Optimization of the load-and-haul operation at an opencast

colliery
O. Pasch; S. Uludag

School of Mining Engineering, University of Pretoria, South Africa

SYNOPSIS

The current coal mining climate is characterized by coal price volatility, political
instability, high labour costs, and increasing operational costs. This is exacerbated by a
steady decline in the growth of global coal demand due to the increased use of
alternative and renewable fuels in the energy industry. Locally, the overall mining cost
inflation indices shows a yearly increase of 2% over the national consumer inflation. In
order for coal mines to survive and mine profitably, they need to capitalize on the
opportunity to improve their productivity and focus on one factor they can control:
operational efficiency. Increasing productivity is one of the key drivers to counter
diminishing profit margins. Increasing production effectively reduces operating costs.
However, the emphasis should not only be on increasing output with the same input, but
increasing the output while decreasing the input, and ultimately adding optimum value
to current resources. Research shows that an increase in production will ultimately
decrease the operation's unit cost, especially fixed costs.
In this study a load-and-haul fleet optimization approach has been used to identify the
opportunities for operational improvement at an opencast colliery. The study combines
the results of a literature review, on-site time studies, and statistical data analysis in
order to determine the best loader-truck fleet combinations for increased production.
Several relevant key performance indicators (KPIs) for the evaluation and identification
of productivity improvement opportunities were defined during this study. These KPIs
are bucket fill factor, loading conditions, loading cycle time, utilization, and deviations
from schedule. The priority delays determined by on-site time studies compared to the
time book for each delay showed that idle or waiting time by the loaders, face
preparation and relocation, and process delays had significant deviations. However, the
results showed that this operation is under-trucked, hence optimizing the loader-related
inputs proved less effective than optimizing truck-related inputs. The results indicated
that a homogeneous truck fleet consisting of five Caterpillar 789C trucks, combined with
a Caterpillar 994K loader, is the most efficient fleet option and will produce 1455 t/h.
The combined optimized effect of each identified KPI of production led to a tonnage
improvement opportunity of 5421 t per shift.

Keywords: optimization, productivity, load-and-haul fleet, KPIs, under-trucked, tonnage

improvement opportunity.

Mine background

The opencast coal mining complex extends over two mines within the Witbank Coalfield,
to the south of Emalahleni, Mpumalanga Province. The target pit is currently a mature
opencast strip-mining operation that extracts multi-layered coal seams, exercising a
throw-over method to the low-wall side to the north and advancing in a southerly
direction. The area was previously mined using a conventional underground bord-and-
pillar method.
The mining complex produced a total of 30.4 Mt in the preceding financial year, about
8.5 Mt/ of which was produced at the opencast colliery. The mine was established in
2010 and consists of a mixture of greenfield and brownfield operations that supply
energy coal to both local and export markets. Coal is supplied to Eskom's power station
via a conveyor system, while a higher quality export grade is supplied to mainly
European and Indian markets.

Project background

In recent years, the global effort towards utilization of alternative and renewable energy
resources has driven the demand for energy from coal downward. In addition to this,
various sources have highlighted that fact that coal is a finite resource; it is inevitable
that coal resources are heading towards depletion. Meanwhile, supplying coal has
become increasingly more challenging. This is because mining operations usually extract
the reserves nearest to the surface or the most conveniently accessible reserves. Coal
deposits that have been left in situ by previous operations may now be sterilized, or if
the extraction opportunity exists, will require sizeable financial investments. Operational
costs for South African mines are increasing at a rate that diminishes the margin to mine
profitably. The developing South African economy has had difficulty maintaining a
positive GDP growth rate, weakening the rand to US dollar exchange rate. Not only does
this play a decisive role in the mining industry's export sector, but it also increases
mining operational input costs significantly. Most operational expenses like consumables,
electricity, equipment, fuel, and various overhead costs are heavily inflated by a weaker
rand.

Production challenges

Pillar mining is one of the most challenging coal extraction methods in an opencast
environment. Remnant coal pillars in old workings are also susceptible to spontaneous
combustion. Not only does this degrade the quality of the coal, it also creates an
extremely hazardous operating environment. As a result, overheating of equipment
occurs frequently, which leads to regular production stoppages. The ambient
environment caused by the smoke reduces operator visibility and generates
uncomfortable operating conditions. The uneven and undulating country rock results in
water accumulation on the pit floor. This contributes to productivity losses in a number
of ways, including trapping of equipment by mud, pumping arrangements, unseen
potholes damaging tyres, and material losses from loaded haul trucks. Another factor
that affects productivity is operator proficiency. Operators may be tempted to ignore
reporting and logging responsibilities as a result of working in a strenuous environment.

