cane
Production data
2007/08; PSMA production and supply to mills
Energy consumption (process ENERCON Energy Audit
level) dataset
ENERCON Energy Audit
Technology characteristics
dataset
Cement sector
Pakistan Economic Survey
Production data
2007/08
Energy consumption (process Energy Year Book 2007, ADB
level) 2009b
Technology characteristics ADB (2009b)

Engagement with trade association will be important for maintaining datasets in future years.
Key associations are listed below in Table 34.
Table 34: List of Sector Associations

Sector Web address
Pakistan Steel http://www.paksteel.com.pk/
Pakistan Steel Melters Association http://steelmelters.com
Pakistan Steel Re-Rolling Mills
Association
Pakistan Sugar Mills Association http://www.psmaonline.com/psma/home.aspx
All Pakistan Cement Manufacturers http://www.apcma.com/
Association
All Pakistan Textile Mills Association http://www.aptma.org.pk/
D. Residential Sector
The residential sector, like industry and transport, is one of the largest consumers of energy,
accounting for 20% of total energy consumed. The published reports and data available to
characterize the residential sector are either quite old, inadequate, or do not provide a
consistent set of information that can be use to characterize these sectors and understand
how energy is currently used.
Information describing the general energy supply situation is available from the Energy
Yearbook, and distribution companies. However, important details are lacking when it
comes to understanding time-of-use consumption patterns and the relative shares of energy
utilized by end-use appliances, such as air conditioners, refrigerators, water heaters, lighting,
etc. The last major surveying activity of the residential sector, the World Bank sponsored
Household Energy Strategy Study was undertaken in the early 1990s (ESMAP 2003).
However, new research as part of the modeling analysis has enabled a reasonable
characterization of energy use within this sector. This includes development of biomass
estimates to ensure this important energy type for rural communities is captured, and
survey-based activities to better understand the technologies on the market, and the stock in
use.
Approach to residential sector modeling

The types of information needed for a proper characterization of the residential sector are
illustrated in Figure 30. They include data on regional energy use patterns, demographic
information to determine how the current energy use may change in the future, and a
breakdown of the energy use patterns by end-use application. Energy demand should be
defined according to useful energy requirements rather than final energy use, so that the
impacts of fuel switching and end-use device efficiency can be factored into analyses of the
sector.

Residential Sector
Characterization
Regional Demographic Energy Services

Characteristics Characteristics
• End-use applications
• Province • Income
• Energy use and timing
• Urban vs. Rural • Age
• Device stocks and
• Season • Migration trends characteristics
• Time of Day • Energy Intensity trends • New device options
Figure 30: Information Needed to Characterize Residential Sector Energy Use

Due to a lack of up-to-date information for these sectors in Pakistan, a large surveying
program has been proposed by the project team to be funded by USAID as part of their
Energy Efficiency project. However, plans for this survey are currently on hold as the USAID
project focused on activities with greater near-term impact.
In the absence of new survey data, development of this first version of Pak-IEM relied on a
variety of existing and recent data sources, using these to develop a decomposition
approach to residential sector modeling.
The approach for this sector is illustrated in Figure 31. It takes the following steps:
• Determining the number of urban and rural households
• Estimating the number of households in urban / rural areas using specific fuels
• Weighting the above household numbers to reflect differences in energy intensity
• Estimating shares to disaggregate the urban / rural energy consumption by energy
service, based on a range of sources

No. of rural / urban households FBS
% of urban / rural households Pakistan Household

using a specific fuel Expenditure Survey (cited in
ESMAP 2006)
Rural / urban household

energy intensity by energy HESS (to weight urban/rural
service type household consumption)
HESS (Biomass)
Shares for rural / urban energy ADB / HB (Electricity)
use by service type SGCC / SNGPL (Gas)
Experts (LPG / Kerosene)
PJ use for the following energy services:
Space Heating Cooking

Water Heating Refrigeration
Space Cooling (Fans) Miscellaneous Electric
Space Cooling (AC) Other
Lighting
Figure 31: Approach to modeling the residential sector

In future versions of Pak-IEM, there is an intent to expand this approach to disaggregate
rural and urban classes into income classes, if possible. The rural-urban distinction does
reflect some of the differences in the type of consumption but not to the extent that an
income group disaggregation would.
The differences in base year consumption levels between rural and urban areas as a
percentage of overall energy consumption are shown in Figure 32 and Figure 33. The large
amount of biomass use in the rural areas is particularly evident as is the current dominant
energy for cooking services

60%
50%
Other
40% Miscellaneous
Electric
% of urban final energy
Refrigeration
30%
Cooking
20%
Lighting
10% Space Cooling (AC &

Coolers)
Space Cooling (Fans)
0%
Water Heating
Space Heating
Figure 32: Shares of final energy consumption by type in the urban residential sector
70%
60%
Other
50%
Miscellaneous
Electric
% of rural final energy
40% Refrigeration
30% Cooking
20% Lighting
10% Space Cooling (AC &

Coolers)
Space Cooling (Fans)
0%
Water Heating
Space Heating
Figure 33: Shares of final energy consumption by type in the rural residential sector
Technologies used in the residential sector

The approach described so far allocates fuel consumption to different types of uses (cooling,
heating, etc). The next step in constructing this sector was first, to understand the existing

stock in the sector of appliances, and second, the new technologies that might be available
for investment in later years in the model.
Significant assistance was provided by the Planning Team, in particular the Mechanical
Engineering Department at the University of Engineering and Technology – Taxila (UETT),
to help develop this dataset, particularly in the absence of other data. Much of the
information to characterize the appliances was done based on UETT Taxila survey work, as
shown in Table 35 below. Due to the surveys being focused on the Punjab region, a future
priority will be to expand these survey activities across Pakistan.
Table 35: Data sources to characterize residential sector technologies
Data Parameter Data Source
Residential Energy Survey UETT Taxila 2009 & 2010

Appliance price www.metro.com.pk
www.prices.pk
Avg. Power Rating (W) – for Residential Energy Survey UETT Taxila 2009 & 2010
electric appliances
www.metro.com.pk

Avg. Life(yrs.)
Interviews and Local Market Survey 2010

Avg. Use/day
http://www.gallup.com.pk/
Interviews and Local Market Survey 2010

Annual maintenance cost
Estimating Residential Electricity Demand Responses in

Pakistan’s Punjab(Theresa Chaudhry), The Lahore Journal
of Economics, 15: Special Edition: Papers and Proceedings
of the Sixth
Stock of appliances by type Residential Energy Survey UETT Taxila 2009 & 2010
OWNERSHIP OF DURABLES IN PAKISTANI HOUSEHOLD

(results released on: July 30, 1999)
http://www.gallup.com.pk/
Data on lighting technologies, including existing stock, their costs and potential, is sourced
from ADB (2009b). For new technologies, the US NEMS model has been used to provide
information on more advanced technologies.
The type of technologies considered are shown in the RES diagram below, illustrating the
type of energy used (on the left) and the energy services provided (on the right).

Sector fuels Technology Type Energy Service Demand
Lighting
Electricity •Tube Lights Space Illumination
•Incandescent Lamps
•Fluorescent Lamps
•Energy Savers Space Heating
Natural Gas Space Cooling
•Kerosene Lamps
•Candles
LPG
Climate Controls
•Heaters Water Heating
•Fans
Kerosene •Desert Coolers
•Air Conditioners
•Geysers
Dung Cake •Boilers
Agricultural Residues Appliances Cooking

and Firewood •Refrigerator Refrigeration
•Stove Other
•Generators
Diesel/Petrol •Other
Figure 34: Residential sector RES Diagram
Temporal patterns of demand in the residential sector

In Pak-IEM, the demand for energy services is tracked at a sub-annual level, reflecting that
heating is required most during winter periods and air conditioning during the summer. For
those commodities that are characterized sub-annually (gas seasonally and electricity
across periods of the day), their supply has to match the temporal patterns of demand.
Demand fractions (COM_FRs in the model) have been calculated, primarily with electricity in
mind, which allocate annual demand for an energy service across the different timeslices in
the model. This has been done by calibrating the temporal profile of demand with the
electricity load curve, using hourly load data from PEPCO and KESC (as described in
section II.B).
Therefore, at the aggregate level, it is clear how much electricity is demanded in a given
timeslice for all sectors. In addition, due to seasonal demand data, the level of demand in the
residential sector by month is also known. The most difficult information to source is what the
difference in demand is for a specific energy service within a specific sector. This information
has primarily been sourced from a survey52 by UETT and from expert judgment from
stakeholders and the wider Planning Team.
Figure 35 illustrates the survey information on energy service demand by season (as defined
in the model). Using these data and other information, the resulting timing of demand profiles
for space cooling and space heating can be seen in Figure 36. These profiles also apply to
gas supply, which is seasonally-defined commodity.
52
UETT Taxila conducted an energy survey for the residential sector in Punjab, consisting of about 35 questions
regarding life style, income group, and electricity consumption by different categories of end-use devices; data
from utility bills was also collected. The sample size was 1,300 households, of which 41% were rural and 59%
were urban. Income group information was also collected but not used here.

2.5
Refridgeration
Misc. Electric
2.0 Lighting
Space heating
Seasonal fraction Space cooling

1.5
1.0
0.5
0.0
W I S
* Above shares based on UETT survey are fixed at the aggregate level to seasonal data provided by PEPCO.
This figures shows the higher demand in summer (S) relative to other seasons, and the relative contributions by
energy service demand
Figure 35: Season fractions for electricity demand in residential sector across
different end-use applications
0.40
Heating
Fraction of annual energy service demand in a specific
0.35
Space Cooling
0.30
0.25
timeslice
0.20
0.15
0.10
0.05
0.00
SD SS SP SN WD WS WP WN ID IS IP IN
Timeslice period
Figure 36: Timing of Demand Profiles in Residential Sector by space heating and
cooling

Residential sector data sources

A range of data sources have been used for this sector, and are listed in Table 36 below
Table 36: Summary of Data Sources for the Residential Sector
Household expenditure survey
Household use of fuels
2001 (as cited in ESMAP 2006)
Relative energy intensity of urban
HESS (ESMAP 1993)
/ rural households
HESS (ESMAP 1993)
Fuel shares by energy service SGCC / SNGPL Seasonal profile used to identify gas
type demand for different energy services
ADB (2009b) Electricity shares
See Table 35
Technology data Also US NEMS data, ADB
2009b
E. Commercial Sector
The commercial / Other Government (including Military) sector takes a similar approach to
that employed for the residential sector, but as a single aggregate sector (no urban / rural
split). The shares of energy by end-use demand category are currently formed on the basis
of expert judgment, as are the timing of demand profiles.
45%
40%
Share of sector final energy consumption
35%
Other
30%
Miscellaneous Electric
25%
Refrigeration
20%
Lighting
15% Cooking
10% Space Cooling
5% Water Heating
0% Space Heating
Figure 37: Shares of final energy consumption by type in the commercial sector
Energy use is dominated by gas and electricity. Transport fuels are considered to be used in
Government vehicles, although these have been allocated to “Other.”
Due to the lack of in-country data, information on different technologies is sourced from the
US NEMS model.

The current lack of data means that many of the assumptions in this sector will continue to
be based on expert judgment – and potentially some very limited surveying activity. Although
not ideal, the concerns over the strength of the data are less because this sector is actually a
relatively small user of energy (6% of final energy consumption as per the Energy Year Book
2007). However, as there is a real potential for energy saving, and this is a growing sector,
better characterization of the sector is desirable. Further discussion of areas for
development can be found in section VIII.
F. Transportation Sector
The transportation sector includes various forms of transport for passengers and freight in
Pakistan. Table 37 provides the sub-sector splits in the model between road and non-road
transport vehicles, many of which provide both passenger and freight transport.
Table 37: Transportation Sub-sectors in Pak-IEM
Sub-sector Vehicle type / Mode
1
Road transport
Passenger Cars
Taxis
Buses
Mini-buses
Three wheelers
Two wheelers
Freight Trucks
Vans
Non-road transport
Passenger Rail electric
Rail diesel
Freight Rail diesel
Non-differentiated Aviation
Shipping
1
Road transport sub-sectors determined on the basis of vehicle stock statistics
Energy use in transport

Transport energy use in the base year (2006-07) is sourced from the Energy Year
Book (HDIP 2008).
Kerosene, 0.001
Heavy Oil, 0.009

Natural Gas,
Av. Fuel,
Electricity, 1.32
0.53
0.001
Gasoline, 1.20
Diesel, 4.70
Figure 38 illustrates the split of total energy used in the transport sector, which totals 348 PJ.
The sector is dominated by high speed diesel consumption. The figure excludes the HSD
that has been allocated to the agricultural sector, which the Energy Year Book includes in
the transport energy consumption totals.

Kerosene, 0.001
Heavy Oil, 0.009

Natural Gas,
Av. Fuel,
Electricity, 1.32
0.53
0.001
Gasoline, 1.20
Diesel, 4.70
Figure 38: Fuel Consumption (PJ) in the Transport Sector, 2006/07 (HDIP 2008)
Table 38 shows the HSD balance from the Energy Yearbook and how it has been allocated
in the model. The bottom-up estimates (in Pak-IEM) for agriculture combined with transport
suggest a shortfall of 35 PJ, which we assume to be diesel imported on the black market.
The assumptions used in our bottom-up estimates require further scrutiny by experts to
ensure that this surplus relative to the Energy Yearbook is representative. Discussions to
date suggest that there is a significant amount of cross-border smuggling; therefore a 10%
surplus may be a reasonable estimate.
Table 38: HSD Energy Balance (PJ) for 2006/07, and Allocation in Pak-IEM
Energy Yearbook 2007
Production Refining output 150
Import 185
Export -2
Bunkers -0.2
Total available to system 333
Consumption Industry 36
Transport 294
Agriculture 0.5
Other 3
Total consumed 333
Pak-IEM consumption
estimates
Agriculture 118
Transport 208
All other sectors 42

Energy Yearbook 2007

Deficit (assumed to be Black
35
Market Diesel, 11% of EYB total)
Transport sector modeling approach

Although the high level energy use in the transport sector is known from the Pakistan Energy
Year Book, understanding energy use by road vehicle type and mode requires bottom-up
estimation using stock information, and various associated technical parameters – notably
efficiency and activity (e.g. annual vehicle km). The basic approach is shown in Figure 39.
The resulting estimates are cross-checked against the energy balance values to ensure
base year calibration. Useful demand estimates are then developed based on vehicle
occupancy (passengers) or load (freight).
‘On road’ estimates compared

Stock (no. of vehicles) against registrations
X
Check against aggregate vkm
Activity (vkm/yr) estimates
/
Efficiency (MJ/vkm)
Energy use by vehicle Ensure consistent with EYB

type estimates
Useful demand by Vehicle occupancy /

vehicle type freight load
Figure 39: Approach to Bottom-up Estimation of Energy Use in Road Transport Sector
Based on the above approach, Figure 40 and Figure 41 show the Reference Energy System
structure of the passenger and freight transport sub-sectors, respectively. These RES
diagrams show how the sector fuels can be used by different vehicle types for the provision
of the various transport sector energy service demands.

