Techno-Economic Feasibility of Grid-Independent Residential Roof-Top Solar PV Systems
Techno-Economic Feasibility of Grid-Independent Residential Roof-Top Solar PV Systems
Techno-Economic Feasibility of Grid-Independent Residential Roof-Top Solar PV Systems
a
Centre for Environmental Policy, Imperial College, London, United Kingdom
b
Centre for Process Systems Engineering, Imperial College, London, United Kingdom
Keywords: Oman is a country characterised by high solar availability, yet very little electricity is produced using solar
Solar energy. As the residential sector is the largest consumer of electricity in Oman, we develop a novel approach,
Photovoltaics using houses in Muscat as a case study, to assess the potential of implementing roof-top solar PV/battery
Battery technologies, that operate without recourse to the electricity grid. Such systems target the complete dec-
Energy
arbonisation of electricity demand per household and are defined in this study as grid-independent systems. The
Microgeneration
approach adopted starts with a technical assessment of grid-independent systems that evaluates the character-
istics of the solar panel and the battery facility required to provide grid-independence. This is then compared to a
similar grid-connected system and any techno-economic targets necessary to enhance the feasibility of re-
sidential roof-top PV systems in Muscat are identified. Such an analysis was achieved through developing a
detailed techno-economic mathematical model describing four sub-systems; the solar panel DC source, the grid-
independent sub-system, the grid-connected sub-system and the economic sub-system. The model was im-
plemented in gPROMS and uses real hourly weather and climate conditions matched with real demand data,
over a simulated period of 20 years. The results indicate that, in the context of the system studied, grid-in-
dependent PV systems are not feasible. However, combined with a sufficiently high electricity price, grid-in-
dependent systems can become economically feasible only with significant reductions in battery costs (> 90%
reductions).
⁎
Corresponding author at: Centre for Environmental Policy, Imperial College, London, United Kingdom.
E-mail address: niall@imperial.ac.uk (N. Mac Dowell).
https://doi.org/10.1016/j.enconman.2018.10.021
Received 20 August 2018; Received in revised form 7 October 2018; Accepted 8 October 2018
0196-8904/ © 2018 Elsevier Ltd. All rights reserved.
J. Al-Saqlawi et al. Energy Conversion and Management 178 (2018) 322–334
maximising the daily operational savings that accrue to customers. This 23.45° and the rotation of the earth around the sun. Therefore, is
was explored for a range of financial incentives offered for PV uptake determined by the day of the year (d) and can be described by Eq. (1)
where it was concluded that the use of battery technologies is ad- below [13]:
vantageous in terms of maintaining grid voltages. However, the feasi- 360
bility of the proposed PV/battery system was not discussed. = sin 1 (sin(23.45°)sin( (d 81)))
365 (1)
Given that the cost of electrical energy storage systems plays a pi-
votal role in future low-carbon energy systems, Schmidt et al. [11] The at solar noon is the maximum and can be determined by Eq. (2).
constructed experience curves to project future prices for eleven elec- This maximum is used in simple PV system design and is used for this
trical energy storage technologies, including batteries. They found that model [13].
regardless of technology, although capital costs currently range from 90 ( ), for locations in the Northern hemisphere
150 to 2000 USD/kWh, they are on a trajectory towards 175 ± 25 USD/ =
90 + ( ), for locations in the Southern hemisphere (2)
kWh for battery packs once 1 TWh of capacity is installed for each
technology. Although this price range was confirmed to be feasible by Therefore, given GSI, GSIT can be calculated [13].
performing a bottom-up assessment of material and production costs,
GSIsin( + )
there was no discussion on how these projected costs would impact the GSIT =
feasibility of roof-top PV/systems.
sin( ) (3)
Since complete independence from the grid using roof-top PV/bat- Although Eq. (3) is underestimating GSIT in terms of not accounting for
tery systems represents a case whereby the electricity demand per the increase from the diffuse and reflected irradiance, this is assumed to
household is completely decarbonised while any excess power pro- be acceptable given that this model uses at its maximum at solar
duced is stored by a battery instead of being exported to the grid noon, as described by Eq. (2), and therefore is already overestimating
thereby eliminating issues emerging from connecting to the grid such as GSIT.
reverse power flows, this study develops a novel step by step approach Once GSIT is calculated, the power output of a PV module (Pm) can
on grid-independent systems. This is achieved by developing a model in be calculated using:
gPROMS [12] and using Muscat, Oman as a case study, to explore the
following gaps that exist in assessing residential roof-top PV systems,
Pm = m GSIT Ar [1 n (PVD)] (4)
mainly PV/battery systems: where Ar is the roof area, is the proportion of the roof area covered
with PV modules and m is the efficiency of the module usually ap-
• Including physical sizing to the overall sizing analysis of PV/battery proximated as [15]:
systems thereby allowing for the assessment of the practicality of
such systems. m = ref [1 ref (Tc Tref )] (5)
• Assessing the impact of varying technical parameters specific to the where the values of the reference efficiency ( ref ), reference tempera-
individual PV and battery technologies on PV/battery system fea- ture (Tref) and the temperature coefficient ( ref ) are normally given by
sibility. the PV manufacturer. Given that cell temperature (Tc) is one of the most
• Setting economic targets necessary to make residential roof-top PV important parameters used in assessing the performance of PV systems
systems economically feasible. and their power output, several cell temperature prediction models
• Evaluating the impact of demand size and demand profile on PV/ have been developed. After a review of several models [14,16–22], the
battery system size. nominal operating cell temperature (NOCT) model was chosen for this
• Assessing the impact of varying the installation date on PV/battery study owing to its simplicity, adequacy of its predicted temperatures for
system size and start-up conditions. PV applications and wider availability of input data. In the NOCT
• Understanding how PV/battery system size evolves over time due to model, the cell temperature is calculated as [14,16–18]:
factors such as degradation and variations in demand.
