Multi-factor analysis of haul truck fuel consumption in open-pit mining: Insights from continuous real-time monitoring

Abstract

As fuel consumption is a major operating cost in open-pit mining, identifying and quantifying controllable inefficiencies requires continuous monitoring under real production conditions. This study presents a comprehensive analysis of haul truck fuel consumption using an integrated onboard monitoring system installed on a Komatsu 785 truck at a copper mine. The system continuously measured fuel flow, payload, position and speed over 90 days and captured 1780 complete haul cycles over 150 shifts under varying operational, environmental and behavioural conditions. Specific fuel consumption (SFC) analysis revealed that road gradient was the dominant factor, with uphill waste-haul routes consuming 75–80 gr/t.km compared to 65–70 gr/t.km on crusher routes. An optimal payload range of 92–97 tonnes minimised the SFC, while both under-loading and over-loading reduced fuel efficiency. Spatial mapping identified specific high-consumption road segments that required maintenance. Driver behaviour significantly affected fuel use, with aggressive driving increasing the SFC by approximately 7–10% (75–77 gr/t.km) compared to normal driver behaviour (70–72 gr/t.km). Rainfall events increased SFC by 2–3 gr/t.km above the baseline due to elevated rolling resistance. Operational delays, particularly loading queues (45% of total idle time), contributed substantially to unproductive fuel consumption. The findings demonstrated that fuel inefficiencies were largely controllable through targeted operational improvements, including payload optimisation, road maintenance, driver training and dispatch co-ordination. The spatial mapping methodology provided a practical diagnostic tool transferable to other mine sites for identifying energy-intensive haulage segments and prioritising fuel-reduction interventions.

Introduction

Open-pit mining is one of the most widely used methods for exploiting large, near-surface mineral deposits. To maintain competitive production rates, current operations rely on fleets of large diesel-powered equipment, including hydraulic shovels, wheel loaders, drills, bulldozers and especially haul trucks. While they have enabled rapid increases in productivity, they also have consumed very high levels of energy. In most hard-rock open pits, diesel fuel is one of the largest single operating cost items.

Open-pit mining relies on large fleets of diesel-powered equipment to maintain competitive production rates, with haul trucks typically representing the largest single category of fuel consumption. Studies of mine energy balances show that truck haulage accounts for one-third or more of total site diesel consumption (Dindarloo and Siami-Irdemoosa, 2016; Dumakor et al., 2017; Siami-Irdemoosa and Dindarloo, 2015). As fuel prices rise and environmental regulations tighten, improving fuel efficiency has become critical for both economic competitiveness and environmental sustainability. However, the complex interaction of operational, environmental and behavioural factors makes fuel consumption difficult to predict and to control without detailed monitoring under actual production conditions.

Haul truck fuel consumption is governed by multiple interacting factors that can be broadly categorised into vehicle design parameters, route characteristics, operational practices and environmental conditions. Vehicle-related factors include gross vehicle weight, payload, engine efficiency and tyre conditions (Tadubana et al., 2024). Route characteristics such as haul distance, road gradient, rolling resistance and surface quality directly influence engine load and energy demand (Rodovalho et al., 2016; Wang et al., 2019). Operational factors, including travel speed, acceleration patterns, idle duration, loading and dumping times and traffic congestion, affect fuel use throughout the haulage cycle (Soofastaei et al., 2016). Environmental conditions such as ambient temperature, altitude and precipitation can further modify fuel consumption by altering engine performance and rolling resistance (Rodovalho et al., 2016).