Sector fuels Vehicle type Technology type Energy Service demand
2‐wheelers Petrol ICE 2‐wheeler pkm

Petrol
3‐wheelers Petrol ICE 3‐wheeler pkm
Petrol ICE
CNG Taxi pkm
Taxis Dual‐fuel
Cars Petrol ICE
Dual‐fuel Car pkm
Diesel ICE
Diesel Minibuses Diesel ICE Minibus pkm
Buses Diesel ICE Bus pkm
Electricity Rail Diesel

Rail pass. Rail pkm
Rail Elc
Aviation fuel
Air pass. Air Av. fuel Air pkm
Figure 40: Pak-IEM Transport Sector RES Diagram – Passenger Vehicles
Sector fuels Vehicle type Technology type Energy Service demand
Diesel Trucks Diesel ICE Truck tkm
Vans Diesel ICE Van tkm
Rail freight Diesel Rail Rail tkm
Fuel oil
Fuel oil shipping
Shipping Shipping tkm
Diesel shipping
Aviation fuel Air freight Aviation fuel Air tkm
Figure 41: Pak-IEM Transport Sector RES Diagram – Freight Vehicles

Sector data requirements and sources

The approach presented earlier in this section requires the following types of data:
• Vehicle stock data
• Stock activity
• Characterization of existing vehicle technologies
The focus of the data requirements presented here are for base year modeling. Table 39
lists the main data requirements for different transport technologies.
Table 39: Parameters Required for Characterization of Transport Sector Technologies

Vehicle type
Existing stock Number
Annual distance travelled Km/yr
Input energy carrier
Average occupancy Passenger/vehicle
Remaining life (existing) Years TIMES can also use the start year for existing vehicles
Lifetime (new) Years New vehicles
Efficiency passenger-km/PJ or
tonne-km/PJ
Investment cost $/vehicle New vehicles only
Fixed O&M Cost $/vehicle/a
Variable O&M Cost $/vehicle/km
Capacity factor 1
Input share Fraction For dual fuel vehicles
Output passenger-km Calculated (not data)
Vehicle stock
Table 40 provides estimates of vehicle stock, which are sourced from “On the road”
estimates provided by NTRC (2008).
Table 40: Vehicle Stock Statistics, 2006/07 (NTRC 2008) and Fuel Split Assumptions
Diesel / Petrol CNG conversions (petrol /
Vehicle type* 000s vehicles
splits CNG split)
Buses 108.4 1/0
Cars 1872.2 0.15/0.85 82% (0.02/0.98)
Trucks 182.1 1/0
Vans 148.9 1/0
Two-wheelers 4463.8 0/1
Three-wheelers 79.0 0/1

Diesel / Petrol CNG conversions (petrol /

Vehicle type* 000s vehicles
splits CNG split)
Taxis 119.1 0.05/0.95 80% (0.04/0.96)
Minivans 173.7 1/0
Other 38.5
Total 7185.8
*
NTRC tractor estimates have not been included; they are dealt with in the agriculture sector
An alternative dataset of stock is also published by FBS (2008). Significant efforts went into
understanding which is the most appropriate dataset to use. We ended up using the NTRC
estimates, as we believe these provide the most accurate estimates of vehicles actually on
the road. According to NTRC, they use records from regional excise and taxation
departments so as to estimate not simply registered vehicles but those actually being used.
FBS clarified what they believe to be the differences between the datasets – FBS data on
road transport includes only those vehicles which are registered with the excise department
while NTRC figures are based on a survey which includes three categories of vehicles: (i)
registered with excise department, (ii) exempted to registration from excise department, and
(iii) smuggled vehicles running on roads.53 This explanation appears to make sense, given
that the NTRC estimates are higher at the aggregate level. However, for specific road
vehicle categories, FBS stock estimates are higher but not for others.
No official source on the split between diesel and petrol vehicles has been identified. The
estimates are therefore based on expert judgment and a paper by ENERCON (2003).
Compressed Natural Gas (CNG) vehicles in 2006/7 were estimated at 1.4 million vehicles
(OGRA 2008),54 and all assumed to be petrol-vehicle conversions. (This value has been
cross-checked against other sources, notably the Hydrocarbon Development Institute of
Pakistan). Currently, conversions are assumed to be only on car / taxi vehicle stocks, with
expert judgment on petrol / CNG usage splits for converted vehicles.
CNG conversions have taken off significantly in recent years due to the price differential
between CNG and petrol. These vehicles have been set up as dual-fuelled vehicles in the
model, allowing for fuel switching depending on price.
There has been a recent emergence of CNG vehicles across other vehicle types (rickshaws,
buses) – e.g., the government of Punjab has mandated that all public-transport vehicles will
use CNG by 2007.55 However, these are in the stock numbers post-2006/7 so do not
feature in the base year.
Stock activity and technology characterization
A range of different technology assumptions are required to characterize different vehicles,
notably vehicle efficiency, annual activity and load / occupancy. These have been used to
develop the bottom-up estimates of energy use by vehicle type. The primary source of data
is an ENERCON paper (2003), shown in Table 41. NTRC (2005) provides a comparable
dataset with similar assumptions.
53
Based on personal communication with FBS.
54
Estimate of 1.55 million on IANGV website, www.iangv.org, (Accessed 2nd April 2009, 14.00); (NB. Website
also has time series information on numbers of CNG vehicles that may help inform future trends).
55
PakCNG.com, www.pakcng.com/cng_info.asp; (Also see www.cng.com.pk/index.html for additional
information.)

Table 41: Efficiency and Activity Assumptions for Different Vehicle Types, ENERCON
2003; Lifetime and Occupancy Values Sourced from NTRC (JICA 2006b)
Activity Efficiency Life Occupancy /
Vehicle type** Size Fuel
(km/yr) (mvkm/PJ) (Years) Load
Gasoline 32,500 403 12 3
Taxi cabs Small
CNG 32,500 400 12 3
Buses All Diesel 50,400 109 10 45
Minibuses All Diesel 50,400 305 10 17
Trucks All Diesel 50,400 97 12 3 (ton)
Gasoline 12,600 314 12 2
Medium Diesel 12,600 377 12 2
Motor cars
CNG 12,600 315 12 2
Large Diesel 12,600 226 12 2
Vans All Diesel 28,000 305 10 0.5 (ton)
Motorcycles All Gasoline 10,000 1531 10 1.4
Rickshaws All Gasoline 40,000 349 10 2
** Includes jeeps / pickups, classified as large cars

New transport technologies

A range of new technologies have been included in the model for road transport, and are
shown in the Box below.
Cars (in Pak-IEM, cost / efficiency factors relevant for Buses (in Pak-IEM, cost / efficiency factors relevant
all LDVs) for minibuses)
INCLUDED: INCLUDED:
Conventional ICE Diesel / Gasoline ICE Diesel
Advanced ICE Diesel / Gasoline Hybrid Diesel
Hybrid Diesel / Gasoline Electric
Electric Vehicle CNG
Flex-ethanol (fuel sugar cane derived)
CNG EXCLUDED:
Fuel cell
EXCLUDED:
Hydrogen ICE / Fuel cells TO BE CONSIDERED IN FUTURE MODEL
VERSIONS:
TO BE CONSIDERED IN FUTURE MODEL Non-technical options (driving behaviour etc)
VERSIONS:
Small cars (e.g. Tata)
Non-technical options (driving behaviour etc)
Blending biofuels in mainstream fuels (in upstream
sector)
Trucks 2 wheelers
INCLUDED: INCLUDED:
ICE Diesel ICE Gasoline (2-stroke / 4-stroke)
Diesel hybrid
3 wheelers
EXCLUDED:
Hydrogen ICE / Fuel cells INCLUDED:
ICE Gasoline (2-stroke / 4-stroke)
TO BE CONSIDERED IN FUTURE MODEL Gasoline Hybrid
VERSIONS: Electric
Hybrid electric / CNG (these are in the UK model CNG
although not sure if serious options) CNG Hybrid
Non-technical options (driving behaviour etc, better
maintenance, reducing unnecessary weight)
The technology assumptions by road vehicle type are discussed in turn below.
Cars
Conventional technology
These technologies represent current vehicles on the market; future improvements in ICE
technologies are incorporated in the advanced technology type.
The capital cost of diesel / gasoline cars are taken from the NTRC published VOC study
(2005). All other technologies are factored against the costs of these vehicles, and the basic
efficiency values.
Advanced technology
These options include improvements to vehicle powertrain systems that improve overall
efficiency. For the purposes of the model, non-engine improvements (e.g. aerodynamics,
material weight) will be assumed to be incorporated in these vehicles – and other future
technologies.

Current year cost differential and efficiency improvements are based on IEA (2008). Future
efficiency improvements out to 2050 (of 20-25%) are based on IEA (2005), and cross
checked against IEA/WBCSD SMP Model (2004).
Hybrids
These are vehicles that incorporate an electric motor, running off a battery that recharges
while the gas engine is being used and also through regenerative braking. The electric motor
is used whilst the vehicle is moving slowly or cruising, and power requirements are less. This
combination allows for a much more efficient use of the internal combustion engine, allowing
it to operate steadily at near-optimal loads. (Hybrid technologies can be categorized into
partial or full hybrid systems; only generic (no distinction between series or parallel) full
hybrid system used in model).
The cost increase for hybrids is based on IEA (2008), as is the typical efficiency
improvement of between 46-48%. Future cost reductions are based on improved and more
cost-effective battery technologies. Efficiency improvements in future years are based on
Ricardo estimates, as cited in UKERC (2007).
Plug-in hybrid vehicles (PHEVs)
These vehicles allow charging from the grid rather than the internal recharging system, and
therefore improve energy use per KM. The main challenge is battery range and cost.
Capital costs are again based on IEA (2008), and a battery range of 50km (12.5 kWh battery
storage). These decrease significantly in future years, by 35-40%, driven by improvements
to battery technology (IEA 2008, IEA 2005). Efficiency of electric mode is based on UKERC
(2007), assuming 10% less efficient than full electric vehicles due to additional weight of
gasoline / diesel engine, reducing to 5% in later years. Efficiency of an electric vehicle is
assumed to be four times greater than gasoline car (current year) and by up to five times in
2050 (UKERC 2007, IEA 2005/2008). The improvement in future years is partly due to better
regenerative braking technology. The efficiency of the vehicle when using diesel / gasoline
engine is slightly lower (5%) than that of the advanced engine technology (as described
above).
Electric vehicles
These vehicles have no internal combustion engine, relying entirely on an electric system.
Again the main challenge is battery range and cost. There is a view (ETP 2008) that the
success of PHEV will determine the timing of electric vehicles, and their full
commercialization.
The vehicles have costs 20% higher (in current year) and 10% higher in 2050 than PHEVs
due to an assumed higher travel range of 100 km, rising to 270 km by 2050 (UKERC 2007,
IEA 2005). Efficiency estimates are described in the PHEV section above.
Flex-ethanol 85
These vehicles can use up to 85% ethanol, and are seen in widespread use in countries
such as Brazil. The ethanol supply in Pakistan is likely to be derived from sugar cane
(although this will be influenced by the price of sugar, which has recently been at an all-time
high).
Efficiencies are assumed to be the same as advanced gasoline vehicles in future years.
Based on information in UKERC 2007, a small increment ($200) on capital costs is assumed
relative to cost of a gasoline car.