The exposure of country rock underneath the coal seams also adds to the challenges.
This results in poor road conditions that affect hauling by increasing travel time, causing
excessive wear on haul truck tyres, and increasing safety risks, to mention only a few.
Hauling coal from a pit some 60 m below surface means that loaded trucks must travel
upwards at steep inclines. This reduces the speed trucks can travel at and adds
considerable time to the hauling cycle. As extraction advances with each strip, the
hauling distances to the tipping points increase, in turn increasing the truck cycle time.

Objectives and methodology

The following objectives were formulated.

► Determine and quantify the main factors that affect the productivity of the load-and-
haul fleet at the mine by means of value drivers.
► Identify, compare, and analyse the major delays that have created shortcomings in
the current extraction process to arrive at the potential capacity of the mine's extraction
process.

► Identify the tonnage improvement opportunity associated with each KPI.

► Recommend practical improvements that target priority KPIs in order to increase

productivity at the target pit.

► Develope an optimized fleet solution for the current coal extraction operation through
the basic mining equations.

► Conduct a basic cost evaluation for each improved KPI based on benchmarking and
industry standards.

A literature review was conducted to gain understanding of the background of the

project as well as to limit the context to the milieu of the investigation. In addition, the
literature was used to gain knowledge of industry standards in order to benchmark costs
where applicable. Various literature sources were continuously reviewed to add
knowledge and understanding to the problem statement throughout the investigation.
Preliminary research related to the theme of the study includes conventional data-
gathering techniques like interviews with employees of the mine and group-based
discussions. These insights added qualitative value and support to a mainly quantitative
investigation. Documents, records, and statistical data were also retrieved from the
mine's online Integrated Management System and utilized. Information relevant to the
project scope was made available to the investigator with permission from the General
Manager of the operation. The third method of investigation was based on a time study
where the investigator recorded overall loading and hauling cycle times over 15 twelve-
hour shifts. Simultaneously, side-by-side on-site observations were recorded. This
allowed the investigator to record multiple 'day-in-the-life-of observations whereby half-
hourly events could be logged. These methods can be categorized as visual observations
done over a period of 180 hours. Quantitative data obtained was used to calculate a
benchmarked improvement tonnage opportunity target.

The Time Usage Model, proposed under the ownership of the mine was applied during
the course of the study, which drives a consistent approach to measure productivity.
This model provides a standard methodology to calculate time-related parameters for
equipment at the mine. Time-related parameters are one of the most fundamental
aspects in measuring the relative performance of mines, hence this model contributes to
the continuous iteration in order to achieve optimum productivity.

Scope of study

The investigated extraction processes are limited to the targeted pit's loading and
hauling fleet. Only KPIs and value drivers from extraction processes will be considered in
creating a benchmark for the mine. A generic methodology on how to adapt the in-pit
extraction processes is only applicable to similar operations. i.e. opencast collieries. Due
to the confidential nature of mining financials, relative measures for operational costs of
the mine are used unless stated otherwise.

Literature review

Production
The following key excerpts were taken from various sources relevant to the company's
coal market.

► South Africa supplied 66 Mt for the global coal supply in 2016 (Anon., 2016)

► South African collieries produced 176.9 Mt in 2007 to supply the local market (Anon.,
2016)

► The company produced 14.85 Mt in FY16 for export, of which an estimated 8.25 Mt
were produced at the targeted mine (Anon., (2015). A decrease to 12.8 Mt in FY18 is
expected by the company (Anon., 2016)

► The targeted mine supplied Eskom with 8 Mt in FY15. The company plans to decrease
the mine's contribution to 6.75 Mt in FY 17 (Anon., 2015).

A common focus for a mining operation is to improve productivity to reduce cost

pressures. However, two increasingly important and inevitable factors remain critical,
which mining companies must continuously analyse and evaluate in order to maintain
profitability - the depletion of coal resources and the shift towards alternative energy
(renewable energy sources).

Productivity in mining terms can be generally defined as the same output for less input.
Baxter (2015) argues that productivity gain should be measured as a form of
optimization. In other words, the drive to productivity should focus on increasing the
ratio of output to input. This is evident from Figure 2, which highlights the general
motivation to increase the product or output but reduce the input or cost per unit.