Considerable research has been devoted to modelling and predicting haul truck fuel consumption using diverse analytical approaches. Machine learning methods have proven effective for capturing non-linear relationships between operational variables and fuel use. Siami-Irdemoosa and Dindarloo (2015) successfully applied artificial neural networks to model fuel consumption as a function of payload, speed and road rolling resistance. Other researchers have employed support vector machines (SVMs) and random forests (RFs) to predict fuel consumption from operational data (Alamdari et al., 2022). Statistical regression approaches, including multiple linear regression (Rodovalho et al., 2016) and partial least squares regression (Yousefi Nejad Attari et al., 2021), have been used to quantify the impact of individual parameters on fuel consumption and identify opportunities for operational improvement. From an optimisation perspective, studies have applied mixed-integer linear programming (Yousefi Nejad Attari et al., 2021) and discrete-event simulation (DES; Bajany, 2017; Meneses and Sepúlveda, 2023) to minimise fleet-level fuel consumption through improved truck allocation, routing and traffic flow management. While these modelling approaches have enhanced our understanding of fuel consumption patterns, most rely on aggregate fleet management data or controlled experimental conditions rather than continuous high-resolution monitoring under normal production operations. Additionally, a study from Soofastaei et al. (2016) shows that DES combined with optimisation can reduce fuel consumption by reducing bunching, smoothing traffic flow and improving haul-road design. Furthermore, SVM and RF have been applied by Alamdari et al. (2022) to predict haul-truck fuel consumption. Bajany et al. (2019) presented a realistic model that can be used to minimise fuel consumption in surface mining operations. The developed model dispatches trucks to destinations by optimally determining the paths that result in minimising fuel consumption per truck and shovel cycle time whilst meeting the handling demands of each dumping site. Nobahar et al. (2022) presented an optimum fleet selection method based on a mine's historical data and machine learning algorithms to supply the mineral processing plant demand. Alexandrov et al. (2019) demonstrated that accounting for both external and internal influencing factors when determining the optimal operating mode of a mining truck can further reduce specific fuel consumption (SFC) by selecting appropriate travel speeds in both loaded and empty directions.

Despite these advances, several limitations constrain the practical application of existing fuel-consumption research to operational decision-making. First, most modelling studies rely on aggregated fleet-management data or short-term experiments that do not capture the full range of variability encountered in day-to-day production. Second, while individual factors have been extensively studied, few works have simultaneously quantified the relative contribution of operational, behavioural and environmental drivers using continuous field data from a single instrumented truck. Third, spatial variation in fuel consumption over specific road segments, which directly affect road maintenance and geometric design decisions, has rarely been characterised in a systematic way. Finally, it remains challenging to translate model outputs into actionable improvements without a monitoring system that can attribute fuel use to specific controllable factors such as loading accuracy, driver behaviour and road conditions.

This study addresses these limitations through a comprehensive field implementation of continuous real-time monitoring on an operating haul truck. An integrated onboard system combining fuel-flow measurement, Global Navigation Satellite System (GNSS) positioning and payload sensing was installed on a Komatsu 785HD truck that operated continuously over 90 days, generating a dataset of 1780 complete haul cycles under normal production conditions. This approach provides three main contributions: (i) simultaneous quantification of operational (payload, speed, idle time), behavioural (driver style) and environmental (weather, road condition) factors within a single analytical framework, allowing direct comparison of their relative impact on SFC; (ii) spatial mapping of SFC across the actual pit geometry to reveal energy-intensive road segments and provide a visual diagnostic tool for prioritising maintenance and design improvements; and (iii) demonstration of a practical monitoring methodology that is transferable to other vehicles and mine sites and can underpin data-driven fuel-management programmes. Together, these contributions quantify key controllable drivers of fuel inefficiency and show that many improvements can be implemented at relatively low cost to achieve measurable fuel savings.

Objectives and case study

Building on the monitoring framework described above, this study focuses on identifying the controllable factors that have the largest impact on fuel consumption and where they occur along the haulage network. The overarching objective of this research is to quantify the controllable operational, behavioural and environmental drivers of haul truck fuel consumption and identify practical opportunities for fuel reduction in open-pit mining. This objective is addressed through four specific research questions:

Operational factors: What is the relative contribution of route characteristics (gradient, distance, surface condition), payload variation and travel speed to overall fuel consumption?

Behavioural factors: How does driver behaviour influence fuel consumption, and what is the magnitude of variation between operators?

Environmental and conditional factors: How do weather conditions and operational delays affect fuel consumption?

Spatial patterns and actionable insights: Where along the haulage network is the most highest energy-intensive zones, and what interventions can reduce fuel consumption?

Addressing these questions requires continuous monitoring of fuel flow, payload, position and speed under normal production conditions over an extended period that enables statistical analysis of both cycle-level and segment-level fuel consumption patterns.