CNG
Efficiency factors for new CNG vehicles track the estimates for advanced gasoline vehicles,
scaling for a 5% reduction (IEA/WBCSD 2004).
Vehicle O&M costs
O&M costs are modelled relatively simply. There are two elements (represented as fixed
costs as we assume that each vehicle by category travels the same distance in a year):
• Servicing and general maintenance (including replacing tires) and
• Parts replacement e.g. batteries in electric vehicles
Other options
Another option relevant to the transport sector that will be characterized in the upstream
sector is blending plant, to allow for small proportions of biofuel (probably ethanol from sugar
cane) to be introduced into the gasoline supply.
There is also the potential for a small vehicle option (e.g. Tata) and non-technical options
(slower driving, proper maintenance). These will be further discussed in consultation with
Pakistan transportation sector experts.
Hydrogen-based vehicles have not been considered as an option for the model due to the
small likelihood of these technologies being commercially available in the model timescales,
and the significant infrastructure requirements.
Two- and three-wheelers
For new two-wheelers, the options are divided into:
• 2-stroke (Assume that most of current stock are 2 stroke), and
• 4-stroke
Electric-powered two-wheelers are another option that could be included, although for the
time being this has been omitted (subject to further discussion). TERI (2006) also has
hydrogen ICE engine technologies; however, hydrogen-based technologies have been
omitted from the model.
Data for these technologies have been sourced from TERI (2006). In considering the co-
benefits of specific technologies, it is also worth mentioning the significant benefits that 4-
stroke engines have in reducing air quality pollutant emissions, as compared to 2-stroke
engines.56
For three-wheelers, the options identified and data are sourced from TERI (2006). A 4-stroke
diesel technology was also listed, so some discussion with local counterparts to decide on
whether this is also necessary to include.
Buses
ICE Diesel
The efficiency of conventional diesel buses is taken from the base year technology, and
improved based on LDV ICE efficiency improvements over time. Capital costs are assumed
to remain constant. Note that the cost of buses is very low from NTRC (2005), making the
56
For further information, see PCFV, Cleaner motorcycles: Promoting the use of four-stroke engines, Kjaer
Group / Partnership for Clean Fuels and Vehicles (PCFV),UNEP.

application of mutlipliers for efficiency and cost increase / reduction problematic. For
example, costs of a 40+ seat bus are ~USD 90,000. In the international literature, costs are
generally compared to diesel buses at around USD 250-300,000.
Hybrid Diesel (HEVs)
Hybrid electric buses are assumed to be 40% more efficient than conventional diesel buses,
based on their use in major cities in the UK / USA (Transport for London, US Department of
Transportation (DoT) 2005). The rate of future efficiency improvement is based on hybrid
cars.
Cost estimates (based on the same sources) put HEV at 60% more than conventional buses
for UK and US sources. This is based on US experience, with costs of between ~USD 450-
500,000 (EESI 2007) and in the UK, costs of USD 350-400,000 (UKERC 2007). One factor
explaining the variation will be the number of buses ordered at any given time, resulting in
economies of scale.
TERI (2006) suggests ~USD 200,000, at 3.4 times the cost of a conventional bus. This is the
factor that is being used in the model, to account for the very low cost of a conventional
diesel bus. Future cost reductions are linked to reduced battery costs (as per light vehicles).
Electric buses
Efficiency estimate (1kWh/km) based on UK information (DfT 2005). Future efficiency
improvements based on those cited in UKERC 2007. Cost estimates again based on
UKERC 2007, based on the assumed battery range of vehicle (200km in 2008, increasing to
300 km post-2020). Future cost reductions linked to reduced battery costs (as per light
vehicles).
CNG buses
The fuel economy of a CNG bus is considered to be 25% less than a conventional diesel bus
(US DoT 2005, EESI 2007). Capital costs are assumed to be 1.5 times higher than a
conventional bus (TERI 2006).
Trucks
ICE Diesel
This technology represents efficiency improvements that could be realized for ICE
technologies in future years. The IEA (2008) ETP publication is used as the basis for
efficiency improvement in new vehicles: - 20% by 2030, and 30% by 2050 (for non-OECD
regions). This efficiency improvement is based on IEA ETP baseline efficiency improvement;
potential could be higher depending on:
• Engine/drivetrain efficiency improvements, e.g. advanced higher-compression diesel
engines
• Truck weight reduction
• Aerodynamic improvements
• Low-rolling resistance tyres
• More efficient auxiliary equipment, such as cabin heating/cooling systems and

lighting

It would appear that there may be significantly greater potential in Pakistan than other
countries for reducing truck weight / improving aerodynamics. Local information is therefore
needed to further understand potential in future years.
Hybrid Diesel
Hybrid drivetrain technology has been included as an option, involving adding a
motor/battery system to significantly improve urban cycle operation. Clearly, efficiency is
dependent on urban versus inter-city transport; the IEA put the efficiency improvement range
between 25-45%. Cost reduction is a function of changes predicted in the LDV vehicle
category.
Summary of transport sector data sources

A summary of transport sector sources is provided in Table 42.
Table 42: Summary of Data Sources for the Transport Sector
Road transport
Motor Vehicles on Road by Region Registration data from FBS also
Existing vehicle stock
2006-07, (NTRC 2008) reviewed but not used
Fuel Efficiency, Vehicular Emissions
Activity by vehicle type and Pollution Control in Pakistan
(ENERCON 2003)
Fuel Efficiency, Vehicular Emissions
Fuels consumed and efficiency
and Pollution Control in Pakistan
(l/km) by vehicle type
(ENERCON 2003)
Pakistan Transport Plan Study (JICA
Vehicle life / Occupancy
2006)
Vehicle Operating Costs using HDM
Operating costs
– VOC Version 4.0 (NTRC 2005)
IEA ETP (2008), IEA (2005), See references in previous
Characterization of future
IEA/WBCSD (2004), National section
technology options
Energy Map for India (TERI 2006)
Non-road transport
Pakistan Railway Statistics
Rail
Handbook 2006/07
Aviation Pakistan Oil Report 2006-07
Shipping Pakistan Oil Report 2006-07

VII. DEMAND PROJECTIONS

A. Overview
Useful energy demand projections, or the demand for energy service, are a critical input to
any MARKAL/TIMES model, as the model will identify the least-cost configuration of the
energy system that will satisfy these demands. For Pak-IEM, modeling the future evolution
of the energy system requires projecting the future demand for energy services using
suitable demand drivers (e.g., GDP, population, energy intensity, etc.). The drivers for each
energy service demand are those that most directly impact its potential growth, and the
relationship (e.g., elasticity) between each driver and the associated energy service demand
must also be established. Finally, any external factors affecting the demand growth (e.g.,
autonomous energy improvement) also need to be determined. Table 43 identifies both the
primary energy service demand drivers by sector and some additional factors affecting the
growth of that sector or sub-sector. Most of these additional factors have not been
incorporated into the current demand projections for Pak-IEM because sufficient data is not
available. These factors could be incorporated into future versions of Pak-IEM if suitable
data can be developed.
Table 43: Energy Service Demand Drivers by Sector
Sector Demand Primary Drivers Additional Factors
Agriculture Agricultural energy Sector GDP Increased pumping due to climate

services (pumping / Energy service change
tractor activity) elasticity Changes in farming techniques
(e.g., tilling, fertilizer use)
Commercial/ Buildings energy Sector GDP Growth of service and

Public services government sectors
Sector Decommissioning of existing
building stock
Building standards for new
construction
Industry Industrial product Sector GDP Sector structural change (e.g.

output Construction levels new ventures, re-tooling etc)
Sub-sector specific Changes in management
drivers (e.g. practices
agricultural output for
fertilizer sub-sector)
Residential Household energy Population, Occupancy / demographic issues,

services urbanization, including income levels
Electrification rate, Changes in existing housing
Household energy stock
intensity vs. GDP/Cap Building standards for new
Income dwellings
Climate issues (temperature rise)
Transport Passenger and GDP Vehicle choice

freight demand by Income Occupancy rates
transport mode
Public transport development –
modal shift

The Pak-IEM demand projection workbook was developed and contains projections for GDP
growth by sector, population growth, urbanization rates, and household size for urban and
rural families leading to projections of urban and rural households. All data is sourced from
official GoP sources where available. The demand projection workbook is described in more
detail in the Pak-IEM Final Report: Volume III – User's Guide.
B. Primary Demand Drivers

Reference case GDP growth rates are based on the official GoP targets, as shown in Figure
42. The 2007-2030 aggregate average annual GDP growth rates for this Reference (REF)
scenario is 5.6%, and the actual GDP growth is shown in Figure 43. Medium (MED) and
Low (LOW) GDP growth projections were prepared by the Planning Team based on
adjustments to the GDP component growth rates, an shown in Table 44. The MED
aggregate average GDR growth rate for 2007 to 2030 is 5.0%, and the similar growth rate
for the LOW GDP scenario is 4.2%. As shown in Figure 44 overall GDP growth rate was
kept constant for all scenarios between 2030 and 2045, and the sectoral GDP growth rates
were adjusted to accomplish this.
Year Agriculture Industry Services GDP (fc)
2010 3.8 1.7 3.9 3.3
2011 3.9 4.2 4.5 4.3
2012 4.0 5.0 5.2 4.9
2013 4.1 5.9 5.9 5.5
2014 4.1 6.8 6.6 6.1
2015 4.1 7.6 7.1 6.6
2016 4.1 8.0 6.9 6.6
2017 4.1 8.2 6.8 6.6
2018 4.1 8.5 6.4 6.5
2019 4.1 8.8 6.3 6.5
2020 4.0 10.0 5.5 6.4
2021 4.0 10.0 5.1 6.2
2022 4.0 10.0 4.8 6.1
2023 3.8 10.0 4.7 6.1
2024 3.6 10.0 4.5 6.0
2025 3.5 10.0 4.4 6.0
2026 3.2 9.8 4.3 5.9
2027 3.0 9.6 4.4 5.9
2028 3.0 9.4 4.2 5.8
2029 3.0 9.2 4.2 5.8
2030 3.0 9.0 4.2 5.8
Average 3.7 8.2 5.2 5.9
Figure 42: Official GoP GDP / Sectoral Growth Targets

GDP Projections
35,000
30,000 REF GDP (fc)
25,000 MED GDP (fc)

Billion 2007 Rupees
LOW GDP (fc)

20,000
15,000
10,000
5,000
0
2007 2009 2012 2015 2018 2021 2024 2027 2030
Figure 43: Reference, Medium and Low GDP Growth
Note that all the GDP growth scenarios have a significant increase in the Industry GDP
contribution, and that this sector also shows the most significant decrease in the MED and
LOW growth projection scenarios.
Table 44: GDP Growth Rates Components for Medium and Low Projections
AGRICULTURE 2007 2009 2012 2015 2018 2021 2024 2027 2030
MED Growth rate 2.60% 3.80% 3.80% 3.90% 4.00% 3.80% 3.60% 3.00% 2.75%
LOW Growth rate 2.60% 3.80% 3.60% 3.70% 3.80% 3.50% 3.25% 3.00% 2.50%
INDUSTRIAL 2007 2009 2012 2015 2018 2021 2024 2027 2030
MED Growth rate 5.30% 1.70% 5.00% 7.00% 8.00% 8.50% 8.00% 7.50% 7.50%
LOW Growth rate 5.30% 1.70% 5.00% 6.00% 6.50% 6.00% 6.00% 6.00% 6.00%
COMMERCIAL 2007 2009 2012 2015 2018 2021 2024 2027 2030
MED Growth rate 5.77% 3.92% 3.60% 5.34% 5.65% 4.94% 4.36% 3.90% 3.77%
LOW Growth rate 5.77% 3.92% 2.75% 3.09% 4.62% 4.58% 3.84% 3.05% 3.11%

Overall GDP Growth Rates

REF Growth rate MED Growth rate LOW Growth rate
7.00%
6.00%
5.00%
4.00%
3.00%
2.00%
1.00%
0.00%
2009 2012 2015 2018 2021 2024 2027 2030 2033 2036 2039 2042 2045
Figure 44: Overall GDP Growth Rate Projections
Figure 45 provides a breakdown of the relative sectoral GDP shares for the three demand
projection scenarios. Of interest is the relative share of industry sector relative to services
and agriculture. Note that the economic share of commercial and services is much greater
than its share of the energy use. Also, note that the industry share of GDP grows
dramatically in all the demand scenarios.
Sectoral GDP Contributions

100%
90%
80% Commercial
& Services
70%
60%
Industry
50%
40%
Agriculture
30%
20%
10%
0%
2007
2009
2012
2015
2018
2021
2024
2027
2030
2033
2036
2039
2042
2045
2007
2009
2012
2015
2018
2021
2024
2027
2030
2033
2036
2039
2042
2045
2007
2009
2012
2015
2018
2021
2024
2027
2030
2033
2036
2039
2042
REF Contributions MED Contributions LOW Contributions
Figure 45: GDP Sectoral Contributions for Reference, Medium and Low Projections

The projections for population growth were taken from the draft Population Policy 2010
report.57 As depicted in Figure 46, the overall population in Pakistan is expected to grow
from 158 million in 2007 to 243 million in 2030. The share of rural population is expected to
change from 66% in 2007 to 50% in 2030, and the average household size is expected to
drop for both urban and rural households.
Population and Household Size

300.0 7.00
6.80
250.0
6.60
Persons per household

200.0 6.40
Population (Millions)
6.20
150.0
6.00
100.0 5.80
5.60
50.0
5.40
0.0 5.20
2007 2009 2012 2015 2018 2021 2024 2027 2030
Urban Population Rural Population Urban household size Rural household size
Figure 46: Projections of Population and Household Size
C. Demand Projections by Sector

5. Agriculture
The agriculture sector energy service demands were developed for the following end-use
applications. All demands are in units of PJ.
• Water pumping
• Tractors – Haulage
• Tractors – Farm Operations
• Other Use
The agricultural component of GDP growth is the principal driver for this sector, and each
end-use sector is related to the GDP growth through an elasticity factor, which is defined as
the percentage change in energy consumption divided by the percentage change in GDP
growth. The elasticity factors are either zero or negative, which if zero means that energy
use grows at the same rate as GDP grows, or if negative means that energy use grows more
slowly than GDP growth. The elasticities are used to calculate factors that when multiplied
by the GDP growth gives the effective energy growth. The agricultural sector elasticities are
57
National Population Policy – 2010, Ministry of Population Welfare, Islamabad, Jan 18, 2010.

based partly on expert judgment as well as data from the Energy Yearbook that was used to
calculate a historical elasticity factor for the entire sector, which was -1.05%.
Table 45: Agricultural Sector Energy/GDP Elasticity Factors
Energy/GDP Elasticity Factor (Overall change in energy use relative to GDP
Demand
growth)
Water Pumping - 0.5% per annum (11% less energy growth by 2030 relative to GDP growth)
Farm Operation - 0.75% per annum (16% less energy growth by 2030 relative to GDP growth)
Haulage - 1.0% per annum (21% less energy growth by 2030 relative to GDP growth)
Others - 0.5% per annum (11% less energy growth by 2030 relative to GDP growth)
Given the agricultural sector GDP demand driver, the base year useful energy demands,
and the energy/GDP elasticities, the actual projections for the agricultural sector demands
are shown in Figure 47.
Reference Scenario
Agricultural Useful Energy Demands
0.80
0.70
0.60
0.50 Water pumping
MTOE
0.40 Tractors ‐ Haulage

0.30 Tractors ‐ Farm Op.
0.20 Other Use
0.10
0.00
2007 2009 2012 2015 2018 2021 2024 2027 2030
Figure 47: Agricultural Sector Useful Energy Demands
6. Commercial
The commercial sector energy service demands were developed for the following end-use
applications. All demands are in units of PJ.
• Space Heating
• Water Heating
• Space Cooling (Fans and AC)
• Lighting
• Refrigeration
• Electrical Appliances
• Other

The services component of GDP growth is the principal driver for this sector, and each end-
use application is related to the GDP growth through an elasticity factor, which is defined as
growth. The elasticity factors are either zero or negative, which if zero means that energy
use grows at the same rate as GDP grows, or if negative means that energy use grows more
slowly than GDP growth. The elasticities are used to calculate factors that when multiplied
by the GDP growth gives the effective energy growth. The commercial sector elasticities
are based on expert judgment.
Table 46: Commercial Sector Energy /GDP Elasticity Factors
Energy/GDP Elasticity Factor (Overall change in energy use relative to
Demand
GDP growth)
Space Heating -1.0% per annum (21% less energy growth by 2030 relative to GDP growth)
-0.75% per annum (16% less energy growth by 2030 relative to GDP
Water Heating
growth)
Space Cooling (Fans

No change relative to GDP
and AC)
Lighting -0.5% per annum (11% less energy growth by 2030 relative to GDP growth)
Refrigeration No change relative to GDP
Electrical Appliances No change relative to GDP
Other -1.0% per annum (21% less energy growth by 2030 relative to GDP growth)
Given the commercial sector GDP demand driver, the base year useful energy demands,
and the energy/GDP elasticities, the actual projections for the commercial sector end-use
demands are shown in Figure 48.