NOCT 20
Tc = TA + GSIT
800 (6)
This step by step approach starts with a technical analysis of grid-
independent systems and ends with an economic analysis where the Additionally, since a study by Jordan and Kurtz [23] of approximately
economic feasibility of a grid-independent system is compared to a si- 2000 PV panels has shown that PV panels degrade by an annual rate,
milar grid-connected system. The model developed for this study could PVD (%), the power output of the PV module is dependent on the age of
be divided into four sub-models presented in Section 2 below. These the system (n). Given that a linear degradation model provides a more
sub-models are applicable to any location and various PV and battery conservative estimate than an exponential model [24–26], assessing PV
technologies. Section 3 describes the input data used for this specific panels that degrades linearly was considered more suitable for this
case study of Muscat. The results obtained are presented in Sections 4 study since a conservative estimate of degradation incorporates the
and 5. Finally, these results are analysed in Section 6 where conclusions worst case scenario.
are drawn. Finally, the total energy output of the PV module (Em) in kWh is
simply the integral of Pm between two periods of time, t. This could
2. Model design and construction then be compared to the total energy demand in kWh (ED), calculated as
the integral of PD between two periods of time.
2.1. DC source
2.2. Grid-independent PV system
The core element of a roof-top PV system is the direct current (DC)
power output from the PV module (Pm), also known as the DC source. Fig. 1(a) shows the schematic of the grid-independent PV system
Solar radiation is primarily measured as global solar irradiance on a model developed for this study [7,27–29]. In this system, the energy
horizontal plane (GSI). However, Pm is directly dependent on the solar produced by the DC source is consumed directly by converting it into
radiation incident on it (GSIT) [13,14]. GSIT is determined by geometric AC electricity using an inverter with an efficiency inv and an absolute
parameters such as the tilt angle of the PV from the horizontal surface lifetime Ninv. Any excess energy is stored in the battery as DC power and
( ) and the orientation towards the sun, the latter comprising the de- any shortfall in energy is provided for by the battery. The battery is
clination angle ( ), the elevation angle ( ) and the latitude ( ) [13]. protected from overcharge and deep discharge using a charge controller
varies seasonally due to the tilt of the earth on its axis of rotation by with an efficiency cc [30]. Therefore, within the PV system depicted in
323
J. Al-Saqlawi et al. Energy Conversion and Management 178 (2018) 322–334
Fig. 1(a), two situations could exist; a surplus in which the PV system given by:
produces more energy than household demand (PD) and a shortfall in
(P+)
which the PV system produces less energy than household demand. ID0 =
inv VB (14)
Whether the system is in surplus or shortfall is determined by the fol- cc
lowing equation: The actual discharge current, ID, accounting for the state of the battery,
is limited by ID IDmax and SOC > SOCmin. Therefore, if the battery is
P+ =( cc inv ) Pm PD (7)
empty (SOC = SOCmin), the shortfall is met by the grid. Otherwise, if
+ +
where P > 0 during a surplus and P < 0 during a shortfall [31]. If SOC > SOCmin, it is prioritised over the grid to meet the shortfall by
P+ = 0, the battery does not charge or discharge and all power pro- discharging. In such cases, if ID0 < IDmax, then ID = ID0. On the other
duced is consumed immediately. hand, if ID0 IDmax, then ID = IDmax. Given VB, the battery discharge
power, PBD, is given by:
2.2.1. Surplus (P+ > 0)
PBD = ID VB (15)
Initially, during a surplus, the power generated is used to satisfy
demand (Psat). This can be expressed by: The amount of charge discharged by the battery during a time period t
or the battery discharge capacity, CD, is given by:
Psat = PD (8)
D
The remaining power is then used to charge the battery within its
CD = (ID t ) batt (16)
limits. If the battery is not already fully charged, the battery will receive where D
is the battery discharge efficiency. Since the grid is pre-
batt
a charge current limited by the maximum charge current, ICmax. The sumed available to ensure that the demand is always met when either
value of the available charge current, IC0, at any time (i.e. ignoring the SOC = SOCmin or ID0 > IDmax, the power being met by the grid, Pgrid, is
state of the battery) is given by: given by:
P+ Pgrid = ( P +) (17)
IC0 = inv PBD
cc inv VB (9)
The actual charge current, IC, that accounts for the state of the battery is 2.2.3. Battery state of charge (SOC)
limited by IC ICmax as well as SOC < SOCmax. Therefore, if the battery Tracking the battery’s state of charge at any given time, SOC(t), can
is full (SOC = SOCmax) the battery cannot be charged (IC = 0). be calculated through tracking the change in battery capacity, C,
Otherwise, if SOC < SOCmax and IC0 ICmax, then IC = IC0. If, on the where:
other hand, SOC < SOCmax and I C0 > ICmax, then IC = ICmax. Given the
C = CC CD (18)
battery voltage, VB, the battery charge power, PBC, is given by:
PBC = IC VB (10) Negative values of C result during discharge and positive values
during charge. The battery charge, C(t), at any time t, is therefore:
The amount of charge gained by the battery during a time period t or
the battery charge capacity, CC, is given by: C (t ) = [C (t 1) + C ][1 BSD h] (19)
C where C(t − 1) is the charge of the battery during the previous time
CC = (IC t ) batt (11)
period and BSDh is the battery’s hourly self-discharge rate. Therefore,
where batt
C
is the battery charge efficiency. During a surplus, it is pos- the battery’s state of charge at any given time, SOC(t), can be calculated
sible to export power to the grid if either; SOC = SOCmax or using:
IC0 > ICmax. The power exported as a result of these conditions, Pexp is
C (t )
given by: SOC (t ) =
C (20)
Pexp = P + PBC (12)
This will be limited by the expression SOCmin SOC(t) SOCmax.