 
 

Load-and-haul operation

The load-and-haul operation can be seen as a continuous cycle that is made up of

different production steps or activities. The load-haul cycle time is the total time to
complete a full cycle. The critical cycle steps, as time fractions of the total cycle, have
been identified by Krause (2006):

► Spotting at loading is the time required for a truck, as soon as it arrives near the
vicinity of the loader, to manoeuvre into a stopping position for loading.

► Loading time is the total time for the loader to load the bucket of the truck to its
required payload.

► Hauling-full time is the total travelling time for a loaded truck to reach the dump site
from the loading site.

► Travel empty time refers to the total travelling time to for an empty truck to reach the
load site from the dump site.

► Queuing time is the total time an empty truck has to wait in line before it can
manoeuvre into a position for loading.

Another factor that is commonly ignored is the waiting time of a loader. 'Waiting at
dump', 'Queuing', and lastly 'Waiting time of loader' are the three noteworthy delays that
are not directly caused by the performance of the equipment, but rather due to load-
haul equipment combinations.

Time usage model

The company promotes the use of a standardized method by which productivity can be
measured and calculated, in addition to the reporting and documenting of equipment
usage. Figure 3 illustrates a typical time usage model. The time usage model illustrates
time factors that are related to the total time that equipment is not performing its
required, planned, and intended function. However, the total time under consideration
when equipment is performing its required, planned, and intended function is referred to
as Equipment Production Time or Direct Operating Hours (DOH). The DOH is defined by
the time when measurable throughput in the process is established. This is generally
considered to be the product of the available time for equipment to perform work and
the actual utilization time to maintain a productive cycle. Various equipment
performance metrics that will determine the productivity of an operation can be derived
from the time factors. These relative measures are generally called key performance
indicators (KPIs) (Choudhary, 2015).

Equipment match factor

In a load-and-haul operation, the capacities of the loaders should be compatible with the
capacities of the truck fleet. Choudhary (2015) presents a simplified model to determine
if a fleet is under-trucked or over-trucked. This is represented by the following formula:

If the current fleet operates with a number of trucks exceeding the value determined by
Equation [1], the fleet is said to be over-trucked, and if less than the value, under-
trucked. Equation [1] does not consider the queuing times, thus it can only be applied as
a comparison whereby two ideal fleets are considered.

The equipment match factor (MF) between two interdependent pieces of equipment
takes into account the total cycle time of the equipment. According to Choudhary (2015)
and Krause (2006), the equipment match factor refers to the ideal capacity and number
of trucks that are paired in operation with a loader and how well they are suited to each
other. The variables in Equation [2] are dependent on the equipment specifications as
well as fixed performance capacities, i.e. payload, truck height, and loader reach to
name a few.

The match factor, which provides a measure of the productivity of the fleet, can be
calculated by Equation [2]. Although the MF ratio does not take into account the
equipment capacities and specifications, it is inherently considered by applying the
equipment cycle times in the equation. Choudhary (2015) notes that the MF ratio is not
to be applied to the efficiency of the production itself, but only to the efficiency of the
selected combined fleet. Choudhary (2015) goes on to state that an MF of 1.0 indicates
that the fleet is 100% compatible regarding size and timing, but in this instance the
combination can only produce 10% of the capacity of a fleet that has an MF of 0.5. The
mining industry would rather opt for a lower MF limit, as this will correlate with a lower
operating cost. Choudhary (2015) states that the MF is best applied in combination with
other approaches, such as experiential and iterative methods, for determining the
haulage fleet size.

Equipment specification

The mine has become accustomed to the exclusive use of CAT equipment for their load-
and-haul operation. According to CAT Mining (2017), the 993K loader with a standard
coal bucket is sized to load 90 t in three to four passes to the 777D, and load 136 t in six
passes to the 785C. This amounts to an average of approximately 22.5 t per load. The
994K loader with a standard coal bucket is sized to load 136 t in four passes to the 785C
and 177 t in five passes to the 789C. It is also capable of loading 227 t to the 793 in six
passes. Hence, in theory, the 994K loader can average approximately 36 t per load.
Bucket selection is a critical factor in any extraction process. According to CAT Mining
(2017), in-seam coal is best mined with a rock-type bucket, and they recommend that
serrated edge buckets be used as this provides the highest penetration rates.

Bucket fill factor (BFF) is used to determine how well the volume of a bucket is utilized.
Bucket fill factor is very useful to determine the productivity of a combined fleet.
Mathematically, it is expressed in Equation [3] (Mohammadi, Rai, and Gupta, 2015).
Bucket fill factors are usually impractical to obtain from in-field measures but an
estimation scale can be used to objectively rate each bucket load accordingly.