Commercial Sector Useful Energy Demands

1.40
1.20
Space Heating
1.00
Water Heating
0.80
MTOE
Space Cooling (Fans and AC)

0.60 Lighting
Refrigeration
0.40
Electrical Appliances
0.20 Other
0.00
2007 2009 2012 2015 2018 2021 2024 2027 2030
Figure 48: Commercial Sector Useful Energy Demands
7. Industry
The industry sector demand projections were developed according to the following
categories. Most of the demands are expressed in activity units, such as million tons (Mt),
while a few are expressed in PJ, and one is expressed in billion square meters (Bm2).
Sub-sector demand category Units
• Textiles Yarn production Mt
• Textiles Cloth production Bm2
• Textiles Garment production PJ
• Sugar production Mt
• Cement production Mt
• Iron and Steel Integrated Pig Iron Mt
• Iron and Steel Integrated Cast Billet Mt
• Iron and Steel Integrated Steel Products Mt
• Iron and Steel Integrated Other Processes PJ
• Iron and Steel Independent Billets (Independent) Mt
• Iron and Steel Independent Construction steel Mt
• Iron and Steel Independent Other (Independent) PJ
• Brick kilns PJ
• Fertilizer production Mt
• Other industry PJ
The industrial component of GDP growth is the principal driver for many, but not all of the
industry sub-sectors. Because of the diversity of the various industry sub-sector demands,
each uses a slightly different approach to the demand projection, which is described below.
In addition, because of the different units employed, only the relative demand growth for the
various industry sub-sectors are shown in Figure 49.

Industry Sector Relative Demand Growth

6.00
5.00 Textiles
4.00 Sugar
Cement
3.00
Iron & Steel
2.00
Brick
1.00 Fertilizer
Other
0.00
2007 2009 2012 2015 2018 2021 2024 2027 2030
Figure 49: Industry Sector Relative Demand Growth
The drivers and elasticity factors used in each industry sub-sector are summarized in Table
47 and discussed in the sub-sections below.
Table 47: Industry Sub-sector Elasticity Factors
Demand Driver Energy/GDP Elasticity Factor (Overall change in energy
use relative to GDP growth)
Specific growth factors for yarn (5.1%), cloth (7.7%) & garments (6.4%) based on
Textiles
2010 Economic Survey data. See Table 48)
Sugar Sugar Cane Production growth of 2%
Construction
Cement Variable (See Table 49)
component of GDP
-1.0% per annum (21% less energy growth by 2030 relative

Iron & Steel Industrial GDP
to GDP growth)
Construction -1.0% per annum (21% less energy growth by 2030 relative
Brick
component of GDP to GDP growth)

Fertilizer Agricultural GDP
to GDP growth)

Other Industrial GDP
to GDP growth)

a) Textiles
The textile sub-sector energy service demand was not developed using the industry sector
GDP growth as the driver. Because of the importance of this sub-sector and because of the
international competitive factors that strongly influence its growth, it was decided to develop
independent growth factors for each segment of the textile sub-sector based on data from
the 2010 Economic Survey. Those growth factors are shown in Table 48 and are given in
REF, MED and LOW. The growth rates for each textile industry segment and demand
scenario were developed based on near-term issues as well as expected long-term trends.
Table 48: Textile Industry Component Growth Rates
Textile Industry 2009 2012 2015 2018 2021 2024 2027 2030
REF 1.38% 3.50% 5.00% 6.00% 6.00% 6.00% 6.00% 6.00%

Yarn
MED 1.38% 2.00% 3.00% 4.00% 4.00% 4.00% 4.00% 4.00%
Production
LOW 1.38% 1.50% 1.50% 1.50% 1.50% 1.50% 1.50% 1.50%
REF 5.03% 6.00% 8.00% 9.00% 9.00% 9.00% 8.00% 7.00%

Cloth
MED 5.03% 5.50% 6.00% 6.00% 7.50% 7.50% 6.00% 5.00%
Production
LOW 5.03% 5.00% 5.00% 5.00% 5.00% 5.00% 5.00% 4.00%
REF -4.57% 1.00% 2.00% 3.00% 4.00% 5.00% 5.00% 5.00%

Garment
MED -4.57% 0.00% 1.00% 2.00% 3.00% 3.00% 3.00% 3.00%
Production
LOW -4.57% -1.00% 0.00% 1.00% 1.00% 1.00% 1.00% 1.00%
b) Sugar Production
The sugar sub-sector energy service demand was developed using the projected growth
rate in sugar cane production58 as the principal driver – and limitation on growth in the sub-
sector. No elasticity factor or other factor was used. The report gives 2.5% growth potential,
but for Pak-IEM this projection was moderated to 2% per annum.
c) Cement production
The cement sub-sector energy service demand was developed using the industry sector
GDP growth as the principal driver, a construction sub-component elasticity, and a cement
energy use elasticity factor (defined as the percentage change in energy consumption
divided by the percentage change in GDP growth), which is based on review of cement
sector trends in the 2010 Economic Survey.
Historically, the construction-related GDP growth relative to the industrial GDP growth was
found to be highly variable, but the long-term average elasticity is zero, meaning that the
construction component of industry GDP growth grows at the same average long-term rate
as industry as a whole. Regarding the energy to GDP elasticity, the current trend toward
cement exports is supporting strong near-term growth, but this trend is not expected to last
for the long term. Therefore, the elasticity factor was varied over time (See Table 49) to
allow for adjustment in this trend.
58
Potential of Bagasse Based Cogeneration in Pakistan, Dr. Khanji Harijan, Dr. Mohammad Aslam
Uqaili & Dr. Mujeebuddin Memon, World Renewable Energy Congress (WRECX) Editor A. Sayigh ©
2008 WREC.

Table 49: Cement Industry Energy / GDP Elasticity

Cement Industry 2009 2012 2015 2018 2021 2024 2027 2030
Energy / GDP
2.00% 2.00% 1.00% 0.00% -1.00% -2.00% -2.00% -2.00%
Elasticity
d) Iron and Steel

The steel sub-sector energy service demand was developed using the industrial component
of GDP growth as the principal driver, and an energy use to GDP elasticity factor (defined as
growth). An elasticity of -1%, based on expert judgment, was used for the period up through
2030. While there are seven process-level demand for iron and steel, they are linked and so
use the same values for the sub-sector elasticities.
e) Brick Kilns
The brick kiln sub-sector energy service demand was developed using the industry sector
GDP growth as the principal driver, a construction sub-component GDP elasticity and a brick
kiln energy use elasticity factor (defined as the percentage change in energy consumption
divided by the percentage change in GDP growth). An elasticity of -1% was used, which is
based on expert judgment. Historically, the construction-related GDP growth relative to the
industrial GDP growth was found to be highly variable, but the long-term average elasticity is
zero, meaning that the construction component of industry GDP growth grows at the same
average long-term rate as industry as a whole.
f) Fertilizer
The fertilizer sub-sector energy service demand was developed using the agricultural
component of GDP growth as the principal driver, and an elasticity factor (defined as the
percentage change in energy consumption divided by the percentage change in GDP
growth) of -0.5%, which is based on expert judgment.
g) Other Industry
The Other industry sub-sector energy service demand was developed using the industrial
component of GDP growth as the principal driver, and an energy use elasticity factor
(defined as the percentage change in energy consumption divided by the percentage
change in GDP growth) 0f -1.5%, which is based on expert judgment.
8. Residential
The residential sector energy service demands were developed for the following end-use
applications for both urban and rural households. All demands are in units of PJ.
• Space Heating
• Water Heating
• Space Cooling (Fans and Coolers)
• Space Cooling (AC)
• Lighting
• Cooking
• Refrigeration
• Miscellaneous Electric
• Other

The residential sector energy service demands are primarily driven by the growth in
households, which are divided into urban and rural categories and further classified
according to whether they are connected to the electricity grid. The projected number of
electricity customers (households) is shown in Figure 50 for both urban and rural areas. For
urban areas, the percentage of connected households is projected to go from 98% in 2007
to 100% in 2012. For rural areas, the number of connected households starts at the 2007
level and grows in proportion to the growth in rural population with additions for village
electrification and migration from un-electrified areas to electrified villages. Rural households
start at 81% electrified in 2007 and grow to 91% electrified in 2030.
Two other significant drivers to residential energy demand are the percentage of households
with air conditioners and the percentage of households with refrigerators. Figure 50 shows
the aggregate annual growth rate of these two major appliances for both urban and rural
households. Households with refrigerators are expected to grow at an annual rate of 6% for
urban households and 4% for rural households. Households with air conditioners are
expected to grow at an annual rate of 6% for urban households and 3% for rural households.
Household Electricity Consumers

45.0 20%
40.0 18%
Percentage of Total Households

35.0 16%
14%
Million Households
30.0
12% Rural Household Consumers
25.0
10% Urban Household Consumers
20.0
8% Households with refrigerators
15.0 Households with air conditioning
6%
10.0 4%
5.0 2%
0.0 0%
2007 2009 2012 2015 2018 2021 2024 2027 2030
Figure 50: Number of Urban and Rural Electricity Consumers
For each end-use application, the average household energy use, as shown in Table 50, is
determined from the disaggregation of the energy balance data combined with the
urban/rural household data. As GDP in Pakistan grows, per-household energy use is
projected to grow, and data developed by the World Resources Institute59 was used to
calculate the per capita energy intensity increase as shown in Figure 51, which shows the
relationship between energy use per capita and GDP per capita.
59
EarthTrends (http://earthtrends.wri.org) Searchable Database Provided by the World Resources Institute
(http://www.wri.org): Energy and Resources – Energy Consumption: Residential energy consumption per capita.

Table 50: Average Household Energy Use in 2007

Average Household Energy Use in 2007 (kgoe)
Demand
Urban Rural
Space Heating 42.7 67.0
Water Heating 145.7 76.9
Space Cooling (Fans and Coolers) 51.6 39.2

Space Cooling (AC) 16.5 3.6
Lighting 45.7 42.8
Cooking 542.4 581.4
Refrigeration 22.8 6.4
Miscellaneous Electric 22.3 12.0
Other 10.6 14.1
Comparison by country ‐ GDP vs. Residential

energy use (both per capita)
600
500
y = 0.0189x + 129.69
400
R² = 0.6992
kgoe / capita
300
200
100
0
0 5000 10000 15000 20000 25000
GDP per capita ($US 2000)
Figure 51: Comparison of Energy Use per GDP for Pakistan and Other Countries
The urban and rural household base year final energy demand for each end-use application
was calculated from the projected per capita energy use values multiplied by the relevant
household population. To determine useful energy consumption at the end-use application
level for both urban and rural households, the energy balance data was disaggregated – first

by fuel type and then by end-use fuel share – so that typical device efficiencies could be
estimated and applied to calculate useful energy demands.
The energy balance data was first apportioned between urban and rural households based
on PSLM60 data on percentages of households consuming various fuels. The relative
energy use for urban and rural households is take from HESS61 data to calculate the fuel
shares between urban and rural households, which when multiplied by the energy balance
data gives the urban and rural fuel use. This process is illustrated in Figure 52.
Urban Fuel Type

Residential
Fuel Type
Fuel Use Fuel Use Shares End-Use Fuel
Share by End-Use
Type
Fuel Type Share
Urban Fuel Type End Use Type

Residential Fuel
Use by End-Use
Application
Energy End Use Share
Balance
Residential End Use Type
Fuel Use
Fuel Type Fuel Type

Base Year End- Useful Energy
Use Device Demand by End
Efficiency Use
Device Useful Energy
Efficiency Demand
End Use Type End Use Type
Rural
Residential Fuel Use
{Similar Processes]
Figure 52: Diagram of Useful Energy Demand Development

The energy consumption shares by end-use application were calculated based on the fuel
shares and energy consumption totals by end-use application. The base year efficiency
value for typical devices by application and fuel type were based on expert judgment and
were used to calculate useful energy consumption by end-use and fuel type. The urban and
rural useful energy demands for each end-use application are shown in Figure 53 and Figure
54, respectively.
60
Pakistan Social & Living Standards Measurement Survey PSLMS 2006-07.
61
Pakistan Household Energy Strategy Study, Household Energy Demand handbook for 1991, The World Bank.