324
J. Al-Saqlawi et al. Energy Conversion and Management 178 (2018) 322–334
surplus, some of the power output from the inverter is used to satisfy
the demand (Psat = PD) while the surplus power is exported to the grid
(Pexp = Pinv − PD). On the other hand, during a shortfall the power
output from the inverter is used to satisfy some of the demand
(Psat = Pinv) while additional power is imported from the grid
(Pgrid = PD − Pinv) to satisfy the shortfall. Finally, if Pinv = PD, all power
produced is consumed immediately.
3.1. Hourly input data described by Jenkins et al. [31] for lead-acid batteries as shown in Eq.
(22) below.
Half-hourly demand data for 14 different villas ranging in floor
100
areas between 218 m2 and 858.3 m2, for the year 2013 was collected
C,D
batt = CVB
13.3ln I + 59.8 (22)
from [39] for Muscat, Oman. Given the relative paucity of villa data C,D
325
J. Al-Saqlawi et al. Energy Conversion and Management 178 (2018) 322–334
Fig. 4. Graphs showing if solar PV supply in Muscat can exceed demand at the median PVD of 0.7% where (a) shows the change in annual Em/m2 and ED/m2 with
while (b) shows the monthly Em/m2 and ED/m2 at = 33%. The error bars show how Em/m2 would vary as PVD changes between 0% and 2.5% [23].
326
J. Al-Saqlawi et al. Energy Conversion and Management 178 (2018) 322–334
Fig. 5. Graphs showing the impact of changing PVD and BSD on the grid-independent system characteristics in Muscat for year n = N − 1 where (a) shows how GI is
impacted and (b) shows how CGI is impacted.
Fig. 7. Image depicting the battery specifications needed to achieve grid-independent status in Muscat at four different installation dates of year n = N − 1; January,
March, July and September.
327
J. Al-Saqlawi et al. Energy Conversion and Management 178 (2018) 322–334
Fig. 8. Graph showing the minimum CGI needed for the system in Muscat to be grid-independent at four different points within its 20 year lifetime (i.e., at n = 0,
n = 5, n = 10 and n = 15). The change in SOC of the batteries throughout the 20 years is also shown by the blue line.
every year, the PV panel produces less power causing the battery to cost of batteries is expected to decrease exponentially over the years at
discharge slightly more power in order to satisfy the demand. When an annual rate ranging between 8 and 14% [48,49], the cost of the
PVD = 0%, this decline in SOCinitial does not exist. Since increased batteries was assumed to reduce annually by 11%.
charge/discharge ranges put a strain on battery lifetime [29], it can be As can be seen from Fig. 9, it is clear that both grid-connected and
concluded that increased PVD impacts the lifetime of the battery. Ad- grid-independent systems in Muscat are not feasible since the NPV re-
ditionally, given the fact that the PV panels are less degraded in the mains negative throughout the system lifetime. This can be explained
previous years implies that for GI = 40%, excess energy is produced. by comparing Oman’s residential electricity price with the rest of the
This combination of factors indicate that a 165 kAh battery is not big world. Owing primarily to extensive government subsidy [50], at 2.5
enough to achieve grid-independent status and therefore a bigger bat- U.S. cents per kWh, the electricity price in Oman is very low, where it is
tery is needed (CGI = 202 kAh) in order to be as independent from the 15 times lower than that of Germany and 5 times lower than the United
grid as possible. States [51]. This low electricity price indicates that the benefits that
come from PV investments are insufficient to out-weigh the costs
4.1.6. Twenty years of grid independence therefore making the investment infeasible. Secondly, grid-independent
The analysis in Section 4.1.5 looked at sizing the grid-independent systems are approximately 40 times less feasible than grid-connected
system for the final five years of the system lifetime. Given that the grid- systems. This is largely driven by the high battery costs, which form
independent system is assumed to have a lifetime, N, of 20 years (Table over 97% of the total system cost. Therefore, increasing the price of
C.5), the batteries should be expected to be replaced four times (i.e., at electricity to 5.94 times the current tariff, at an average price similar to
n = 0, n = 5, n = 10 and n = 15). Assuming GI remains at 40% that of United States [51] and Oman’s cost-reflective tariff [1], makes
throughout the system lifetime, Pm will reduce annually due to PVD. grid-connected systems feasible with a pay-back period of ten years. On
Therefore the battery size, CGI, for the first five years will be higher than the other hand, given that grid-independent systems have significantly
that for the final five years due to the excess power produced. This is larger up-front costs, the increase in electricity price needed to break-
shown in Fig. 8 which shows the minimum battery sizes needed for the even is much higher than that for grid-connected systems, with a re-
system to be grid-independent at four different points within its life- quired increase of 338.70 times the current tariff in order to achieve a
time. This chart also shows how the state of charge (SOC) of the bat- pay-back period of ten years. This tariff is over 21 times higher than
teries changes in 20 years depicting clearly how the battery discharge most countries in the world indicating that an increase in electricity
range increases annually due to PVD. price alone is not enough to make grid-independent systems feasible in
Muscat [50].