The case study was conducted at a large open-pit operation exploiting a porphyry copper deposit. The operation required the removal of substantial quantities of waste rock to access the ore zones, which led to long haulage distances between the active benches, the primary crusher and the waste dumps. Annual production was approximately 10–15 million tonnes of combined ore and waste, necessitating the movement of large material volumes using a fleet of diesel trucks and loading units operating on a multi-bench system of ramps and haul roads. The pit reached depths of more than 400 m, and trucks were required to travel along long, steep ramps that imposed high fuel demand, particularly for loaded uphill hauls. The haulage fleet at the mine included Komatsu 785HD and 325HD, and Belaz trucks, and various support vehicles. Among these, the Komatsu 785HD truck, with a nominal payload of approximately 95 tonnes and a gross vehicle weight exceeding 150 tonnes when fully loaded, was selected as the instrumented vehicle for this study. It was representative of the larger-capacity fleet and operated along the main ore-haul circuits between the loading points and the crusher. Focusing on a single truck allowed detailed and consistent monitoring while maintaining manageable installation and data-logging complexity. Haul routes consisted of connections between multiple loading zones in the pit and the primary crusher or waste dumps, linked via ramps with varying gradients. Some road segments were relatively level, while others featured inclines exceeding 10%. Road conditions varied depending on maintenance scheduling, weather and traffic cycles, influencing rolling resistance throughout the haul network.

Operating conditions were characteristic of large open-pit mines, including multiple shifts per day, seasonal temperature variations and episodic precipitation. These environmental factors affected both road surface conditions and engine performance. Capturing fuel consumption in such a dynamic environment required collecting data over an extended period, ensuring coverage of diverse operational scenarios.

Materials and methodology

System architecture

An integrated onboard monitoring system was designed to capture, in real time, the key variables affecting truck fuel consumption. The system comprises the following components.

The onboard monitoring system (see Figure 1) integrates multiple sensors and communication modules to capture all variables affecting truck fuel use. Fuel flow meters installed on the supply and return lines measure real-time and cumulative fuel usage, allowing net consumption to be calculated. A GNSS receiver mounted on the cabin records position, elevation, speed and time, enabling reconstruction of haul routes and gradients. Payload sensors embedded in the suspension or frame estimate gross vehicle weight and, after subtracting the empty weight, provide accurate payload values for each cycle. All sensor signals are collected and time-stamped by an onboard data logger, then periodically sent via wireless communication to a central server. The server and analysis software validate, store and process the data using tools (MATLAB and Python) to compute SFC and perform detailed statistical analysis.

Figure 1.

Architecture of the designed system.

The system is designed to operate continuously during normal production, requiring minimal intervention from the operator. Visual indicators on the dashboard can be used to verify proper operation of the sensors and the communication link.

Sensor installation and calibration

Careful installation and calibration of sensors are essential to obtain reliable measurements.

The laboratory setup illustrated in Figure 2 shows the calibration of flow-meter used. Fuel was fed from a storage tank through the flow meter, and the discharged volume was collected in a graduated container. A voltage meter was connected to the flow meter to continuously record the output signal generated by the sensor during the test. By simultaneously measuring the actual volume collected over a known time period and the corresponding voltage output, a precise relationship between output voltage and flow rate was established. This procedure ensured accurate calibration of the flow meter before installing it on the haul truck for field measurements.

Figure 2.

Schematic of the laboratory calibration setup.

To ensure accurate measurement under our local operating conditions and to eliminate uncertainty, the flow meters had to be re-calibrated experimentally before data collection. As shown below, the calibration curves do not pass through the origin because, upon engine ignition, the flow-meter circuitry produced a baseline (offset) voltage, even when no fuel was flowing. This initial electrical noise and stabilisation current could not be eliminated due to the design of the flow-meter electronics. To address this, the offset voltage was quantified during calibration and subsequently subtracted from all field measurements to ensure that only the flow-dependent voltage was used in calculating the fuel flow rate. This correction ensured reliable and accurate operational measurements despite the inherent offset. And the equation for fuel consumption is calculated based on:

