# Fuel Management in Remote Sites: Reconciling Diesel When Every Litre is Tracked by Hand
TL;DR: On a remote site, diesel is a high-value asset that is consumed continuously and measured imprecisely — the only way to control it is to tie every litre issued to a piece of equipment, an hour-meter reading, and an expected-versus-actual consumption calculation, then systematically investigate any variance.
In an office in Luanda, a cash discrepancy is rare and easy to spot. In a camp four hours from the main road, with a generator running twenty hours a day and a fleet of tanker trucks, backhoes and compressors, diesel behaves differently: it is the one input being purchased, stored, transported, dispensed and consumed continuously, often without a single electronic reading anywhere in the chain. A paper logbook, a fuel point without a calibrated meter, and an operator who writes "full" instead of a number are the norm at many sites — and that is exactly where fuel gets lost, whether through measurement error, evaporation, diversion, or plain negligence in filling out the log.
This guide does not assume widespread fraud — there is no reliable, verifiable statistic on fuel losses in remote operations in Angola worth citing, and inventing a number would be worse than citing none. What is verifiable is the rising cost of every litre lost: in March 2025 the price of diesel in Angola rose 50%, from 200 to 300 kwanza per litre, after the government progressively removed fuel subsidies at the IMF's request (Bloomberg). That alone justifies treating diesel with the same accounting discipline as a high-value inventory item, not as an operating expense that gets topped up without questions.
Fuel as an asset, not an expense line
The most common design mistake is treating diesel as a cost line rather than a stock item. A stock item needs four things: an auditable inbound quantity (supplier delivery note, invoice, offload into the site's master tank), an auditable outbound quantity (an issuance record per piece of equipment), a running theoretical balance at any point in time, and a periodic physical reconciliation against that balance. Without all four, what exists is a tank that "seems" to have enough fuel — until it doesn't.
This is the same problem, structurally, that shows up in any critical field inventory — the same logic of physical count against theoretical balance applies to MRO parts inventory, where not knowing what is physically on hand always surfaces at the worst possible moment.
Designing the fuel issuance record
The issuance record is the central document of the entire system. Every line should capture, at minimum:
| Field | Why it's mandatory |
|---|---|
| Date and time of issuance | Enables interval analysis between fill-ups and detection of anomalous patterns (e.g. off-shift dispensing) |
| Equipment (unique ID) | Ties consumption to a specific asset, not a generic category |
| Operator who requested and operator who dispensed | Two names, never one — the basis of segregation-of-duties control |
| Hour-meter/odometer reading at time of issuance | Starting point for calculating consumption per hour worked |
| Litres issued (metered, not estimated) | Ideally from a pump with a calibrated counter, not a visual mark on the tank |
| Master tank dip reading before and after | Cross-checks pump output against the actual change in tank level |
| Location/site | Necessary when there is more than one dispensing point on the same site |
| Signature or PIN of the operator authorised to dispense | Closes the record as an auditable event, not an informal note |
Without the hour-meter reading on every line, the issuance record is just a cash book of litres — useful for knowing how much went out, useless for knowing whether what went out matched the work actually performed. That link between litre and hour worked is what turns an administrative record into a control instrument.
This is exactly where a purpose-built system beats a spreadsheet: it forces the critical fields to be filled before a record can close, validates the interval between the current and previous hour-meter reading (rejecting, for instance, a value lower than the last one), and keeps a complete history per piece of equipment without relying on manual discipline to copy figures between files.
Linking issuance to equipment and the hour meter
Every diesel-consuming piece of equipment — generator, welding genset, truck, backhoe, compressor, pump — needs its own record in the system with three fixed attributes: the theoretical consumption rate at its normal duty cycle (from the manufacturer's spec, or established from a historical baseline), its own tank capacity, and its current hour-meter/odometer reading.
Every fuel issuance automatically updates that equipment's running balance and sets a new reference point for the next hour-meter reading. The goal is that, at any moment, two questions can be answered without going to the field: how many litres has this equipment consumed over the last 100 hours worked, and is that figure within the expected range for its load profile.
For generators, this is a question with a known numerical answer. The industry-standard reference formula for hourly consumption is:
Consumption (L/h) = 0.25 × rated power (kW) × load factor
where 0.25 litres per kW per hour is the average specific fuel consumption of a diesel engine, and the load factor reflects the percentage of rated power actually being used (DEPCO Power Systems; Diesel Generator Tech). A 100 kVA generator (80 kW effective) at 75% load consumes roughly 15 litres/hour — and it is that number, not a generic site-wide average, that should be recorded against the equipment and used as the baseline for variance calculations. It's worth noting that fuel efficiency per litre worsens significantly below 70-75% load, which matters when sizing a generator for a camp with variable daily demand — oversizing "for safety" carries a direct cost in litres per kWh generated.