Productivity for coal haul trucks is not limited by the nominal payload perse, but rather
by the volumetric design and dimensions of the bucket. Due to the low relative density
of coal, which is amplified by the swell factor of broken coal during loading, there is no
theoretical limitation to the output per ton and the truck payload.

Results and interpretation

The total fleet sizes seen in Table I and Table II indicate the available equipment for the
duration of the study. The sample size includes the equipment from the fleet on which
observations, measurements, and recordings were made and from which relevant data
was collected. However, the most significant amount of data was measured and recorded
from 994K loaders and 789K haul trucks.

 
 

It can be argued from the total and sampled fleet sizes that the load-and-haul coal
extraction operation is under-trucked. However, under-trucking or over-trucking varies
with the time parameters in the truck cycle time. According to CAT (2017), each loader
has been benchmarked to service at least three to five trucks. From Table I and Table II,
looking at the sampled fleet size, it is seen that one 994K loads on average loads two
trucks, which indicates that the operation is under-trucked.

Bucket fill factor

Table III shows that each loader-truck combination deviates significantly from the
benchmarked passes or loader loads prescribed by CAT Mining (2017). The 994K-789C
pairing deviates 26% from benchmarked practice, effectively resulting in a 74.1% BFF
for the loader (see Figure 1).

Optimizing the loader BFF to a mine-ideal standard will inevitably decrease the passes
per load. Increasing the BFF of the loader results in two improvements the overall
productivity. The first is a decrease in passes, which will shorten the loading time while
still maintaining the same mass per loading cycle. This results in a 90 t/h increase (see
Table IV). However, if the cycle is under-trucked, reducing the loader cycle time will not
necessarily add direct production value. This is because the loader has to wait for trucks
in an under-trucked cycle. No additional tonnage can be hauled unless another truck is
added to the cycle. Additional available time 'created' in this improvement can, however,
be spent on face preparation for the next load. A second option may be considered
whereby an additional load can be loaded, making six loads per loading cycle. This will
increase the BFF of the trucks to 90.5% (144.9 t). This results in a total increase of 156
t/h (see Table IV). Applying this method will evidently increase the tonnage produced in
an under-trucked cycle since more tons are added while operating at the same loading
and truck cycle time.

The Australian Department of Resources, Energy and Tourism (2014) has indicated that
with an increase in the average payload, the fuel cost per ton decreases as more tons
are being hauled. The aforementioned source also states that the fuel usage does not
increase significantly when the payload is increased from 80% to a 95% BFF (Australian
Government, 2014). By comparing only fuel costs of the current BFF to the second
tonnage improvement opportunity, the net opportunity profit per hour resulted in a
R19.6 per ton decrease in unit cost.

Loading conditions

Loading conditions in-pit differ greatly from stockpile loading conditions. Loading times
from the pit and from the stockpile will also differ from theoretical loading times, as well
as benchmarked loading times and assumed conditions. In the case of this study,
stockpile loading conditions are presumed to be either ideal or near-ideal for the
following observed reasons.

► The stockpile floor grade is periodically maintained by graders, thus providing a flat
and horizontal load floor that improves loading performance.
► The coal stockpiles are loosely packed against the angle of repose; hence minimum
penetration force is used over the shortest possible time at an angle that is ideal for
bucket designs to maximize bucket fill factors.

► Less spontaneous combustion on stockpiles, as the time the coal spends on the
stockpiles is kept to a minimum. This allows for loader operators to plan and execute
loading methods at a high accuracy and in the shortest possible time.

► If stockpiles are effectively managed, the loader can be ideally positioned for the
loading method used, with sufficient space to move around.

The following conditions and practices were observed at in-pit loading operations.

► Uneven floor conditions as toes and irregularities in the footwall of the coal seam
occur frequently, resulting in low bucket fill factors and consequently increasing loading
time.

► Increased breakout force is required by the loaders to break coal from in situ
conditions, thus increasing loading time.

► Limited space and changing coal face conditions in the pit led to frequent
manoeuvring and positioning for loading trucks. This caused the loading method to
change frequently, adding to time wasted.

► Spontaneous combustion caused extreme loading conditions, preventing the operator

from loading efficiently. Operators have to continuously assess the ambient conditions
and manoeuver into positions that enable the loader to remain at a safe distance from
the burning coal but still access the coal face.