$$V=a*\mathrm{Flow}\mathrm{Rate}+b_{\mathrm{offset}v}$$

Figures 3 and 4 illustrate the calibrated relationships between the flow-meter output voltage and the measured fuel flow rate. In both cases, the data exhibit a strong linear trend (99%), demonstrating that output voltage is a reliable predictor of instantaneous fuel flow. The high coefficients of determination indicate an excellent fit, confirming that the calibrated models can accurately convert voltage readings into fuel consumption values with minimal error. Fuel flow meters were installed on straight sections of the fuel lines to minimise disturbance to the flow profile. After installation, the flow meters were calibrated again by measuring the fuel consumed during controlled fuelling events and comparing the volume recorded by the meters to the volume indicated by the fuel dispenser. Payload sensors were calibrated using known reference loads. The truck was loaded incrementally with known tonnages of ore, determined by the loading unit's calibrated bucket by truck scales. The corresponding sensor outputs were recorded, and a linear regression was fitted to obtain calibration coefficients. The empty truck was also measured to confirm the baseline signal for zero payload. Figure 5 shows the calibration curve used to convert the load-cell output voltage into the corresponding truck payload. The data points follow a clear linear trend, indicating a strong and consistent relationship between sensor voltage and measured payload at 99% accuracy across the tested range.

Figure 3.

Linear calibration curve between output voltage and fuel flow rate (supply line).

Figure 4.

Linear calibration curve between output voltage and fuel flow rate (return line).

Figure 5.

Calibration curve between output voltage and payload.

The GNSS unit was checked against known survey benchmarks to confirm positional accuracy, and time synchronisation between the GNSS and the onboard data logger was verified. Once calibration was complete, test cycles were run to ensure that fuel use, payload and position were recorded consistently and that the data streams could be merged correctly.

Data acquisition and processing

The monitoring campaign was conducted over a total of 90 days, covering 150 production shifts and recording 1780 complete haul cycles. This produced a large and diverse dataset capturing a wide range of operating conditions, including different weather events, shift patterns, loading locations, road gradients and driver behaviours. The scale of this dataset provides strong statistical reliability and allows detailed analysis of both cycle-level and segment-level SFC patterns across the mine. During the monitoring period, the instrumented Komatsu 785 truck was operated on its usual haul routes, loading ore and waste under normal production conditions. The system recorded fuel flow, position, speed and payload at regular intervals, yielding a time series of multi-variable data for each operating shift.

Data processing involved several steps:

Data cleaning: Obvious sensor errors, such as missing values, unrealistic spikes or communication dropouts, were identified and removed or corrected using interpolation where appropriate.

Cycle segmentation: The continuous time series was segmented into individual haul cycles consisting of loading, travel loaded, dumping, travel empty and idle periods. Cycle boundaries were identified using a combination of position, speed, payload and fuel flow patterns.

Route characterisation: For each cycle, the haul distance, elevation change and average gradient were computed from the GNSS data. Segments with distinct gradient ranges (e.g., uphill, level, downhill) were identified.

Fuel aggregation: For each cycle and for each segment within the cycle, fuel consumption was integrated over time to obtain the total fuel used and was computed as:

$$\mathrm{SFC}={\displaystyle \frac{m_{\mathrm{Fuel}}}{m_{\mathrm{material}}*l_{\mathrm{distance}}}}$$

where $m_{\mathrm{fuel}}$ is the total fuel mass consumed (gr) and $m_{\mathrm{material}}$ is the hauled material mass (t) and $l_{\mathrm{distance}}$ is travelled distance of truck over the considered cycle or segment. In this study, SFC is therefore expressed in units of gr/t.km.

• Factor attribution: Each cycle and segment was annotated with values or classes for the key factors of interest (e.g., gradient band, payload band, speed band, tyre pressure status, operator ID, weather category).

The processed data were then ready for statistical analysis to explore how SFC varies with these factors.

Results

The monitoring campaign produced a substantial data set covering many operating hours and numerous haul cycles under a range of conditions. The data included cycles along different routes, with varying gradients, payloads and speeds, and under both favourable and adverse weather conditions. This allowed meaningful analysis of how SFC responds to different operating regimes.