For vehicles and earthmoving equipment, the logic is similar but the reference metric is usually litres per engine hour or litres per kilometre, established from the equipment's own history over its first months of operation, since real-world consumption varies far more with terrain and load than it does for a stationary generator.
Expected versus actual consumption: the calculation that matters
The reconciliation itself is simple to state and hard to operationalise without a system: for each piece of equipment, in each period,
Variance = Actual recorded consumption − (Hours worked × Theoretical consumption per hour)
A persistent positive variance (more fuel issued than the recorded work would predict) is the classic signature of loss — through measurement error, evaporation, misuse, or theft. A persistent negative variance usually points to a faulty hour meter, under-recorded working hours, or actual consumption running above the manufacturer's spec (an engine due for maintenance, a clogged air filter, misadjusted injectors).
What makes the operational difference is the tolerance threshold and what happens when it's crossed. A well-designed system doesn't produce a monthly report nobody reads — it raises an exception the moment a piece of equipment's cumulative variance crosses, say, 8-10% of theoretical consumption over a seven-day window, and requires a logged justification before authorising the next issuance for that equipment. That turns reconciliation from a retrospective audit (which finds the loss three months later) into a preventive control (which catches it in the first week).
Reconciling the site's master tank follows the same logic at a different level: theoretical balance = previous balance + deliveries received − total dispensed by the pumps, checked against a physical dip performed by someone independent of whoever operates the pump. Any fuel management system that doesn't separate those two functions — who dispenses and who verifies the physical balance — is designing its own undetected exception.
Telemetry and sensors: reducing reliance on handwriting
Where infrastructure allows, tank-level sensors and CAN-bus modules on the most critical equipment eliminate much of the manual-reading error, feeding the system near-real-time data instead of records filled in at the end of a shift. The real engineering challenge isn't getting the data — it's designing telemetry ingestion without drowning the database in high-frequency readings nobody will individually review, and deciding correctly which aggregations and alerts should be computed on-site, since a satellite link at a remote camp is not guaranteed. That's also why the paper issuance log can't simply be digitised as an online form: the system has to keep working when the network drops mid-shift, syncing transactions once connectivity returns — the same edge computing problem in remote oilfields that applies to any field system in Angola, which decides what has to be validated locally before any sync.
Who can authorise, and who can see what
A fuel issuance record only has evidentiary value if nobody can alter it after it closes, and if it's possible to later prove who had authority to dispense diesel to a specific piece of equipment on a specific date. That isn't an abstract compliance concern — it's what separates a control system from a logbook anyone can rewrite after the fact. The foundation for this is the same role-based permissions and immutable audit trail design that any sensitive module of an industrial ERP requires: clearly defined roles (who requests, who dispenses, who physically verifies), and a change log that never deletes, only appends a new version with its author and timestamp.
FAQ
Does a fuel management system replace the manual tank dip?
No, and it shouldn't try to. The physical dip remains the source of truth against which the system's theoretical balance is validated. What the system replaces is the manual calculation of that theoretical balance and the spreadsheet reconciliation — cutting human error out of the part that can be automated, without removing independent physical verification.
Is it worth installing level sensors on every tank at a remote site?
It depends on the asset's value and the distance to the main road. On a large-capacity master tank at a camp with no fixed network, a sensor with periodic transmission pays for itself quickly against the cost of a verification trip. On small auxiliary tanks, a disciplined manual dip, logged in the system, is usually enough.
What happens when a piece of equipment's hour meter fails?
The system should block automatic theoretical-consumption issuance for that equipment and force a justified manual entry until the meter is replaced or recalibrated — it should never simply carry forward the last known value, because that masks exactly the kind of variance the control exists to catch.
Does this only apply to oil and gas operations, or also to construction and EPC sites?
The principle is the same at any remote site with generators and heavy equipment — construction site, logistics base, EPC yard. What changes is the number of dispensing points and the criticality of each piece of equipment, not the reconciliation logic.
A note on who's writing this
Wise Hustlers designs and operates its own energy-sector ERP, covering upstream, wells, production, projects and contracts, procurement, suppliers, inventory, MRO, maintenance, fleet, logistics, HSE, quality, HR, training, finance, tax and compliance — including fuel management modules like the one described above. If you're evaluating building this kind of control tailored to your operation in Angola, our custom software development page describes how we approach this type of project.