5. Economic results
5.2. Grid-connected system vs grid-independent system-technology cost
5.1. Grid-connected system vs grid-independent system-increasing electricity reduction
prices
Fig. 10 shows the combined impact of increasing electricity prices
Fig. 9 compares the NPV of a grid-connected system in Muscat with and reducing the unit cost of the PV panel (UCPV) on the internal rate of
that of a grid-independent system and shows the impact of increasing return (IRR) of grid-connected systems (Fig. 10(a)) and grid-in-
the price of electricity on the economic feasibility of both grid-con- dependent system (Fig. 10(b)) in Muscat. The international electricity
nected and grid-independent systems. It must be noted that these results prices corresponding to the price increase factors are marked [50].
are for a system which is assumed to start running in January, mid-day. Additionally, the IRR needed for these investments to become feasible
The battery sizes for the grid-independent system at PVD = 0.7% and for a private investor at IRR = 13% (red-line), a pension fund at
BSD = 5% are those summarised in Fig. 8. Furthermore, given that the IRR = 8% (black-line) and the government at IRR = 4% (blue-line) are
328
J. Al-Saqlawi et al. Energy Conversion and Management 178 (2018) 322–334
Fig. 9. Graphs showing the impact of increasing electricity prices in Muscat on the NPV where the dashed lines show the price increases needed to achieve a pay-back
period of ten years for (a) grid-connected systems and (b) grid-independent systems.
Fig. 10. Charts showing the combined impact of increasing electricity prices, reducing the unit cost of the PV panel (UCPV) and reducing the unit cost of the battery
(UCBatt) on the internal rate of return (IRR) of (a) grid-connected systems and (b) grid-independent systems in Muscat. The IRR needed for these investments to
become feasible for a private investor at IRR = 13% (red-line), a pension fund at IRR = 8% (black-line) and the government at IRR = 4% (blue-line) are marked.
marked. These IRR values were based on previous work [36]. explored by Schmidt et al. [11], where battery packs are forecast to
From Fig. 10(a), it is clear that in addition to the phasing out of drop to 175 ± 25 USD/kWh once 1 TWh of capacity is installed, it is
subsidies and therefore increasing electricity prices, reducing the cost of clear that increased production and economies of scale are not enough
the PV plays a significant role in increasing the feasibility of grid-con- to achieve such price reductions and that further research in technology
nected systems in Muscat. Although at the current tariff large reduc- innovation is necessary.
tions in UCPV are needed to make grid-connected systems feasible,
combined with increasing electricity prices, the reduction in UCPV
doesn’t necessarily need to be large; where a 20% reduction makes grid- 6. Discussion and conclusions
connected systems attractive to private investors when the electricity
price in Muscat is lower than the USA average. Given that PV costs are When compared to other countries in the world, such as the USA at
forecast to drop by 10–23% per year within the next fifteen years due to 12 MWh/hh.yr, Australia at 7 MWh/hh.yr and the UK, Germany and
factors such as the reduction in the price of poly-silicon, improvements Spain all at approximately 4 MWh/hh.yr [54], it is evident that Oman’s
in technology and increases in industry investment [52,53], these re- annual electricity consumption per household at 44 MWh/hh.yr is
sults indicate that these factors should be sufficient to drive PV costs to amongst the highest in the world. In turn, this study has shown that
levels that enhance the feasibility of grid-connected systems in Muscat. household demand in Oman is mainly driven by cooling demand, where
On the other hand, Fig. 10(b) shows that although a reduction in PV the demand profile follows a similar trend to the temperature and hu-
cost increases the feasibility of grid-independent systems, it is the re- midity profiles. This similarity in profiles indicates that there is a be-
duction in battery cost that has a more significant impact on the fea- havioural aspect to the high household demand in Oman, which in turn
sibility of grid-independent systems since battery cost constitutes 97% could be driven by the country’s low electricity prices. Therefore, in
of the total cost of grid-independent systems in Muscat. Additionally, it addition to focusing on the architecture of households, insulation and
is clear from this chart that at the current electricity price, grid-in- efficiency measures as well as cooling technologies used, further
dependent systems are infeasible in Muscat at any UCBatt and UCPV thought needs to be put into the phasing out of electricity subsidies in
where the IRRs are always negative. However, combined with an in- Oman as well as altering the electricity tariff structure in order to in-
crease in electricity price, grid-independent systems can become eco- fluence behaviour and reduce demand [55].
nomically feasible with significant reductions in UCBatt; where in- Secondly, the combination of high household demand, low elec-
creasing prices to Germany’s average requires the battery price to drop tricity prices and high battery cost renders grid-independent systems
by 93–94% (i.e. 6–7% of UCBatt) which is equivalent to 10–11 USD/ infeasible in Muscat. The high demand results in a larger overall system
kWh. When compared to the forecast in battery price reductions as size and therefore larger upfront costs, 97% of which is the cost of the
battery. Even increasing the PV panel efficiency to its maximum
329
J. Al-Saqlawi et al. Energy Conversion and Management 178 (2018) 322–334
theoretical efficiency does little to bring down this capital cost. In turn, large increase in electricity prices is needed to make grid-independent
reducing the unit cost of the battery can make grid-independent sys- PV systems economically feasible implies that in order to make such
tems feasible in Muscat provided electricity tariffs are increased to that investments feasible, in addition to an increase in electricity tariffs,
of Germany’s average. However, this reduction needs to be significant other PV support policies such as grants and loans are needed [61–63].