Urban Residential Sector Useful Energy Demands

8.00
Urban Cooking
7.00
Urban Water Heating
6.00
Urban Space Heating
5.00 Urban Space Cooling (Fans and
MTOE
Coolers)
4.00 Urban Space Cooling (AC)
3.00 Urban Lighting
2.00 Urban Refrigeration
1.00 Urban Miscellaneous Electric
0.00 Urban Other

2007 2009 2012 2015 2018 2021 2024 2027 2030
Figure 53: Urban Residential Sector Useful Energy Demands
Rural Residential Sector Useful Energy Demands

4.00
Rural Cooking
3.50
Rural Lighting
3.00
Rural Space Cooling (Fans and
Coolers)
2.50 Rural Space Heating
MTOE
2.00 Rural Water Heating

1.50 Rural Space Cooling (AC)
1.00 Rural Refrigeration
0.50 Rural Miscellaneous Electric
0.00 Rural Other

2007 2009 2012 2015 2018 2021 2024 2027 2030
Figure 54: Rural Residential Sector Useful Energy Demands
9. Transport
The transport sector energy service demands were developed for the end-use applications
listed below. There are two sub-categories of demands: passenger transport and freight
transport. Most of the passenger transport demand is expressed in million passenger
kilometers (mpkm) except for aviation, which is in PJ. Most of the freight transport demand
is expressed in million ton kilometers (mtkm) except for shipping, which is in PJ.
Passenger transport Units

• Buses mpkm
• Cars mpkm
• Two-wheelers mpkm
• Three-wheelers mpkm
• Taxis mpkm
• Minivans mpkm
• Rail Passenger mpkm
• Aviation PJ
Freight transport Units
• Trucks mtkm
• Delivery Vans mtkm
• Rail Freight mtkm
• Shipping PJ
Overall GDP growth is the principal driver for this sector, and each sub-sector is related to
the GDP growth through an energy use / GDP elasticity factor, which is defined as the
percentage change in energy consumption divided by the percentage change in GDP
growth. The Pakistan Transport Plan Study62 was the principal reference used to calculate
the elasticity factors. Table 51provides the energy use / GDP elasticities used for the
passenger transport sub-sector demand, and Figure 55 shows the resulting passenger
transport demand level and shares. The per capita passenger transport demand goes from
3500 km/yr in 2007 to 6500 km/yr in 2030 with most growth happening for cars and two-
wheelers.
Table 51: Passenger Transport Energy /GDP Elasticity Factors
Energy/GDP Elasticity Factor (Overall change in energy use relative to
Demand
GDP growth)
Buses -1.5% per annum (29% less energy growth by 2030 relative to GDP growth)
Cars No change relative to GDP
Two-wheelers -0.5% per annum (11% less energy growth by 2030 relative to GDP growth)
Three-wheelers -0.5% per annum (11% less energy growth by 2030 relative to GDP growth)
Taxis -0.5% per annum (11% less energy growth by 2030 relative to GDP growth)
Minivans -1.5% per annum (29% less energy growth by 2030 relative to GDP growth)
Rail Passenger No change relative to GDP
62
Pakistan Transport Plan Study, JICA 2006.

Projected Share of Passenger Transport

100% 1,600
90%
1,400
80%
1,200
70% Rail Passenger
Total Activity (1000 PKM)

Peercentage Shares
1,000 Minivans
60%
Taxis
50% 800 Three‐wheelers
Two‐wheelers
40%
600
Cars
30%
Buses
400
20% Total PKM
200
10%
0% ‐
2007 2009 2012 2015 2018 2021 2024 2027 2030
Figure 55: Passenger Transport Demand Projections
The Pakistan Transport Plan Study63 was also the principal reference used to calculate the
elasticity factors for freight transport. Table 52 provides the energy use / GDP elasticities
used for the freight transport sub-sector demand, and Figure 56 shows the resulting freight
transport demand level and shares. The freight transport demand increases by a factor of
2.5 between 2007 and 2030 with freight rail showing an increasing share over time.
Table 52: Freight transport Energy /GDP Elasticity Factors

Demand Energy/GDP Elasticity Factor (Overall change in energy use relative to GDP
growth)
Trucks -1.0% per annum (21% less energy growth by 2030 relative to GDP growth)
Delivery Vans -1.0% per annum (21% less energy growth by 2030 relative to GDP growth)
Rail Freight No change relative to GDP
Increase by 1.0% per annum (21% less energy growth by 2030 relative to GDP
Aviation growth) &
1.2% Autonomous Efficiency Improvement factor
Shipping -0.5% per annum (11% less energy growth by 2030 relative to GDP growth)
63
Pakistan Transport Plan Study, JICA 2006.

Projected share of Freight Transport

100% 120
90%
100
80%
Total Activity (TKM)

70%
Percentage Share
80
60% Rail Freight
50% 60 Trucks
40% Delivery Vans
40
30%
Total MTKM
20%
20
10%
0% ‐
2007 2009 2012 2015 2018 2021 2024 2027 2030
Figure 56: Freight Transport Demand Projections

VIII. Base Year Calibration

The Pak-IEM model was calibrated so that it reflects the published energy balance as
published in the Energy Year Book 2007. The goal of the calibration is to match the total
energy supply, electricity generation, and energy consumption by fuel type across the entire
energy system. This calibration is important, as it ensures that the energy system in the base
year closely represents the ground reality, as reflected in the energy statistics. The first year
of the model also needs to be properly calibrated because the base year energy service
demands derived from the energy balance provide the seed for the demand projection
drivers.
The User Guide describes the process (or mechanics) of checking the base year calibration
so that is not repeated here. Four specific metric have been checked:
• Primary energy production
• Final energy consumption
• Power sector consumption / production
• Temporal electricity demand profile.
Current calibration checks show that both primary energy production and final energy
consumption in the model base year vary less than 0.2% from what is reported in the Energy
Year Book.64 The power sector fuel consumption is also subject to checking; this varied only
slightly from the published statistics by 0.5% due to small levels of under consumption by
gas plants. Note that these calibration checking metrics (described in the User Guide) are
not exactly the same as the standard primary and final energy tables, as they have been
specifically set-up for calibration checks.
An important aspect of the electricity system that has been subject to some calibration work
to get a reasonable representation in Pak-IEM is the electricity load curve. The load curve is
calibrated using data from both PEPCO and KESC systems. To replicate this in the model,
the timing and level of all of the individual sector demands at an aggregate level must match
the reported data.
Figure 57 shows the load curves in 2006/07 for each of the three seasons, showing demand
for electricity (in PJ) in each hour. The dashed lines represent the model output and the solid
lines, the actual reported data from KESC / PEPCO. There is currently a reasonable match
in the model concerning shape of the curve and level of demand. However, some
improvements are needed to better calibrate summer night demands, and improve the
summer peak representation.
64
See worksheet Calibration in model workbook PakIEM_EnergyBalance-2007 for additional information.

14
13
12
11
Hourly PJ demand per
10
7
Intermediate (R) Intermediate
6
Summer (R) Summer
5
Winter (R) Winter
4
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24
Hours of the day
Figure 57: Reported load curve values versus model outputs
While other metrics were subject to checking, namely carbon emissions, these have not
been introduced systematically into the calibration process. This should be done as part of
future model development.

IX. AREAS FOR FURTHER MODEL DEVELOPMENT

This first version of Pak-IEM provides significant analytical capability for assessing
Pakistan’s different energy policy issues. However, like with all models, there are potential
areas for improvement, both structural and with respect to data availability. This section of
the report outlines the areas for improvement although does not definitively state what
should be the priorities; this is to be determined by the Planning Team, and their Advisory
Committee.
Cross-cutting model enhancements
• Development of model regions, to provide additional insights into regional energy
issues and policy options
• Track investments made by public / private sector in the model so that different
constraints can be put on the different financial flows
• Explore alternative objective function formulations
• Increased use of economic constraints to control technology uptake through higher
technology specific discount rates, representing barriers. These could replace the
more rigid UC share constraints used in the model.
Potential improvements by sector are described below.
Resource supply
Issue Description
The resource is included in the model but further work is required to
Tight gas resource develop extraction cost estimates, and timing of resource
availability.
The structure is set up in the model but further assessment of
CCS transport and storage
storage potential and costs is needed.
Gas network expansion costs need further review, particularly the
Gas network use of cost data provided as investment per KM (see DG GAS data
file Response of DG Gas 4-10-2010.pdf.
Gas processing technologies (cost and processing efficiency) could
Gas processing
be added if data can be obtained.
Refinery sector65
Issue Description
This is based on information provided by different refineries, relating
to future upgrades and the costs of oil transportation. An obvious
Potential further development improvement would be to differentiate crude transport costs
of existing refinery depending on the crude input mix (based on domestic region),
characterization taking care that these were not already included in the operating
costs.
65
The notes of the July Refinery Sector Task Force Meeting suggest that it would not be a priority to enhance
Refinery to multi-stage and that a single block representation suffices. At the September meeting, there was no
real support for developing API representation by crude type. At the present time, there is not the data to model
this anyway.

Power sector
Issue Description
To increase use of discrete investment, a list of plants need to be

identified for inclusion, and the permitted discrete investments. This
Discrete investments
also applies to specific investments in the resource supply sector
e.g. import gas pipelines, refineries, LNG terminals, etc.
Power plant that require additional transmission network due to

Additional transmission costs their probable location should be identified and costs estimated.
The structure has been set up in the model.
Further development of options for upgrade of existing network
could be considered. KESC has provided information on options for
T&D system upgrade
reducing losses and T&D investment costs (see file KESC's
Response 23-9-10.pdf).
Further assessment of electricity prices in different timeslices,
Electricity prices considering whether load curve data used needs to be reviewed to
ensure peak period prices.
Residential / Commercial sector

A significant need for the residential and commercial sectors is improved data to better
characterize the level and timing of fuel use types for different energy services. This is
needed for different income groups in the residential sector and sub-sectors for the
commercial sector. In addition, a better understanding of the types of technologies used is
also important.
No significant work has been done on these issues since the HESS study. A large-scale
survey to better underpin the data assumptions in the model for these sectors is seen as a
major priority.
Specific issues pertinent to the residential sector are listed below:
Issue Description
Income classes Further disaggregation of urban / rural into income classes.
Uptake of AC Information on recent trend in uptake of AC needed, and data on
future trends.
Biomass substitution Consideration of rate of switching away from biomass needed for
rural / urban households.
Future energy intensity Estimates needed of changing household energy intensity over
time.
Timing of device use Further review of current timings used in model – based on load
curve calibration.
Backup generation Percent of households w/ gensets and amount (time) of use.
Industry sector
Further engagement with the industry sector is important for the further improvement of the
model. In particular, the textile sector, one of the most important economic sectors in
Pakistan, requires additional development, with greater engagement with the sector.
Specific areas for further model development are listed below:

Issue Description
Textile sector Improved understanding of the sector composition and technology
characterization stock. In addition, further information on the evolution of the sector
could be important to capture – to understand future energy needs /
technology uptake.
Fertilizer sector development This sector only tracks different types of natural gas use (fuel vs
feedstock). If deemed important, the structure of this sector could
be further detailed
Other industry This “catch all” sector is relatively large. Further consideration
should be given to further break out of other sectors

X. REFERENCES
These references are those which are published, and do not include specific in-house
datasets supplied by different organizations. Such datasets are referenced in footnotes in
the body of the report.
ACO (2005), Pakistan Agricultural Machinery Census 2004, Statistics Division, Agriculture
Census Organization, Government of Pakistan, June 2005, Lahore, Pakistan
ADB (2010), Islamic Republic of Pakistan: Rental Power Review, Asian Development Bank,
January 2010
ADB (2009), Central Asia South Asia Regional Electricity Market Project: Techno-Economic
Feasibility Study for the Central Asia-South Asia Transmission Interconnection (CASA –
1000), Prepared by SNC-Lavalin International Inc, Project Number: 40537-01, August 2009
ADB (2009b), Sustainable Energy Efficiency Development (SEED) – Sector Investment
Plan: Approach, Estimates and Schedule, Project Number: ADB TA 7060, May 2009
ADB (2008), TA 4665 PAK Power Transmission Enhancement Project – Pakistan, British
Power International, Funded by Asian Development Bank, November 2008
AEO (2010) Assumptions to the Annual Energy Outlook 2010, EIA, 2010
Bom GJ. and van Steenbergen F (1997) Fuel efficiency and inefficiency in private tubewell
development, Energy for Sustainable Development, Volume III No. 5, January 1997
BPI (2008), TA 4665 PAK - Power Transmission Enhancement Project – Pakistan, British
Power International on behalf of Asian Development Bank & National Transmission and
Despatch Company, November 2008
EESI (2007), Hybrid buses costs and benefits, Environmental and Energy Study Institute
EIA (2010), Annual Energy Outlook 2010, Energy Information Administration (EIA),
http://www.eia.doe.gov/oiaf/aeo/
ENERCON (2003), Fuel Efficiency, Vehicular Emissions and Pollution Control in Pakistan,
Authored by Iftikhar A. Raja, Part of Fuel Efficiency in Road Transport Sector (FERTS)
Project, ENERCON, Islamabad, April 2003
ESMAP (1993), Household Energy Demand Handbook for 1991, Report under Household
Energy Strategy Study (HESS), Energy Sector Management Assistance Programme,
Prepared for Government of Pakistan under the UN Development Programme (PAK/88/036)
ESMAP (2006), Pakistan Household Use of Commercial Energy, Energy Sector
Management Assistance Programme, Report 320/06, February 2006, IBDR / World Bank,
Washington D.C.
FAO (2009), Pakistan Forestry Outlook Study, Working Paper No. APFSOS II/WP/2009/28,
Authored by Office of the Inspector General of Forests, Ministry of Environment,
Government of Pakistan, Published by Food and Agriculture Organisation of the United
Nations, Bangkok, 2009
FBS (2007), Census of Electricity Establishments (CEE) 2005-06, Government of Pakistan,
Statistics Division, Federal Bureau of Statistics, September 2007
GoP (2009) Economic Survey of Pakistan 2008-09, Finance Division, Economic Advisors
Wing, Finance Ministry, Government of Pakistan, http://www.finance.gov.pk