Cycle times

A mine ideal standard was developed to compare actual results against a targeted mine
potential. The mine ideal was determined by benchmarking equipment specifications,
supported by the upper quartile data distribution of recorded data obtained from the
mine. This means that the mine ideal is now standardized and based on the best cycle
activity time that had been reached in a quarter of the time over which the data were
recorded. The cycle times were recorded (see Table V) for 161 truck cycles for the
combination of 994K loader and 789C haul truck. The most significant contributor to the
overall cycle time deviation is the loading activity, which contributed 37%. This is mainly
due to mitigating the hazardous loading conditions i.e. water-spraying, floor grading,
and face preparation. It was also determined that 27% of the total deviation is due to
the queuing time. The results indicated that the current extraction technique or loading
cycles could possibly be over-trucked, or the loader is not fully utilized for loading the
trucks. This is in contrast with the results from the fleet sizes, which showed an under-
trucked operation. The queuing times can be explained by a few localized events where
either all trucks were dispatched to one loader when the other loaders were unavailable,
or downstream process downtime due to crusher maintenance.

 
 

Table VI highlights the tonnage opportunity per shift or production improvement

opportunity when only the cycle time is optimized to a mine ideal. It is important to keep
in mind that the above summary is presented on the assumption that there is no
queuing time, which is expected in an under-trucked operation, but frequently
impractical. Yet, with the reduction in cycle time of four minutes and four seconds, the
total improvement is 154 t per shift per truck for the 994K-789C combination.

In the case of improved cycle times, the payload per cycle will remain constant, in
conjunction with all other factors, ceteris paribus, i.e. maintenance, tyres, and
operator/labour costs. The fuel costs may increase due to the increase in the total
haulage distance per shift, but this may be considered as negligible as the fuel cost
associated with idling at queuing are assumed to be used for hauling. The total cost
opportunity due to the increase in tonnage per shift amounted to a net decrease of
R1.94 per ton.

Equipment match factor

According to Equation [1], the optimal number of 789C trucks paired with one 994K
loader is five. Currently, the fleet is operating at an average MF of 0.59. This confirms
that the operation is currently under-trucked. According to Equation [2], the mine ideal
MF ratio should be 0.96. The MF ratio can now be used as an index of the overall fleet
efficiency or as a relative efficiency measure. Including the waiting and queuing times
recorded in the study will provide a more accurate insight into cases where mixed fleets
are used, as is the case at this mine. Results show that the optimum fleet in terms of
tons produced per hour consists of one 994K loader and five 789C haul trucks. This
yielded a total of 1455 t/h. Alternatively, removing one 789C truck in this fleet will
decrease the tonnage by approximately 75 t/h. From the results, it can be deduced that
there may be a linear relationship between the MF and the queuing times. In reducing
the queuing time, the MF will tend to 1.0, which indicates a perfect MF. Combinations
with the lowest queuing times does not necessarily reflect the best possible production
fleet. The MF ratio for the optimized fleet yielded the largest MF ratio in this specific
case, yet it resulted in the best possible production fleet. Thus, these MF calculations
cannot be used in isolation from other fleet optimizing measures.

Productivity

Due to the nature of time studies and the associated methodology, the study focused on
activities associated with delays in the extraction process. The average reported
availability over the study period was 86.6%, according to data collected from the mine's
Integrated Management System (IMS). Figure 5 compares the observed and reported
utilization. In eight out of ten days, the observed utilization was lower than what was
reported. The average IMS reported utilization is 83.2%, whereas the average measured
results show a 71.3% utilization. For the period between Christmas Day and New Year's
Day, utilization reduced significantly. This period is generally known in the mining
industry as the 'Silly Season', which typically reflects increased lost-time injury (LTI)
trends.

Table VII shows the contribution of each delay to the total measured lost time. The main
contributors to these delays are idle or waiting time. Using the rate of production, 1445
t/h, this amounts to a total of 10 405 t per shift and consequently 20 810 t/d. If delays
could be reduced so that the full 602 minutes available could be utilized, a total of 26
688 t/d could theoretically be achieved for the targeted pit.

Figure 6 highlights the opportunity tonnage associated with each delay. An additional 2
940 t per shift is achievable, which could increase revenue by 18%.

Conclusions

The targeted pit is a high-producing asset for the mine, and the company is heavily
dependent on this pit as an operation and a company. Several relevant KPIs for the
evaluation and identification of productivity improvement opportunities were defined
during this study. These KPIs are bucket fill factors, loading conditions, cycle and loading
time, time utilization, and deviations from schedule. The combined tonnage
improvement opportunity can be viewed in Figure 7, which reflects a theoretical
improvement of 49.4%. From the research, it is evident that an increase in production
will ultimately decrease the operational unit cost, which could increase profit margins
significantly.