(93–94%) indicating that an increased installation capacity alone is not Finally, the fact that an increase in electricity price to an average
sufficient to achieve such price reductions and that further research in price lower than other PV successful countries such as Germany and
technology innovation is necessary. Spain makes grid-connected PV systems economically feasible with a
Furthermore, a larger system size not only makes grid-independent pay-back period of less than ten years despite there being no PV support
systems economically infeasible, but also makes them practically in- policies such as feed-in-tariffs (FITs) and the household demand in
feasible. It was found that in order to have a 302 kAh system for five Oman being very high demonstrates the potential of residential grid-
years, assuming that each battery has a capacity of 2.4 kWh and weight connected systems in Oman [61–63]. Furthermore, small reductions in
of 75 kg/kWh [56], 6040 batteries are needed that have a total weight PV cost as a result of increased installation capacity and industry im-
of 1087 tonnes occupying about 13% of the house [57–60]. Not only provements would only make grid-connected systems more feasible in
would storing such a large and heavy amount of batteries prove difficult Oman. However, the successful implementation of such systems re-
but the need to maintain them at ambient temperature could result in quires well thought out policies as part of a national energy and re-
further increasing the electricity demand. Similarly, changing the bat- newable energy strategy, with well thought out action plans and de-
tery type to one with a higher energy capacity such as a lithium-ion signated institutions which assist in permit attainment, finance,
battery (8–10 kWh) implies that although a reduced amount of batteries behavioural awareness, and research. To date, such policies still do not
are needed, i.e. approximately 1610 batteries weighing about 181 exist in Oman [61–64].
tonnes (12.5 kg/kWh), these numbers are still quite high. Additionally,
lithium-ion batteries tend to be 2.5–3 times more expensive than lead-
acid batteries [56]. Furthermore, lithium itself is a non-renewable, fi- Acknowledgments
nite product which makes this option less attractive in the long term.
On the other hand, low electricity prices imply that the benefits from We thank the Ministry of Higher Education-Oman (MOHE) for the
grid-independent systems cannot outweigh the costs. The fact that a funding of this project.
Appendix A. Nomenclature
Acronyms
AC Alternating current
c Si Crystalline-silicon
DC Direct current
FIT Feed-in-Tariff
OMR Omani rial
PV Photovoltaic
Greek symbols
Roman symbols
330
J. Al-Saqlawi et al. Energy Conversion and Management 178 (2018) 322–334
331
J. Al-Saqlawi et al. Energy Conversion and Management 178 (2018) 322–334
Table B.2
List of key inputs required for DC source model.
Key Model Inputs Value Unit Nomenclature References
2
Global Solar Irradiance – kW/m per GSI [40]
hour
Ambient Temperature – °C per hour TA [40]
Day of Year – per hour d [40]
Latitude 23.6 Degrees (°) [40]
Tilt Angle – Degrees (°) [40]
Electrical Efficiency of PV 12 % ref [15]
module
−1
Temperature Coefficient 0.004 K ref [15]
Reference Temperature 25 °C Tref [15]
Nominal Operating Cell 48 Degrees (°) NOCT [18]
Temperature
Power Demand of – kW per hour PD [1,39,58]
Residential Building
PV Degradation Rate 0.7 % per year PVD [31]
Roof Area 503 m2 Ar [57,58,65]
Proportion of Roof Area – % [13]
Covered with PV Module
Table B.3
List of key inputs required for grid-independent and grid-connected PV system models.
Key model inputs Value Unit Nomenclature References
Table B.4
List of key inputs required for economic model.
Key model inputs Value Unit Nomenclature References
332
J. Al-Saqlawi et al. Energy Conversion and Management 178 (2018) 322–334
Table C.5
List of model/input data assumptions and reasoning.
Assumptions Reasoning
IDT and IRT are negligible Most monitors do not measure IDT and IRT as these are difficult and expensive to obtain [13,40]
is represented by the maximum at solar noon This maximum is used in simple PV system design and is considered acceptable since GSIT is
underestimated [13]
The sun’s orbit is a perfect circle Allows for the factor of 360/365 to convert the day number (d) to a position in the orbit [13]
PV system can be fitted on the roof-top Roof-top PV systems are widely available [13]
PV module area (Am) is a function of roof area (Ar) PV modules are to be installed on the roof [13]
Humidity, air velocity, dust, shading, ohmic losses of conductors and Many of these variables are difficult to obtain without experimentation [13,14]
mismatch losses have no effect on PV performance
Tc is modelled at steady state This approach was considered more suitable for this study due to lack of data availability
NOCT model is suitable to determine cell temperature Gives adequate predicted temperatures for PV applications [17]
Grid connection is easily accessible Grid-connected systems are assessed in this study
The extent to which batteries can be charged and discharged is limited by Depending on the battery technology, full charge or discharge could permanently destroy the
SOCmin and SOCmax battery [29]
Battery charge/discharge rates are limited to ICmax/IDmax Battery capacity is affected by the rate at which the battery is charged/discharged [29,31]
Monthly BSD could be evenly converted to an hourly rate (BSDh) This is typical practice and allows for hourly assessment [66]
‘Cycle life’ impact on battery capacity is negligible therefore battery has an Initial assessments of the grid-independent system model for Oman have shown that batteries only
absolute lifetime, Nbatt [29] go through one complete charge and discharge cycle in a year
Charge controller is a standard switched controller These are the most common controllers used [28]
Inverter has an absolute lifetime, Ninv Several studies have assessed inverters as having an absolute lifetime [7,28]
inv of grid-connected system is the same as that for grid-independent To make both systems comparative to one another