In broad terms, SFC values were found to vary considerably between cycles, reflecting the combined influence of route characteristics, loading practice, driver behaviour and environmental effects. Some cycles exhibited relatively low SFC, indicating efficient operation, while others showed much higher SFC, pointing to potential inefficiencies or difficult conditions.

Influence of road gradient

The spatial distribution of SFC across the haul-road network provides a valuable insight into how road geometry, gradient and surface condition influence truck energy demand. By mapping fuel-use intensity directly onto the pit layout, the analysis highlights where trucks consistently burn more fuel and identifies segments of the network that contribute disproportionately to overall haulage cost. The following figures present SFC variations along the main routes to the crusher and waste dump, revealing distinct high- and low-consumption zones that can guide targeted road maintenance and operational improvements.

Figure 6 illustrates the variation of SFC along the haul roads leading to the primary crusher. Road segments are classified into two fuel-consumption ranges, shown in green (65–70 gr/t.km) and yellow (70–75 gr/t.km). The green sections represent relatively fuel-efficient areas where trucks experience lower rolling resistance and smoother geometric alignment, resulting in reduced engine load. In contrast, the yellow portions correspond to haul segments with elevated fuel consumption, typically associated with steeper gradients, tighter curves or rougher road surfaces. The pattern clearly shows that even within a single haul route, fuel demand fluctuates according to local road and terrain characteristics. This visualisation demonstrates the usefulness of spatial SFC mapping as a diagnostic tool for identifying high-cost zones and prioritising operational or maintenance interventions.

Figure 6.

Fuel consumption on roads leading to the crusher.

On the other hand, Figure 7 shows the fuel-consumption profile along the haul roads that lead to the waste dump. Compared with the crusher route, these roads exhibit a broader range of SFC values, represented by yellow (65–70 gr/t.km), orange (70–75 gr/t.km) and red (75–80 gr/t.km) segments. The widespread presence of orange and red sections indicates that waste-dump routes impose heavier energy demands on trucks, which are likely due to longer uphill travel, greater elevation gain or poorer road quality. The concentration of high-SFC zones near the upper benches and dump approach suggests that road gradient is the dominant driver of fuel consumption in this area. These spatial patterns reinforce the importance of targeted road maintenance, slope optimisation and traffic-flow management to minimise fuel penalties on heavily used waste-haul routes.

Figure 7.

Fuel consumption on roads leading to the waste dump.

Analysis of SFC as a function of road gradient confirmed that uphill hauling was significantly more fuel-intensive than level or downhill travel. As shown in Figure 8, for segments with higher positive gradients, trucks required greater engine power to overcome the component of vehicle weight acting against motion, which led to increased fuel flow. The effect was particularly pronounced when gradients approached the upper limits recommended for haul-road design. Downhill segments, by contrast, generally exhibited lower SFC, provided that truck speeds were controlled and excessive braking was avoided. However, very steep downhill sections could also be problematic, as they required frequent braking and engine retarding, which increased fuel use and caused additional wear. The results highlighted the importance of good haul-road design and maintenance, including limiting maximum gradients, smoothing vertical curves and avoiding unnecessary elevation changes. Gradient had a statistically significant impact on SFC, with an average increase of 0.95 gr/t·km for each 1% increase in road gradient.

Figure 8.

Upward vs downhill roads specific fuel consumption (SFC).

Influence of payload

Payload analysis revealed that both under-loading and over-loading are detrimental to fuel efficiency. Figure 9 illustrates the variation of SFC with respect to hauled material tonnage for both upward and downhill travel. In both cases, the SFC curve exhibits a characteristic U-shaped pattern, indicating that fuel efficiency improves as the truck approaches an optimal payload range and then decreases once the load becomes either too low or too high. For uphill travel, SFC is consistently higher due to the increased engine load required to overcome the road gradient. Downhill conditions show lower SFC across all payloads, as gravitational forces assist vehicle movement and reduce fuel demand. The minimum SFC occurs at approximately 92–97 tonnes for both directions, highlighting the existence of an optimal payload range where the truck operates with maximum fuel efficiency. An intermediate payload band close to the truck's rated capacity was identified as providing the best balance between productivity and fuel efficiency. Maintaining payloads within this target band requires consistent loading practice, accurate payload measurement and appropriate dispatching policies. The presence of the onboard payload sensor and the data collected in this study provide a basis for monitoring and improving loading accuracy over time.