GoP (2008) Economic Survey of Pakistan 2007-08, Finance Division, Economic Advisors
Wing, Finance Ministry, Government of Pakistan, http://www.finance.gov.pk
HDIP (2008), Pakistan Energy Yearbook 2007, Hydrocarbon Development Institute of
Pakistan, Ministry of Petroleum and Natural Resources, Government of Pakistan, January
2008
IEA. (2009), Energy Balances of Non-OECD Countries, 2006-07, International Energy
Agency, France
IEA. (2009b), World Energy Outlook (WEO) 2009, International Energy Agency, France
IEA (2008), Energy Technology Perspectives 2008: Scenarios & Strategies to 2050,
International Energy Agency, Paris
IEA (2005), Prospects for hydrogen and fuel cells, International Energy Agency, Paris
IEA (2005b), Projected Costs of Generating Electricity: 2005 Update, International Energy
Agency, OECD/IEA, Paris
IEA (2010), Projected Costs of Generating Electricity, 2010 Edition, International Energy
Agency, Nuclear Energy Agency, Organisation for Economic Co-Operation and
Development, Paris
IEA/WBCSD (2004), IEA/SMP Model Documentation and Reference Case Projection, World
Business Council on Sustainable Development Sustainable Mobility Project (SMP), July
2004
IPCC (2006), 2006 IPCC Guidelines for National Greenhouse Gas Inventories, Prepared by
the National Greenhouse Gas Inventories Programme, Eggleston H.S., Buendia L., Miwa K.,
Ngara T., and Tanabe K. (eds). Published: IGES, Japan
JICA (2006a), A Fact Book on Pakistan Transport, JICA-NTRC publication based on PTPS
study, May 2006
JICA (2006b), Pakistan Transport Plan Study (PTPS) in the Islamic Republic of Pakistan –
Final Report, In collaboration with NTRC and on behalf of the Government of Pakistan,
Islamabad, March 2006
Memon M, Harijan K and Uqaili A (2008), Assessment of Wood Fuel Production Potential in
Pakistan, in World Renewable Energy Congress (WRECX), Editor A. Sayigh
NREL (2008), Clean Cities Factsheet: Flexible Fuel Vehicles: Providing a Renewable Fuel
Choice, Sponsored by U.S. Department of Energy Energy Efficiency and Renewable Energy
Vehicle Technologies Program, June 2008
NREL (2008), Clean Cities Factsheet: Natural Gas, Sponsored by U.S. Department of
Energy Energy Efficiency and Renewable Energy Vehicle Technologies Program, October
2008
NTDC (2008), Electricity Marketing Data (Power System Statistics), 32nd Issue, National
Transmission and Despatch Company, WAPDA, Lahore
NTRC (2008), Motor Vehicles on Road by Region 2006-07, National Transport Research
Centre, Islamabad. Also published in Economic Survey of Pakistan 2007-08 by Accountancy
(T13.4)
NTRC (2005), Vehicle Operating Costs using HDM – VOC Version 4.0, Study 270, National
Transport Research Centre, September 2005

OCAC (2008), Pakistan Oil Report 2006-07, Oil Companies Advisory Committee, Karachi,
Pakistan
OCAC (2009), Pakistan Oil Report 2007-08, Oil Companies Advisory Committee, Karachi,
Pakistan
OGRA (2008), State of the Regulated Petroleum Industry 2007-08, Oil and Gas Regulatory
Authority
Pakistan Railways (2007), Pakistan Railways 2006/07 Statistics Year Book,
http://pakrail.com/ybook.asp
PSMA (2009), Pakistan Sugar Mills Association Annual Report 2009,
http://www.psmaonline.com/psma/home.aspx
PSMA (2007), Pakistan Sugar Mills Association Annual Report 2007,
http://www.psmaonline.com/psma/home.aspx
PCFV, Cleaner motorcycles: Promoting the use of four-stroke engines, Kjaer Group /
Partnership for Clean Fuels and Vehicles (PCFV),UNEP
Platts (2008), UDI World Electric Power Plants Data Base (WEPP)
PPIB (2008), Pakistan’s Thar Coal Power Generation Potential, Private Power and
Infrastructure Board, July 2008
Qureshi A. S, M. Akhtar and A. Sarwar (2003), Effect of Electricity Pricing Policies on
Groundwater Management in Pakistan, Pakistan Journal of Water Resources, Vol.7(2) July-
December 2003
TERI (2006), National Energy Map for India, Technology Vision 2030, TERI Press, ISBN 81-
7993-099-8
TCEB (2008), Facts and Figure: Thar Coalfield, Thar Coal & Energy Board, July 2008
UK Department for Transport (2005), Economics of Bus Drivelines,
http://www.dft.gov.uk/pgr/roads/environment/research/cqvcf/economicsofbusdrivelines
UKERC (2007), UK MARKAL Documentation: Transport Sector Module, Updated June
2007, http://www.ukerc.ac.uk/support/ESMMARKALDocs08
USDA Foreign Agricultural Service (2007), Pakistan Biofuels 2007, GAIN Report Number:
PK7009, December 2007
U.S. Department of Transportation (2005), Analysis of electric drive technologies for transit
applications: Battery electric, hybrid electric and fuel cells, August 2005, Reference. FTA-
MA-26-7100-05.1
WADE (2004), Bagasse Cogeneration – Global Review and Potential, World Alliance for
Decentralised Energy, June 2004
World Bank (2003), Pakistan: Oil and Gas Sector Review, Report No. 26072-PK, Oil and
Gas Policy Division, Oil, Gas, Mining and Chemicals Department, Private Sector
Development and Infrastructure, World Bank, Washington D.C., July 2003

XI. APPENDICES
Appendix 1. Model Naming Conventions
This section describes the naming convention guidelines that are being used in the Pak-IEM
model.
1. Demand sectors
The five major demand sectors use the following names.
• Agriculture (AGR)
• Commercial (COM)
• Industrial (IND)
• Residential (RSD)
• Transportation (TRN)
The sub-sectors in each of these sectors begin with the first letter of the sector name, and
the next 2-4 characters identify the various end-use services within the sector, or in the case
of industry, the specific sub-sector.
2. Commodity Names and Descriptions
The names for the core energy carriers (NRG) and materials employed in the model are
listed in Table 53. For most commodities a root identifying the nature of the fuel (e.g., BIO,
COA, ELC, GAS, OIL), then qualified by the form or particular instance of the fuel.
Table 53: Core Names of Energy Commodities
Core Name(s) Resource
GASRAW Raw Gas
GASNGL Gas Liquids
GASASS Assoc. Gas
GASNGA Processed Natural Gas
OILCRD Oil Crude
OILGSL Gasoline
OILHSD Diesel
OILHFO Heavy Oil
OILKER Kerosene
OILLDO Lt Diesel
OILNAP Naptha
OILAVG Av. Fuel
OILLPG LPG
COAD Domestic Coal
COAI Imported Coal
RNWHYD Hydro
RNWSOL Solar
RNWWIN Wind
NUC Nuclear
ELCC Electricity Connected
ELCR Electricity Remote
HTH High-temp
BIOWDF Wood (Free)
BIOWDP Wood (Purch.)
BIODNF Dung (Free)
BIODNP Dung (Purch.)
BIOCRF Crop Residues (Free)
BIOCRP Crop Residues (Purch.)
BIOBAG Bagasse
For sector fuels, six-letter commodity names have the first three letters removed and
replaced by the sector label. Three- or four-letter names simply have the sector label added
e.g. INDCOAD. The exception is ELCC / ELCR which would become for example INDELC.

Three other types of commodities are used in Pak-IEM: DEM (demands), ENV
(environmental indicators, emissions in Pak-IEM), and MAT (materials).
The list of demand sectors (DEM commodities in TIMES) is provided below. The general rule
is that the first letter will denote the sector, while the following letters will identify the energy
service being provided or product being produced. In the residential sector, the second letter
also denotes the sector, distinguishing between urban and rural.
Table 54: Names of DEM Commodities
DEM Name Description
AOE Agriculture Other Use
ATF Agriculture Tractors - Farm Op.
ATH Agriculture Tractors - Haulage
AWP Agriculture Water pumping
CA Commercial / Other Govt - Electric Appliances
CC Commercial / Other Govt - Space Cooling
CH Commercial / Other Govt - Space Heating
CK Commercial / Other Govt - Cooking
CL Commercial / Other Govt - Lighting
CO Commercial / Other Govt - Other
CR Commercial / Other Govt - Refrigeration
CW Commercial / Other Govt - Water Heating
FTP Fertilizer
IBK Brick kilns
ICTP Cement
IIS Iron and Steel
IOT Other Industry
ISGP Sugar
ITXC Textiles Cloth production
ITXG Textiles Garment production
ITXY Textiles Yarn production
RRA Residential Rural - Space Cooling (AC)
RRC Residential Rural - Space Cooling (Fans and Coolers)
RRH Residential Rural - Space Heating
RRK Residential Rural - Cooking
RRL Residential Rural - Lighting
RRM Residential Rural - Miscellaneous Electric
RRO Residential Rural - Other
RRR Residential Rural - Refrigeration
RRW Residential Rural - Water Heating
RUA Residential Urban - Space Cooling (AC)
RUC Residential Urban - Space Cooling (Fans and Coolers)
RUH Residential Urban - Space Heating
RUK Residential Urban - Cooking
RUL Residential Urban - Lighting
RUM Residential Urban - Miscellaneous Electric
RUO Residential Urban - Other
RUR Residential Urban - Refrigeration
RUW Residential Urban - Water Heating
T2W Transport - Two-wheelers
T3W Transport - Three-wheelers
TAV Transport - Aviation
TBU Transport - Buses
TCA Transport - Cars
TMV Transport - Minivans
TRF Transport - Rail Freight
TRP Transport - Rail Passenger
TSH Transport - Shipping
TTR Transport - Trucks
TTX Transport - Taxis
TVN Transport - Vans

The emissions being tracked in Pak-IEM include CO2, NOX and SO2. All are tracked at the
sector level, and labeled, XXXCO2 e.g. INDCO2. Materials flows are still being finalized in
the model; corresponding commodity names will be listed in subsequent Pak-IEM model
reports.
3. Technology Names and Descriptions
Technology name rule definitions are described in Table 55, where the Technology type
refers to the TIMES set designation.
Table 55: Naming Conventions for Technologies
Technology type Description Naming Convention
IRE (inter-regional Resource supply technologies, • 3 letters – IMP, EXP or MIN
exchange) providing energy to a region • 6 letters – Full commodity Name
• Indicating region or other information
PRE Energy processes converting
energy from one form to another
Refineries • ‘P’ for process technology
• 3 letters – ‘REF’ for refineries
• 4 letters – Refinery name abbreviation
• 2 letters – 00 for existing, ‘N’ or 08 for new
X Processes (Dummy process • ‘X’
technologies to track sector fuels • 3 letters – 3 letter sector name
in all end-use sectors) • 3 letters – 3 letter commodity name
• 2 letters – ‘00’
ELE Electricity generation technologies Existing Tech
• ET,ER,EH to denote thermal, renewable or
CHP
• 6 letters – Full commodity Name (for thermal
plant
• 6 letters – Technology type (for renewable
plant
• 3 letters – Plant Name (abbreviated)
New tech
• ET,ER,EH to denote thermal, renewable or
CHP
• 3 letters – Technology identifier
• 3 letters – Fuel identifier
• ‘N’ to indicate new
DMD Demand devices that are a subset • Starts with single letter to denote sector
of PRE that consume energy to • 2 letters denoting energy service being
meet the demand for energy provided
services • 3 letters denoting fuel consumed
• Final letter ‘E’ to denote existing or ‘N’ to
denote new
There are exceptions to the rules in the above table, but in general they hold across the
majority of technologies.

Appendix 2. Model Templates

The templates constitute structured Excel workbooks that contain all the data embodied in
Pak-IEM, from original data sources reference material, to the various component
declarations and characterization required for a TIMES model. Each workbook is organized
along these basic lines, taken from the power sector workbook. A fuller discussion of the
workbooks can be found in the Pak-IEM Final Report: Volume III – Users' Guide.
The complete set of templates constituting the Reference scenario are listed in Table 56. In
addition a series of policy scenarios (prefixed with Scen_Pol-) complete the initial Pak-IEM
model dataset.
Table 56: Pak-IEM Reference Scenario Workbooks
Model Template Description
Base Year Templates Characterize sectors in the base year only
VT_PAK_SUP Resource Supply Sector
VT_PAK_ELC Power Generation Sector
VT_PAK_AGR Agriculture Sector
VT_PAK_COM Commercial Sector
VT_PAK_RSD Residential Sector
VT_PAK_IND Industry Sector
VT_PAK_TRN Transport Sector
SubRes Characterizes technology options for future years
SubRES_NewTechs New technology database

Model Template Description

Demands Future demands by energy service type
DEM_PIEMRef Demand projection drivers (plus supporting data)
DEM_PIEMLow Demand projection drivers (plus supporting data) – Low growth
DEM_PIEMMed Demand projection drivers (plus supporting data) – Medium growth
Dem_Alloc+Series Driver allocation to specific drivers
Scenario files For implementing scenarios, or introducing additional model
parameters that cut across different BY templates
Scen_REF-FR Demand FRs to characterize timing of demand (plus supporting data)
Scen_REF-DMD Set-up to control fuel and technology shares in end use sectors
Scen_REF-ELC Scenario to ensure power plant operation, introduce build rates and
model load shedding
Scen_SectorPrices Sector fuel prices (to undertake financial approach

Appendix 3. List of Unpublished Data Sources

In this appendix, all of the data files sent directly by data providers and used in the model are listed. As these are not published sources in the
main, they have all been provided electronically to the Energy Wing (EW) / Planning Team, categorized by sector (as shown in the table
below). These files reside in the Sources folder on the Pak-IEM FTP site established for this project.
Sector Data Provider Data File Name Date provided Sent to IRG by
Resource DGPC Supply by field to 2035 Long term Oil Projections-DGPC.xls EW