system
can be changed manually on a monthly basis This is not too time consuming and allows for maximum monthly Pm generation
The PV technology is crystalline-silicon This is a mature technology which dominates the market [13]
NOCT is set at 48 °C NOCT ranges between 33 °C and 58 °C For a typical module, NOCT is 48 °C [18]
Battery used is a lead-acid battery Due to its maturity, familiarity and low cost [43]
Battery internal resistance is assumed to be zero and therefore VB is Lead-acid batteries have a very low internal resistance (< 100 milli ohms) [30] For systems greater
assumed to remain constant at 48 V than 3–4 kWh, it is recommended to set VB at 48 V [28]
Temperature effects on battery voltage and capacity are ignored Battery is assumed to be stored in-doors at constant temperature
voltage is set at 100% therefore Batt is equivalent to coulomb VB is constant [29]
SOCmax is set at 100% and SOCmin is set at 30% Lead-acid batteries can be fully charged while full discharge will permanently destroy the battery
[29]
NBatt is set at 5 years and Ninv is set at 10 years Typical lifetime of lead-acid batteries and inverters [7,31,74,75]
ICmax and IDmax are set at C/10 Charge is added and removed more efficiently at lower currents [29,31]
cc is set at 98% Typical efficiency of charge controller [31]
inv is set at 94% Inverter typically have efficiencies ranging between 92% and 98% [14]
N is set at 20 years Most PV systems have lifetimes ranging between 20 and 25 years [7,13,31,67]
Financial exchange rates do not change throughout the system lifetime Change in exchange rates in the future is difficult to predict
Grid power export tariff is the same as the current residential electricity To date, Oman has no renewable energy policy
tariff
Behavioural changes in electricity demand as a result of increase in Data on behavioural changes in residential electricity demand in Oman are minimal and require
electricity tariff are negligible extensive data collection and surveys which is beyond the scope of this study
Last available data for GSI and TA for the year 2007 are representative of The variation of GSI and TA between 2002 and 2013 is small [41,42]
2013
PD for January, February, March and December could be extrapolated Method devised due to lack of data availability
Model villa in Muscat is rectangular A large portion of the villa archetype in Oman is rectangular [58]
Model home wall thickness is 20 cm This is a standard thickness set by the buildings regulations for Muscat [57]
Hourly annual weather and demand data could be repeated annually for Due to the lack of hourly data availability, this allows for the modelling of PV systems which
20 years typically have a lifetime of 20 years
All simulations start on the first day of the month, mid-day In order to ensure consistency in the results
References battery hybrid plant for off grid electricity supply. Energy Proc 2015;75:381–7.
https://doi.org/10.1016/j.egypro.2015.07.396.
[7] Ghafoor A, Munir A. Design and economic analysis of an off-grid PV system for
[1] AER. Annual report, Tech. rep. Authority of Electricity Generation; 2011/2014. household electrification. Renew Sustain Energy Rev 2015;42:496–502. https://doi.
[2] COWI. Partners, Study on renewable energy resources, Oman. Tech. rep. Authority org/10.1016/j.rser.2014.10.012.
of Electricity Regulation; 2008. [8] Bastholm C, Fiedler F. Techno-economic study of the impact of blackouts on the
[3] Parra D, Patel MK. Effect of tariffs on the performance and economic benefits of PV- viability of connecting an off-grid PV-diesel hybrid system in Tanzania to the na-
coupled battery systems. Appl Energy 2016;164:175–87. https://doi.org/10.1016/j. tional power grid. Energy Convers Manage 2018;171:647–58. https://doi.org/10.
apenergy.2015.11.037. 1016/j.enconman.2018.05.107.
[4] Dufo-Lopez R, Perez-Cebollada E, Bernal-Agustin JL, Martinez-Ruiz I. Optimisation [9] Khoury J, Mbayed R, Salloum G, Monmasson E. Optimal sizing of a residential PV-
of energy supply at off-grid healthcare facilities using monte carlo simulation. battery backup for an intermittent primary energy source under realistic con-
Energy Convers Manage 2016;113:321–30. https://doi.org/10.1016/j.enconman. straints. Energy Build 2015;105:206–16. https://doi.org/10.1016/j.enbuild.2015.
2016.01.057. 07.045.
[5] Shezan SKA, Julai S, Kibria MA, Ullah KR, Saidur R, Chong WT, et al. Performance [10] Ratnam EL, Weller SR, Kellett CM. An optimization-based approach to scheduling
analysis of an off-grid wind-PV-diesel-battery hybrid energy system feasible for residential battery storage with solar PV: assessing customer benefit. Renew Energy
remote areas. J Clean Prod 2016;125:121–32. https://doi.org/10.1016/j.jclepro. 2015;75:123–34. https://doi.org/10.1016/j.renene.2014.09.008.
2016.03.014. [11] O. Schmidt, A. Hawkes, A. Gambhir, I. Staffell, The future cost of electrical energy
[6] Yilmaz S, Ozcalik HR, Aksu M, Karapinar C. Dynamic simulation of a PV-diesel- storage based on experience rates, Nature Energydoi:https://doi.org/10.1038/
333
J. Al-Saqlawi et al. Energy Conversion and Management 178 (2018) 322–334
nenergy.2017.110. storage to investigate the benefits of energy storage for single dwelling. Sustain
[12] PSE, gPROMs, accessed on 01 January 2014 (2014). URL <http://www. Cities Soc 2014;10:1–10. https://doi.org/10.1016/j.scs.2013.04.006.
psenterprise.com/>. [44] Dynamic, Oman electricity tariffs, accessed on 01 December 2015 (2015). URL
[13] Stenning A. Renewing corporate power: a techno-economic assessment of the <http://www.dynamic-ews.com/Tariffs/Electricity%20Tariffs/Oman.pdf>.
electricity supply of commercial sector buildings to be renewably resourced in an [45] Byrd, How does humidity affect air conditioning and heating, accessed on 06
urban environment case study Thesis London: Imperial Collage; 2014. February 2017 (2013). URL <http://byrdheatingandair.com/articles/how-does-
[14] Heinrich H. Photovoltaics system design and practice. Wiley; 2012. humidity-affect-air-conditioning-and-heating>.