Figure 9.

Influence of payload in both upward and downward roads.

Influence of speed and traffic conditions

The relationship between speed and SFC is complex. At very low speeds, as may occur in congested traffic, poor road conditions or during frequent stops, fuel consumption per tonne-kilometre tends to be high because the engine operates inefficiently and much of the fuel is burned while the truck is not covering any significant distance. At very high speeds, fuel consumption may also rise due to increased aerodynamic drag and more frequent acceleration and braking events.

The monitored data showed (see Figure 10) the combined effect of truck speed on SFC and hauled material tonnage. As speed increases, the hauled material per hour initially rises sharply due to shorter cycle times, reaching a plateau at higher speeds. In contrast, SFC follows a U-shaped trend: fuel efficiency improves at moderate speeds but deteriorates at both low and high speeds. At very low speeds, inefficient engine operating conditions increase fuel use per tonne, while at high speeds, elevated engine load, rolling resistance and aerodynamic drag cause SFC to rise again. The interactions of these trends highlight the existence of an optimal speed range where hauled material is maximised and SFC is minimised simultaneously, offering a rational operating window for improving overall haul-truck productivity and fuel efficiency.

Figure 10.

Effect of truck speed on specific fuel consumption (SFC) and hauled material rate.

Operator behaviour

Differences in driving style among truck operators were reflected in the fuel consumption data. For the analysis of driver behaviour, operating data were grouped into three driving-style categories: normal, passive and aggressive. Classification was based on speed and acceleration patterns derived from the GNSS data, supported by the expert knowledge of site supervisors. Passive driving was characterised by lower average speeds and gentle acceleration, with frequent operation below the recommended speed band. Aggressive driving showed higher average speeds, more frequent operation near the upper speed limit and larger acceleration peaks, indicating rapid acceleration and braking. Normal driving fell within the recommended speed band, with moderate and relatively smooth acceleration/deceleration behaviour. Operators who accelerated smoothly, avoided unnecessary idling, maintained speeds within recommended limits and anticipated traffic conditions tended to achieve lower SFC than those who frequently accelerated and braked sharply or left the engine idling for extended periods. In some cases, differences in average SFC between operators on similar routes and under similar conditions were substantial. These findings underline the value of targeted operator training and feedback. By sharing fuel performance indicators with operators and highlighting best practices, mine management can encourage more fuel-efficient driving habits. Over time, such behavioural changes can yield significant cumulative fuel savings. The results clearly demonstrate the strong influence of driving behaviour on both SFC and hauled material over the 50-day monitoring period. As shown in Figure 11, trucks that operated under aggressive driving conditions consistently exhibited the highest SFC, ranging from approximately 68–75 gr/t.km. This pattern reflects the increased fuel demand associated with rapid acceleration, hard braking and frequent speed fluctuations. In contrast, passive driving produces intermediate SFC values (around 58–75 gr/t.km), while normal driving shows the most stable and efficient behaviour, maintaining SFC values predominantly between 60 and 68 gr/t.km. The clear and persistent separation between the three curves indicates that driver behaviour alone can account for substantial variation in fuel efficiency, even when trucks operate on similar routes and under comparable environmental and operational conditions.

Figure 11.

Effect of driving behaviour on specific fuel consumption (SFC).

A similar behavioural pattern for hauled material is observed in Figure 12. Trucks driven aggressively consistently achieve the highest hourly haulage rates (270–290 t/h), primarily due to higher average speeds and reduced cycle times. However, these productivity gains come at the cost of significantly higher fuel consumption, as shown in the SFC plot. Trucks operating under normal driving conditions maintain a moderate and stable haulage rate (265–280 t/h), while passive driving results in the lowest throughput (260–280 t/h). When considered together, the two plots highlight a critical operational trade-off: while aggressive driving may increase short-term productivity, it does so with a disproportionate increase in fuel intensity, reducing overall energy efficiency. Conversely, normal driving achieves a more favourable balance between fuel use and production rate, suggesting that controlled driving strategies can optimise both cost and efficiency in haulage operations.

Figure 12.