Supply
DGPC Supply by field to 2035 Long term Gas Projections-DGPC.xls EW
th
DGPC Estimate of tight gas reserve Response of DGPC 7-10-2010.pdf Letter dated 7 EW
October 2010
DG Minerals Supply by field to 2015 Long term Coal Projections.xls EW
NARC Biomass potential estimates Biomass resource-Munir.jpg MEC
Oil pipeline information
Gas Supply SSGC Pipeline cost estimates SSGC pipeline estimates 14-02-10.doc
DG Gas Pipeline utilization, based on Letter & Data from DG Gas.pdf Letter dated 26th EW
information provided by SSGC / May 09
SNGPL
DG Gas Costs of grid expansion Response of DG Gas 4-10-2010.pdf 10.10.08 Mr Latif (EW)
SSGC Monthly consumption data by sector Categ-det_Jul'06-Jun'07.xls (Industry) EW
Comm_2006-2007-1.xls (Commercial - unused

in model)
Domestic-overall-fy0708.xls (unused in model)
SLAB WISE Domestic CONSUMPTION 2007-

08.xls (unused in model)
SNGPL Monthly consumption data by sector SNGPL Consumer Data 2007 and 2008.xls EW
Slabwise Breakup of Domestic Consumers.pdf

(unused in model)
ISGS Information on timing and costs of ISGS's Response 28-9-10.pdf 10.10.08 Mr Latif (EW)
gas import pipelines
Oil Refining Attock PIEM Refinery Data_Attock.xls 09.11.02 Mr Latif (EW)

Refinery
Attok Refinery Response 17-9-2010.pdf 10.09.17 Mr Latif (EW)
Bosicor PIEM Refinery Data_Bosicor.pdf 09.11.02 Mr Latif (EW)

Refinery
National PIEM Refinery Data_National.xls 09.11.12 Mr Latif (EW)

Refinery
NRL Response 9-9-2010.pdf 10.09.09 Mr Latif (EW)
Pakistan PRL_Refinery_Data.pdf 10.09.20 Mr Latif (EW)

Refinery
PARCO CP-Reminder_letter_for_Refineries_ 10.05.24 Nauman Bhutta

Refinery (PARCO).doc
Electricity KESC Comprehensive dataset on KESC PIEM KESC data.pdf Mr Latif (EW)
Generation generation and distribution
KESC Hourly demand data for KESC Hourly_Demand_KESC.xls 10.03.24 Nauman Bhutta
system
KESC Load curve graphs (based on above Load Curve _030310_KESC.doc 10.03.24 Nauman Bhutta
file)

KESC Additional data from KESC KESC data sheet.doc 10.03.24 Nauman Bhutta
KESC KESC information on options for KESC's Response 23-9-10.pdf 10.10.08 Mr Latif (EW)
reducing losses and T&D investment
costs
PEPCO PEPCO_Consumption_Patterns.xls MEC provided XLS

data (based on
information provided by
PEPCO)
PEPCO PEPCO 24 hour and seasonal load profile MEC provided XLS
data_v03.xls data (based on
information provided by
PEPCO)
PEPCO Information on load shedding on Power Sector Load Shedding & Captive Power 09.08.19 Mr Latif (EW)
PEPCO system and captive Data.pdf (+ XLS copies)
generation plus generation
PEPCO IPP generation data Data of Existing IPPs.pdf (+ XLS copy) Mr Latif (EW)
PEPCO PEPCO information on plant retrofit, Response to Remaining Power Sector Issues 3- 10.02.03 Mr Latif (EW)
hydro potential, plant retirement, 2-10.pdf
planned and recent new build and
grid improvement
PEPCO PEPCO thermal generation data PEPCO Thermal Power Station.pdf (+ XLS Mr Latif (EW)
copy)
PEPCO Additional information regarding Planning Commission 20-03-09.xls Mr Latif (EW)

PEPCO thermal generation data
PEPCO Data on transmission losses and PEPCO Response 17-9-2010.pdf 10.09.17 Mr Latif (EW)
costs of transmission lines (in km)

PEPCO Alternative load profile data from Hourly Load Data 2002-03.xls September 2010 Mr Latif (EW)
PEPCO
Hourly Load Data 2003-04.xls
Hourly Load Data 2005-06.xls
Agriculture Independent Provision of data to validate tractor Independent Expert TRACTORS 30-03-09.pdf 09.03.30 Provided by MEC
expert stock and fuel consumption
estimates
Independent Data on stock of combines provided Agricultural & 04.08.09

expert at Task Force meeting Transport_subgroup_Summary_v03.doc
NARC Estimates of tractor and biomass NARC 29.01.10 Dr Munir (NARC)

stove costs Correspondance_Tractor&StoveCosts.doc
Industry – I&S Pak Steel Fuel consumption and production Pak Steel Data 12-2-2010.pdf 10.02.12 Mr Latif (EW)
statistics
Pak Steel Additional responses to queries PakSteel Qry Responses _032010.rar 10.03.24 Nauman Bhutta
PSRMA Response to request for information Pakistan Steel re-Rolling Mills Association.pdf Mr Latif (EW)
from EW
Magna Steel Steel sector presentation Steel Sector Presentation Final on 10-8-09.ppt MEC
Industry - ENERCON ENERCON MIS database MEC

General
Economic Planning GDP forecasts GDP Projections of Pl Commission.pdf 09.11.10 Mr Latif (EW)
Indicators Commission
Emissions GCISC Emission factors for GAINS (by Emission Factors (CO2,SO2,NOx)_GAINS.xls 10.01.27 GCISC (Kaleem
technology-fuel) for CO2, NOx, SO2 Anwar)

GCISC Emission factors for GAINS in 2005 Emissions by Source Category GAINS 2005.xls 10.02.17 GCISC (Kaleem
for CO2, NOx, SO2 Anwar)

Appendix 4. Selected Examples of Energy Planning Applications Using

MARKAL/TIMES
Developing/Transition Countries Developed Countries

“Energy Policy and Systems Analysis Project,” Second National Communication of Italy for
eight-country (3.5 year) capability building the United Nations Framework Convention on
undertaking sponsored by AusAID under Phase III Climate Change, G.C. Tosato, M. Contaldi,
of the ASEAN-Australia Economic Cooperation and D. Gaudioso, prepared by ENEA for the
Programme, http://www.epsapforum.com/. [See Ministry of the Environment, 1999. Available
Exhibit C-3.] from the Italian Ministry of the Environment,
SIAR, Via Cristoforo Colombo 44, I-00100
Roma; also www.unfccc.de.
USAID is supporting the Athens Forum process Options for a Low Carbon – Phase 2,
for Southeast Europe by promoting capacity prepared by AEA Technology Plc. for the UK
building in eight countries of the region to look at Department of Trade and Industry, 2002,
the potential for and implications of increased http://www.dti.gov.uk/energy/whitepaper/phas
access to natural gas, energy efficiency, and e2.pdf.
accelerated economic growth.
Modeling China’s energy future: A China national Final Report on DTI-DEFRA Scenarios and
MARKAL model including a large number of Sensitivities using the UK MARKAL and
renewable energy, energy efficiency, and MARKAL-Macro Energy System Models,
advanced clean energy technologies was used to 2007,
determine the impact of advanced technologies to http://www.ukerc.ac.uk/ResearchProgrammes
meet China’s future energy demand, improve /EnergySystemsandModelling/ESM.aspx.
urban air quality, limit oil and gas imports, and
reduce GHG emissions. Energy for Sustainable
Development l Volume VII No. 4 l December
2003.
Examining energy issues in South Africa EC New Energy Externalities Development

for Sustainability (NEEDS) pan-European
including:
model integrated with life-cycle analysis,
• South Africa’s first integrated energy plan.
www.needs-project.org.
http://www.dme.gov.za/pdfs/energy/planning/int
egrated_energy_plan_dec03.pdf, and
• Clean Energy and Development for South
Africa,
http://www.erc.uct.ac.za/publications/Report%201
-%20Haw-Hughes.pdf
• An Energy Model for a Low Income Rural
African Village, Stanford University PESD, June
2003, http://iis-
db.stanford.edu/viewpub.lhtml?pid=20219&cntr=
cesp
Models to Assess the Implications of the Kyoto The Third National Communication of the
Protocol on the Energy System and Economy of Republic of Latvia Under the UNFCCC,
Colombia, Cadena Angela, HEC, School of Ministry of Environment Protection and
Economy and Social Sciences, University of Regional Development, 2001,
Geneva, Switzerland, 2000. http://www.varam.gov.lv/vide/publik/Epub.htm
.
Including New and Renewable Technologies in System for the Analysis of Global Energy
Economy-level Energy Models, Asia-Pacific markets (SAGE), Model Documentation, US
Economic Cooperation, September 2002, Energy Information Administration, 2003 -
APEC#202-RE-01.011, ISBN: 0-9726293-0-0. Volume 1,
http://www.eia.doe.gov/bookshelf/docs.html.
The Future of Natural Gas vs. Coal Consumption IEA Energy Technology Perspectives–
in Beijing, Guangdong and Shanghai: An Scenarios and Strategies to 2050, 2006 and

Developing/Transition Countries Developed Countries

assessment utilizing MARKAL, BinBin Jiang, bi-annually,
Program Coordinator, http://www.iea.org/Textbase/publications/free
http://pesd.stanford.edu/news/chinagasreport/, _new_Desc.asp?PUBS_ID=1693.
September 2007.
Enhanced Economic Modeling Capacity for Northeast States for Coordinated Air Use
Management, New England MARKAL Model,
Kazakhstan http://www.sofreco.com/projets/c886/
Reports.htm, Task 6. ongoing, http://www.nescaum.org/projects/ne-
markal/index.html.
Of the undertakings mentioned above, the AusAID EPSAP is worth elaborating on a bit more
owing to their proximity to Pakistan, and in many ways similar energy system challenges. To
this end, the table below provides further evidence of the merits of MARKAL/TIMES for both
capacity building and meaningful policy analysis for Pakistan.
AusAID EPSAP Undertaking Using MARKAL/TIMES
“Country” Participating Study Description/Status

Institution
• The Trans-ASEAN Energy Network - Analysis of gas

ASEAN ASEAN Centre for
pipelines and electricity interconnections
Energy
• ASEAN Energy Market Integration
• National Energy Policy - Analysis of Options
Cambodia Ministry of Industry,
Mines and Energy
• Gas Utilization: National Gas Pipelines, Alternative Fuel Mix

Indonesia Badan Pengkajian Dan
for Power Plant and Demand Sectors
PenerapanTeknoiogi
• The Future Demand for Natural Gas in Indonesian Regions
with Particular Reference to the Use of CNG in Transport
• Future Technologies for Power Plants in Indonesian Regions
with Particular Reference to the Use of Renewable Energy
and Small Scale Coal Steam Plant
• Energy Security and Diversity.
Laos Ministry of Industry
and Handicrafts,
Electricity
• Energy Self-Sufficiency - Options and Strategies

Myanmar Ministry of Energy
• Fuel Diversification - Economic and Environmental Impact of

Malaysia PTM Malaysia Energy
Alternative Fuel Mix
Centre
• Cost and Environmental Impact of Renewable Energy
Technologies
• GHG Mitigation Options with Emphasis on Energy Efficiency
and Renewable Energy Strategies
• Impact of Natural Gas Expansion
Philippines Department of Energy
• Promoting Renewable Energy in a Restructured Electricity
Market
• Increasing Renewable Energy Utilization by Full Cost
Accounting of Electricity Supply
• Removing the Subsidy on LPG and Implementing a Policy to
Thailand National Energy Policy
Increase the Use of CNG in Transport
Office
• Fuel Options for Power Generation
• Renewable Energy

• The Strategy Orientation for Electricity Supply

Vietnam Ministry of Industry
• Analysis of Power Development Strategies in Compliance
with Environmental and Energy Security Issues
• Energy Pricing and its Implication for Energy Efficiency and
Environment
For other references to the application of MARKAL/TIMES in support of policy see the links
posted on the ETSAP website (http://www.estap.org)for the various partner activities.

Appendix 5: Justification for Selection of the MARKAL/TIMES Modeling Platform

There are predominantly four models in use around the world today in the realm of general
integrated energy planning models, and many others that have come and gone. There are
also many models that are not full-sector, integrated energy models, such as Wien
Automatic System Planning (WASP) package and the Integrated Planning Model (IPM), and
others that model only the electric sector. The following four major frameworks provide full-
sector, integrated modeling capability:
♦ MARKAL/TIMES, developed and maintained under the auspices of the IEA-ETSAP
consortium
♦ LEAP, developed and maintained by the Stockholm Environment Institute (SEI,
Boston)
♦ ENPEP, developed and maintained by Argonne National Laboratory
♦ MESSAGE, developed by International Institute for Applied Systems Analysis (IIASA)
and now maintained by the International Atomic Energy Agency.
Of these, none has the longevity or breadth of usage on an ongoing basis of
MARKAL/TIMES. The table below provides a summary of the characteristics and
capabilities of each of the modeling frameworks. The table is an updated version of an
assessment prepared by the Stockholm Environment Institute for the United Nations
Framework Convention on Climate Change (UNFCCC)66.
Overview of Integrated Energy Planning Models
CHARACTERIST
LEAP MESSAGE EPNEP (BALANCE) MARKAL/TIMES
IC
Developer / Stockholm Environment
IIASA / IAEA Argonne / IAEA IEA / ETSAP
Support Group Institute
www.energycommunity. www.iiasa.org /
Home page www.dis.anl.gov www.etsap.org
org www.iaea.org
Methodology - Accounting & - Optimization
Equilibrium simulation Optimization
– Model type spreadsheet–No - Linear
Iteration Linear programming
– Solution algorithm optimization solver programming
– Foresight Myopic Perfect or myopic
- Not applicable - Perfect
Simulate the response
Maximize consumer/
Minimize total of consumers and
Adds flows through a producer surplus while
Solution goal discounted producers to changes
rigid network minimizing overall total
system costs in energy prices and
system cost
demand levels
Data
Medium Medium-high High Medium-high
requirements
Global models from the
Technology Energy CO2DB with
IEA, EIA, and ETSAP
Database TED with ranged values for
Default data provide a repository of
costs, performance and technologies, and IPCC Emission factors
included existing data for
emission factors (inc. IPCC emissions
technologies and
IPCC factors). factors
emissions
User Controlled
User controlled Up to 75 years. Variable length time
Time Horizon User Controlled
Annual results Annual results periods spanning any
number of years
Other model None MACRO version Links with MAED Integrated MACRO
66 UNFCCC: Mitigation Methods and Tools in the Energy Sector (ftp://forums.seib.org/UNFCCC, Module
5.1.PPT).