[15] Dubey S, Sarvaiya JN, Seshadri B. Temperature dependent photovoltaic (PV) effi- [46] Alex, The Shockley Queisser efficiency limit, accessed on 01 November 2016
ciency and its effect on PV poduction in the world; a review. Energy Proc (2017). URL http://solarcellcentral.com/limits_page.html.
2013;33:311–21. https://doi.org/10.1016/j.egypro.2013.05.072. [47] Cengiz MS, Mamis MS. Price-efficiency relationship for photovoltaic systems on a
[16] Muzathik AM. Photovoltaic modules operating temperature estimation using simple global basis. Int J Photoenergy. doi:https://doi.org/10.1155/2015/256101.
correlation. Int J Energy Eng 2014;4:151–8. [48] Nykvist B, Nilsson M. Rapidly falling costs of battery packs for electric vehicles. Nat
[17] Trinuruk P, Sorapipatana C, Chenvidhya D. Estimating operating cell temperature Clim Change 2014;5:329–32. https://doi.org/10.1038/nclimate2564.
of BIPV modules in Thailand. Renew Energy 2009;34:2515–23. https://doi.org/10. [49] IRENA, The power to change: Solar and wind cost reduction potential to 2025.
1016/j.renene.2009.02.027. Tech. rep. International Renewable Energy Agency; 2016.
[18] Ross RJ, Smokler M. Flat-plate solar array project final report; 1986. p. 86–31. [50] Krane J. Stability versus sustainability: energy policy in the Gulf monarchies Tech.
[19] Wang Y, Tian W, Ren J, Zhu L, Wang Q. Influence of a building’s integrated-pho- rep. Cambridge University; 2013.
tovoltaics on heating and cooling loads. Appl Energy 2006;83:989–1003. https:// [51] EIA, Residential electricity prices in Europe and the United States, accessed on 15
doi.org/10.1016/j.apenergy.2005.10.002. May 2016 (2013). URL <http://www.eia.gov/todayinenergy/detail.cfm?id=
[20] Mattei M, Notton G, Cristofari C, Muselli M, Poggi P. Calculation of the poly- 18851>.
crystalline PV module temperature using a simple method of energy balance. Renew [52] Mayer JN. Current and future cost of photovoltaics Tech. rep. Fraunhofer-Institute
Energy 2006;31:553–0567. https://doi.org/10.1016/j.renene.2005.03.010. for Solar Energy Systems (ISE); 2015.
[21] Jain B, Ansari M, Patil VP. Evaluate PV-panel performance enhancement with [53] Pillai U. Drivers of cost reduction in solar photovoltaics. Energy Econ
matlab. Int J Electr Eng Technol 2014;06:23–37 <http://www.iaeme.com/ 2015;50:286–93. https://doi.org/10.1016/j.eneco.2015.05.015.
currentissue.asp?JType=IJEET&VType=6&IType=7> . [54] Enerdata, Average electricity consumption per electrified household, accessed on
[22] Zervas PL, Sarimveis H, Palyvos JA, Markatos NCG. Model-based optimal control of 15 August 2015 (2013). URL <https://www.wec-indicators.enerdata.eu/
a hybrid power generation system consisting of photovoltaic arrays and fuel cells. J household-electricity-use.html>.
Power Sources 2008;181:327–38. https://doi.org/10.1016/j.jpowsour.2007.11. [55] Al Shehri AM. The Report: Saudi Arabia. Oxford Business Group; 2010.
067. [56] Segen, Photovoltaic battery guide, accessed on 01 August 2016 (2016). URL
[23] Jordan DC, Kurtz SR. Photovoltaic degradation rates; an analytical review Tech. <http://www.segen.co.uk/co.uk/batteryguide>.
rep. National Renewable Energy Laboratory; 2012. [57] M. Muncipality, Building regulations for Muscat: Local order no. 23/92; 1992.
[24] McPherson J. Reliability physics and engineering. Springer; 2013. [58] Sweetnam T. Residential energy use in Oman: a scoping study. Tech. rep. Passive
[25] Zhu J, Astawa K, Betts T, Gottschalg R. A modelling approach for long-term de- Systems; 2014.
gradation of thin film silicon photovoltaic modules. Sol Energy Soc. URL <https:// [59] Anuphappharadorn S, Sukchai S, Sirisamphanwong C, Ketjoy N. Comparison the
dspace.lboro.ac.uk/2134/8167>. economic analysis of the battery between lithium-ion and lead-acid in PV stand-
[26] Manuel Vazquez IR-S. Photovoltaic module reliability model based on field de- alone application. Energy Proc. 2014;56:352–8. https://doi.org/10.1016/j.egypro.
gradation studies. Wiley InterScience. doi:https://doi.org/10.1002/pip.825. 2014.07.167.
[27] Akikur RK, Saidur R, Ping HW, Ullah KR. Comparative study of stand-alone and [60] Mehrabankhomartash M, Rayati M, Sheikhi A, Ranjbar AM. Practical battery size
hybrid solar energy systems suitable for off-grid rural electrification: A review. optimization of a PV system by considering individual customer damage function.
Renew Sustain Energy Rev 2013;27:738–52. https://doi.org/10.1016/j.rser.2013. Renew Sustain Energy Rev 2017;67:36–50. https://doi.org/10.1016/j.rser.2016.
06.043. 08.050.