Effect of driving behaviour on material haulage.

Idle time, dumping delays and engine warm-up

Non-productive periods such as idling at loading or dumping points, waiting in queues or performing extended engine warm-up at the start of shifts were identified as important contributors to fuel consumption. While some idle time is unavoidable, the data revealed cases where trucks spent considerable time idling with little or no productive activity.

Engine warm-up practices varied between operators, with some allowing the engine to idle for only a few minutes before commencing work and others waiting much longer. Given that an idle truck engine still consumes a measurable amount of fuel, excessive warm-up translates directly into wasted fuel. Establishing standard warm-up procedures based on engine manufacturer recommendations and local climate conditions can help reduce this waste.

The distribution of idle time across different haulage activities reveals that operational delays represent a major source of unproductive fuel consumption in the mining cycle. As shown in Figure 13, loading queue delays dominate the idle-time structure, accounting for 45% of total non-productive time. This highlights significant congestion at the loading points and indicates a mismatch between shovel productivity and truck arrival rates. The crusher queue is the second-largest contributor, representing 22%, followed by waste-dump queue delays of 17%, both of which suggest that downstream bottlenecks also play a substantial role in reducing haulage efficiency. Additionally, unexpected delays such as short operational interruptions, road blockages or temporary standstills constitute 8% of total idle time, while driver-related faults contribute another 8%. Together, these results demonstrate that nearly all idle-time categories stem from avoidable or partially controllable operational factors, indicating a strong opportunity for system-wide optimisation through better dispatching, improved coordination between loading and dumping units as well as targeted operator training.

Figure 13.

Breakdown of total truck idle time during haulage operations.

Similarly, efforts to streamline dispatching, reduce queuing at crushers and dumps as well as coordinate loading more effectively can decrease the time that trucks spend idling or moving slowly in congested areas, thereby improving overall fuel efficiency.

Weather conditions on fuel consumption

Weather conditions, particularly rainfall, were found to exert a noticeable impact on haul-truck fuel consumption. These data were collected directly on-site through daily operational observations recorded by mine personnel. Rain events were identified based on visible precipitation during the shift and confirmed through supervisor logs. A day was classified as ‘rainy’ if rainfall occurred at any point during active truck operations, regardless of intensity or duration. Figure 14 presents the daily average SFC over a 60-day monitoring period. A clear pattern emerges during rainy periods, SFC consistently increases to its highest values in the dataset, reaching levels above 77–78 gr/t.km, compared with typical dry-weather values of 73–76 gr/t.km. This behaviour aligns with known physical mechanisms: rainfall reduces haul-road surface quality, increases water content in load materials and elevates rolling resistance, all of which force trucks to operate under a higher engine load and reduced traction. These combined effects result in greater fuel burn per ton of hauled material, even when travel speed and payload remain unchanged.

Figure 14.

Average daily specific fuel consumption (SFC) over 60 days, with rainy days (red).

The repeated spikes during each rain event also indicate that the impact of rainfall is immediate and persistent, with SFC rising sharply on wet days and returning to normal once roads dry. This suggests that weather-related fuel penalties are not merely random fluctuations, but systematic responses to degraded road conditions. Furthermore, the magnitude of SFC increases during rainy days to around 2–3 gr/t.km above the baseline, which is a substantial fuel penalty when extrapolated across an entire fleet and a full haulage cycle. These findings reinforce the importance of proactive wet-weather haul-road management, such as improved drainage, rapid grading and surface reinforcement to minimise rolling resistance during rain events. They also highlight the need to integrate weather data into short-term dispatching and fuel-efficiency planning, ensuring that production and truck allocation strategies account for the operational constraints imposed by adverse weather.

Shift-level performance and potential savings

By aggregating SFC and fuel consumption over whole shifts, the study provided a picture of how fuel is spent across different operational activities. A significant portion of total shift fuel was associated with hauling material uphill, as expected, but another notable portion corresponded to idle and low-speed operation. This breakdown helps mine management prioritise fuel-saving initiatives.