CHARACTERIST
IC
features allows for demand services non-linear
coupling with the projection module, plus programming version
economy, WASP power allows for coupling with
separate non- expansion module, and the economy, without
linear Impacts, requiring iteration
programming additional iteration
module requiring
iteration
Integrated “smart”
Excel input workbooks /
ASCII tables and allows full
Manual data input /
Manual data input / manual input / customization of
Data / Results analysis module
flexible reporting tables standard set of analysis tables, and
Handling supporting reporting
and graphs results tables and intelligently links to
tables and graphs
graphs Excel to automatically
update presentation
tables/graphs
Policies can be
Policies cannot be Policies can be Policies cannot be introduced by means of
readily represented in tried by means of readily represented in flexible user-defined
the model; rather the constraints in the the model; rather the constraints in the form
analyst must adjust form of: analyst must adjust of:
assertions as to how the • emissions assertions as to how • emissions targets (on
Representation of
system will evolve over targets on the the system will evolve plants types, sectors,
policies
time and review the overall system over time and review system)
results, tweaking the • shares for the results, tweaking • energy security goals
assumptions until the renewable the assumptions until • shares for renewable
desired result is energy the desired result is energy
obtained. obtained. • imposing efficiency
standards
Expertise required Low High High Medium
Level of effort
Low-Medium High High Medium-high
required
Medium, owing to its
powerful user interface
How Intuitive?
Low, owing to its Low, owing to its bulky with embedded
(matching High, owing to its flexible
very poor user nature and complex modeling assistance
analyst’s mental graphical user interface
interface user interface feature, as well as its
model)
dynamic linkage with
Excel
Reporting
Advanced Basic Basic Advanced
capabilities
Data
management Advanced Basic Basic Advanced
capabilities
Windows, model
Software Windows, source code,
Windows, executable Windows, executable
requirements executable GAMS/solver, user
interface executable
Free to NGO, $8,500–$15,000
Available free for Free to NGO,
government and (including GAMS,
Software cost NPT states government and
researchers in non- solver & VEDA
through the IAEA researchers
OECD countries. interface)
On request: 5
Typical training days/$5,000 No charge for 5 days 5 days
required & cost Also regular international NPT states $10,000 $10,000
workshops
Phone, email,
Phone, email or web Phone or email
Technical support free limited Phone or email
forum $500–$2,500 for one
& cost support for NPT $10,000 for 80 hours
Free limited support year
states
Reference Manual & training Manual provided Manual available to Manual available free

CHARACTERIST
IC
materials materials free on web with the training registered users on website
site
English, French,
Languages Spanish, Portuguese, English English English, customizable
Chinese
The following additional observations are provided to supplement the information presented
above.
♦ LEAP has a great, intuitive user interface. But since it is only an accounting
framework, the analyst must provide “the Answer” to explore alternate futures. That
is, at each point in the energy system the user must assign the splits (market
allocation of technology choices) which then control the flow of energy through the
network. Thus, there is no way to inform the model of a policy or goal and ask it to
reconfigure appropriately. The user has to decide, and then see what the outcome is,
and iterate until the model meets the criteria sought. So it can’t be used to
economically optimize across policy goals. Thus, while LEAP is an excellent tool for
learning the principles of the RES, it is not suited for policy analysis.
♦ ENPEP is a very complex modeling framework. As a simulation model it is even
more data-intensive than MARKAL/TIMES, requiring elasticities at every node of the
network. The complete package involves several models that must be run in
sequence, often requiring iteration between the modules. It views the energy sector
as consisting of autonomous energy producers and consumers, each optimizing
individual objectives, as opposed to the MARKAL/TIMES model which looks to
optimize the entire energy system from the point of view of a central planning
authority that has the overall goal of maximizing social welfare. The model does not
provide an easy way to formulate alternate futures and evaluate policies. It has a
bulky user interface that is rather complex due to the various different model
components. The history with this modeling framework shows a correspondingly high
“drop” rate, even after extensive training. Thus ENPEP is comprehensive, but very
difficult to maintain and use, and therefore not a viable long-term framework in which
to invest.
♦ MESSAGE is a least-cost optimization framework (a cousin of MARKAL/TIMES)
developed at the International Institute for Applied Systems Analysis (IIASA) for its
own use. It was initially never intended that MESSAGE would be used by others, and
it has a very weak user interface and a poor “outreach” track record. Also,
MESSAGE uses a general purpose solver that can take quite a lot of time (hours) to
solve a large linear program.
♦ MARKAL/TIMES has been developed over the past 30 years under the auspices of
ETSAP, which has a mandate to continually develop, support, enhance, and
encourage dissemination to others. Thus, the MARKAL/TIMES modeling framework
has always been intended to enable a wide range of users to employ least-cost
optimization as an integral part of their planning process. The modeling tools and
techniques have been continually developed to improve the process of building,
maintaining, and applying what is a rather sophisticated tool. The use of this
framework in more than 60 countries and 200 institutions attests to the success of
the methodology and approach, and the relevant insights gained from the application
of the system. A typical national model solves in seconds to a couple of minutes. In
addition, the input and output can be implicitly linked to Excel workbooks. This is a
huge advantage in terms of setting up new scenarios and extracting results in “report

ready” format. MESSAGE, LEAP, and ENPEP all require “export” of results which
then require further work before going into a report.
In summary, the most obvious difference between LEAP and MARKAL/TIMES is that the
former is an accounting framework and the latter an optimization model. As such, LEAP
requires the analyst to decide the future exogenously (e.g., the market shares of each
technology/commodity), while MARKAL/TIMES does so based upon life-cycle economics.
For ENPEP this is accomplished by means of an extensive array of elasticities (often very
subjective) and user preferences (again, subjective). For different future scenarios (e.g.,
emission reduction levels), LEAP and ENPEP require an iterative approach where the
analyst “estimates” the anticipated changes in the future energy system, while
MARKAL/TIMES accepts the goal as input and reconfigures the energy system to meet the
stated goal. An important result of this is that MARKAL/TIMES produces the marginal cost of
reduction (e.g., value of carbon rights) as a standard output, as well as the relative ranking of
the various technologies and the prices for energy. In addition, MARKAL/TIMES has facilities
to allow the demand service levels to respond to prices, whereas in LEAP these are strictly
exogenous and ENPEP requires rerunning the MAED module. Note that, if so desired,
MARKAL can be set up as a simulation tool, accepting the market shares as input, but
neither LEAP nor ENPEP can employ optimization, and in such mode the other benefits of
MARKAL/TIMES would still hold (e.g., marginal reduction cost, endogenous demands).
All four models – each with its inherent strengths and limitations – can examine the entire
energy system. But the robustness and power of the user interface is a crucial aspect for
successful transfer of the capability for the Planning Team to use the model on an ongoing
basis. Only LEAP and MARKAL/TIMES can meet this requirement, with the former falling
short of being able to serve as an integrated energy system planning and policy analysis
framework.
IRG has extensive experience with projects developing and transferring MARKAL/TIMES
models around the world. Selected examples of notable MARKAL/TIMES applications are
noted in the Appendix 4. It is this experience and the resulting tools, techniques, and
approach that distinguish this integrated energy modeling framework more than anything
else.
All four of these frameworks require the current energy system and future alternatives to be
described by means of the input data. Traditionally, the approach most often taken is to have
the eventual user (in this case the Planning Commission) gather the underlying data and
construct the model. However, IRG has come to the conclusion that a more efficient and
effective approach is to have local experts guided through the data collection process and to
have the underlying model constructed by experienced modelers (in concert with the local
experts). With a viable relevant model in hand, the IRG experts then embark on the capacity
building of the Planning Team. This ensures that the starting point reflects current best
practices, and that the training of the Planning Team can focus on understanding and using
a quality Pakistan integrated planning model.

Appendix 6. Job description for role of Lead Energy Modeler and Analyst
Requirements:
Education: Masters of Science Degree or Bachelors Degree with
Computer Modeling Experience
Prior Work Experience: Minimum of 5-10 years of relevant experience preferably
in energy sector evaluation, computer model
development, data collection and evaluation. Familiarity
with optimization models a plus.
Software Experience: MicroSoft Office (particularly Excel), Microsoft Outlook,
computer software and programming language skills is
beneficial.
Travel: Domestic travel to Karachi and Lahore likely
Job Summary:
This position provides lead analytical and data management expertise to the Chief of the Energy Wing
in conjunction with the Pakistan Integrated Energy Model (PIEM). Responsibilities include ongoing
stewardship and application of the PIEM on behalf of the Planning Commission of the Government of
Pakistan (GOP). The incumbent will manage Energy Wing staff working on the PIEM, coordinate with
the PIEM Support Institution teams, provide data and interact directly with the data source institutions
(PEPCO, NTRC, WAPDA, ENERCON, etc.). The candidate will be working with an international team
of experts, providing data management and coordination assistance to a multi-national team of
experts. General project support skills and a willingness to travel to a variety of GOP and key energy
sector entities are required. Candidate must be capable of working independently or with minimal
supervision on technical and energy modeling related tasks. Data collection, data management, and
energy modeling are all integral to this position.
Duties and Responsibilities:

Project Management and Support Tasks:
• Support/ensure timely delivery of technical outputs (i.e. contract deliverables and data).
• Participate in energy modeling and capacity building exercises within the Energy Wing.
• Develop in-depth MARKAL/TIMES energy modeling expertise.
• Support data collection and model development.
• Serve as a central point of contact for a multi-national team of energy modeling experts
on technical issues. Assisting with project management tasks as assigned.
• Provide guidance to project personnel, consolidate and manage large databases, support
data collection, and serve as the lead energy modeling expert for the Energy Wing.
• Oversee the activities of the Energy Wing PIEM staff, and coordinate activities with the
various PIEM Support Institutions (PAEC, UETT, PIAS, etc.)
• Assist in coordinating energy modeling workshops, capacity building exercises, and other
managerial tasks as assigned.
Technical Requirements:
• Must have a science or engineering background.

• Level of comfort working with databases and computer models. Knowledge of Operations
Research a major plus.
• Provide technical input, data management, and analysis.
• Demonstrated ability to manage and coordinate support staff and other institutions in a
collaborative manner.
• Ability to communicate effectively by preparing briefing material and reports relating to the
PIEM analyses being undertaken.


PakIEM - Model Design Report

Uploaded by

Copyright:

Available Formats

PakIEM - Model Design Report

Uploaded by

Document Information

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

PakIEM - Model Design Report

Uploaded by

Copyright:

Available Formats

Pakistan Integrated Energy Model

ADB TA-4982 PAK

Final Report Volume I

Asian Development Bank

Ministry of Planning and Development

Pak-IEM Final Report Volume I – Model Design Page ii

Pak-IEM Final Report Volume I – Model Design Page iii

Pak-IEM Final Report Volume I – Model Design Page iv

E. Transmission and Distribution System ...................................................................... 45

Pak-IEM Final Report Volume I – Model Design Page v

Figure 1: TA Organizational Structure .................................................................................... 2

Pak-IEM Final Report Volume I – Model Design Page vi

Pak-IEM Final Report Volume I – Model Design Page vii

Table 6: Prices of Energy Commodities in 2006/07 .............................................................. 25

Pak-IEM Final Report Volume I – Model Design Page viii

Pak-IEM Final Report Volume I – Model Design Page ix

Pak-IEM Final Report Volume I – Model Design Page i

LPG Liquefied Petroleum Gas

Pak-IEM Final Report Volume I – Model Design Page ii

Pak-IEM Final Report Volume I – Model Design Page 1

Pakistan Planning Advisory Committee

Planning Team Advisory Committee

Host Institutions University Research

Figure 1: TA Organizational Structure

Planning Team Composition

Figure 2: Institutional Composition of the Planning Team

B. MARKAL/TIMES Modeling Framework

Pak-IEM Final Report Volume I – Model Design Page 2

industrial process heat, lighting, passenger kilometers traveled), subject to physical

PRIMARY ENERGY CONVERSION END-USE DEMAND FOR

(PRIMARY ENERGY) (FINAL ENERGY) (USEFUL ENERGY)

Renewables, e.g. Fuel Processing Industry, e.g. Industry, e.g.

Figure 3: MARKAL/TIMES Model Overview

The underlying components of a MARKAL/TIMES model include:

Pak-IEM Final Report Volume I – Model Design Page 3

demand for energy services (e.g., passenger transportation, lighting/heating/air conditioning

C. A Framework to Institutionalize Energy Modeling Capability

Pak-IEM Final Report Volume I – Model Design Page 4

Pak-IEM Final Report Volume I – Model Design Page 5

international MARKAL/TIMES community that can be tapped to provide modeling

3. Challenges to Sustaining Integrated Energy Analysis

Pak-IEM Final Report Volume I – Model Design Page 6

4. Conceptual Institutional Structure

Pak-IEM Final Report Volume I – Model Design Page 7

Pak-IEM Final Report Volume I – Model Design Page 8

Pakistan Planning Pak-IEM Advisory

Planning Unit Advisory Task Force

Research Centers &

Figure 4. Conceptual Institutional Structure for the Pak-IEM Modeling Capability

D. Model Design Overview

Pak-IEM Final Report Volume I – Model Design Page 9

Energy Model Horizon & Demand

Power Industry Transport

Figure 5: Structure of the Pak-IEM Data Inputs

E. Scope and Structure of the Report

Pak-IEM Final Report Volume I – Model Design Page 10

Pak-IEM Final Report Volume I – Model Design Page 11

II. OVERVIEW OF Pak-IEM DATA REQUIREMENTS

Energy service demands