[28] Kempener R, Lavagne d’Ortigue O, Saygin D, Skeer J, Vinci S, Gielen D. Off-grid PV [61] IEA. Trends 2014 in photovoltaic applications Tech. rep. International Energy
power systems Tech. rep. IRENA; 2013. Agency; 2014.
[29] Linden D, Reedy TB. Handbook of batteries. McGraw Hill; 2002. [62] IEA. Technology roadmap: solar photovoltaic energy Tech. rep. International
[30] Jossen A, Garche J, Sauer DU. Operation conditions of batteries in PV applications. Energy Agency; 2014.
Sol Energy 2004;76:759–69. https://doi.org/10.1016/j.solener.2003.12.013. [63] Dusonchet L, Telaretti E. Comparative economic analysis of support policies for
[31] Jenkins DP, Fletcher J, Kane D. Model for evaluating impact of battery storage on solar PV in the most representative EU countries. Renew Sustain Energy Rev
microgeneration systems in dwellings. Energy Convers Manage 2008;49:2413–24. 2015;42:986–98. https://doi.org/10.1016/j.rser.2014.10.054.
https://doi.org/10.1016/j.enconman.2008.01.011. [64] IRENA. Sultanate of Oman, Renewables readiness assessment Tech. rep.
[32] Jung J, Tyner WE. Economic and policy analysis for solar PV systems in Indiana. International Renewable Energy Agency; 2014.
Energy Policy 2014;74:123–33. https://doi.org/10.1016/j.enpol.2014.08.027. [65] AER. Existing data report Tech. rep. Authority of Electricity Generation; 2014.
[33] Ma WW, Rasul MG, Liu G, Li M, Tan XH. Climate change impacts on techno-eco- [66] B. University. What does elevated self-discharge do?, accessed on 01 July 2016
nomic performance of roof PV solar system in Australia. Renew Energy (2003). URL <http://batteryuniversity.com/learn/article/elevating_self_
2016;88:430–8. https://doi.org/10.1016/j.renene.2015.11.048. discharge>.
[34] Rodrigues S, Torabikalaki R, Faria F, Cafofo N, Chen X, Ivaki AR, et al. Economic [67] Powerthru. Lead acid battery working technical paper, accessed on 11 July 2016
feasibility analysis of small scale PV systems in different countries. Sol Energy (2016). URL <http://www.power-thru.com/documents/The%20Truth%20About
2016;131:81–95. https://doi.org/10.1016/j.solener.2016.02.019. %20Batteries%20-%20Powerthru%20White%20Paper.pdf>.
[35] Li C, Yu W. Techno-economic comparative analysis of off-grid hybrid photovoltaic/ [68] IRENA. Renewable energy technologies: cost analysis series-solar photovoltaics
diesel/battery and photovoltaic/battery power systems for a household in Urumqi, Tech. rep. International Renewable Energy Agency; 2012.
China. J Clean Prod 2016;124:258–65. https://doi.org/10.1016/j.jclepro.2016.03. [69] Dufo-Lopez R, Zubi G, Fracastoro GV. Techno-economic assessment of an off-grid
002. PV-powered community kitchen for developing regions. Appl Energy
[36] Mac Dowell N, Fajardy M. Inefficient power generation as an optimal route to ne- 2012;91:255–62. https://doi.org/10.1016/j.apenergy.2011.09.027.
gative emissions via BECCS? Environ Res Lett. doi:https://doi.org/10.1088/1748- [70] Fraunhofer. Photovoltaics report Tech. rep. Fraunhofer Institute for Solar Energy
9326/aa67a5. Systems; 2015.
[37] Goetzberger A, Hoffmann VU. Photovoltaic solar energy generation. Springer; [71] Google, Google shopping-PV panels, accessed on 01 March 2016 (2016). URL
2004. <https://www.google.co.uk/>.
[38] Okoye CO, Taylan O, Baker DK. Solar energy potentials in strategically located [72] Alibaba, Alibaba-PV panels, accessed on 01 March 2016 (2016). URL <www.
cities in Nigeria: review, resource assessment and PV system design. Renew Sustain alibaba.com>.
Energy Rev 2016;55:550–66. https://doi.org/10.1016/j.rser.2015.10.154. [73] XE, Currency converter, accessed on 01 April 2016 (2016). URL <http://www.xe.
[39] Sweetnam T, Al-Maskari B. Re: Oman data, Private communication, type to Al- com/currencyconverter/>.
Saqlawi, J, October 2015. [74] Dufo-Lopez R, Lujano-Rojas JM, Bernal-Agustin JL. Comparison of different lead-
[40] Al-Zakwani S. Re: Muscat meterological station data, Private communication, acid battery lifetime prediction models for use in simulation of stand-alone pho-
Directorate General of Meteorology, type to Al-Saqlawi, J, September 2015. tovoltaic systems. Appl Energy 2014;115:242–53. https://doi.org/10.1016/j.
[41] NCSI. Climate statistics bulletin. Tech. rep., National Centre of Statistics and apenergy.2013.11.021.
Information, 2011–2013. [75] Guinot B, Bultel Y, Montignac F, Riu D, Borgne IN-L. Economic impact of perfor-
[42] NCSI. Statistical yearbook. Tech. rep. National Centre of Statistics and Information, mances degradation on the competitiveness of energy storage technologies-Part 2:
2010–2014. application on an example of PV production guarantee. Int J Hydrogen Energy
[43] Parra D, Walker GS, Gillott M. Modeling of PV generation, battery and hydrogen 2013;38:13702–16. https://doi.org/10.1016/j.ijhydene.2013.08.023.
334