Figure 15 illustrates the proportional distribution of total fuel consumption across the main segments of the haulage cycle. The largest share, 36%, occurs during travel from the shovel to the waste dump, reflecting the long uphill distances and heavy payloads typically associated with waste-haul routes. The next major contributor is the shovel-to-crusher segment (22%), followed by the return travel from the crusher to the shovel (17%) and the travel from the waste dump back to the shovel (13%), highlighting the significance of empty-haul movements in the overall fuel demand. Operational delays also contribute meaningfully, with idle time, including loading and dumping delays, accounting for 11% of total fuel use. Warm-up periods represent a small but non-negligible portion at 1%. Overall, the distribution underscores that both loaded and empty hauls dominate fuel consumption, and that reducing cycle delays and optimising haul-road conditions could yield measurable fuel savings.

Figure 15.

Distribution of total fuel consumption across haulage-cycle segments.

Discussions and conclusions

The findings of this case study demonstrate that systematic monitoring of truck fuel consumption can reveal both obvious and subtle inefficiencies in haulage operations. Some of these inefficiencies, such as steep gradients and long hauls, are inherent to the mine layout and may be difficult to change in the short term. Others, including poor tyre maintenance, inconsistent loading, unnecessary idling and suboptimal driving habits, can be addressed more readily through improved procedures, training and maintenance.

From a management perspective, one of the most valuable aspects of the monitoring system is its ability to convert previously unobserved behaviours into quantitative metrics. Instead of relying on general perceptions about which routes or drivers are more or less efficient, mine managers can use actual data to identify high-consumption routes, assess the impact of road maintenance, compare operator performance and evaluate the effectiveness of interventions over time.

The methodology described here is also compatible with broader fleet management initiatives. Fuel consumption data can be integrated with production and maintenance databases to support decisions about equipment replacement, dispatching strategies and even potential introduction of alternative fuels or hybrid technologies. As mines move towards more data-driven and automated operations, such integrated monitoring systems will become increasingly important components of digital mine architectures.

This article has presented a comprehensive case study on the monitoring and optimisation of fuel consumption for haul trucks in a copper mine using an onboard data acquisition system. The main conclusions are as follows:

It is technically and operationally feasible to equip a large haul truck with an integrated system that measures fuel flow, payload, position and speed continuously under normal production conditions, with minimal disruption to operations.

Specific fuel consumption, defined as the amount of fuel used per tonne, provides a meaningful metric for comparing fuel performance across different routes, payloads and operating conditions.

Specific fuel consumption is strongly influenced by road gradient, payload, travel speed, weather, operator behaviour, idle time and engine warm-up practice. Uphill hauling, under- and over-loading, poor tyre maintenance, frequent stop-start driving and long idle periods all contribute to higher fuel use.

Many of the factors that increase fuel consumption can be mitigated by low-cost operational measures, such as improving road design and maintenance, enforcing target payload and speed bands, maintaining correct tyre pressure, standardising warm-up procedures and providing operator training and feedback.

The monitoring approach developed in this study is transferable to other trucks and other mines. When extended across the fleet, it can form the basis of a mine-wide fuel management programme that reduces operating costs and environmental impacts.

Based on these findings, several practical actions can be implemented to improve fuel efficiency at open-pit operations. Expanding the monitoring system to additional trucks and other diesel-intensive equipment would provide a comprehensive view of mine-wide fuel behaviour. Regular tyre pressure maintenance supported by real-time tyre pressure monitoring systems can help minimise rolling resistance and prevent unnecessary fuel loss. The existing payload measurement system should be used proactively to detect under- or over-loading trends and provide direct feedback to loading operators. Haul roads and ramps may also require geometric refinement, including reducing steep gradients, smoothing transitions and improving surface quality to lower engine load. Operationally, standardised procedures for warm-up, idling and speed control, supported by targeted operator training, can reduce behavioural variations in fuel use. Finally, integrating fuel-consumption data with production and maintenance records will strengthen decision-making for fleet scheduling, equipment allocation and long-term mine planning.

As well, future work could include the development of predictive models that estimate fuel consumption as a function of the monitored variables, allowing simulation of ‘what-if’ scenarios for alternative road designs, dispatching strategies or equipment choices. In addition, combining fuel consumption monitoring with emissions modelling would provide a basis for evaluating and managing the environmental footprint of haulage operations.