Fuel monitoring is a chain: the liquid changes, the sensor measures a physical property, the device samples the signal, software converts it to a level or volume, and an operator decides whether the pattern needs action. Error or uncertainty at any stage can produce a convincing but incorrect conclusion.
The practical goal is not to label every change as theft. It is to produce consistent measurements, identify events that deserve review, and preserve enough evidence for a supervisor to distinguish consumption, refuelling, movement effects, maintenance and unexplained removal.
1. Inspect the tank and operating environment
Record tank dimensions, shape, material, capacity evidence, internal partitions, filler position, available sensor location and the full range of vehicle tilt or vibration. On generators and fixed tanks, note mounting level and return lines. On vehicles, inspect whether paired tanks equalise and whether fuel can move between compartments. These details determine whether one probe and one calibration table can represent the usable volume.
2. Select and install the sensing method
Capacitive level sensors such as the Omnicomm LLS family are documented for measuring fuel level and transmitting readings to monitoring terminals. Suitability still depends on fuel type, tank geometry, measurement length, interface compatibility and installation rules. The sensor must be cut, mounted and electrically connected according to the relevant manual.
Keep installation records: sensor serial number, interface, probe length, device input, seal points and photographs. Verify stable power and grounding. Route cables away from damage and interference. A physical installation that cannot be audited later weakens every event that follows.
3. Build a calibration table
A raw sensor value is not automatically a volume. Calibration maps readings to known additions across the tank’s usable range. OIML R 85 describes metrological and technical requirements for automatic level gauges used in fixed storage tanks; it is a useful reference for understanding why measurement conditions and traceability matter, even when a fleet installation is not an OIML-certified custody-transfer system.
Calibrate the tank in its normal orientation using a known reference quantity and regular increments appropriate to its geometry. Allow the liquid to settle, record the reading at each step, and protect the completed table from casual editing. Omnicomm’s CalibrateX documentation also presents calibration as a distinct process for converting sensor measurements into a tank table.
Calibration acceptance checklist
- Tank and sensor identity recorded.
- Known reference quantity and increment documented.
- Readings allowed to settle at each calibration point.
- Empty and usable-full conditions verified safely.
- Completed table protected and linked to the installed device.
4. Define fuel events carefully
| Observed pattern | Possible explanations | Evidence to review |
|---|---|---|
| Sustained rise | Refuelling, tank equalisation or configuration change. | Location, ignition, supplier record, amount and stable readings before/after. |
| Sharp drop | Removal, leak, tank transfer, slope, sensor disturbance or power interruption. | Time series, movement, voltage, site access, physical inspection and later level. |
| Repeated oscillation | Fuel movement, loose connection, unsuitable filtering or mechanical contact. | Speed, road condition, wiring, mounting and raw readings. |
| Unexpected flat line | Stable level, stuck signal, disconnection or device input fault. | Consumption context, diagnostics, power and a controlled level change. |
5. Investigate before classifying
An alert should open a review, not close it. Preserve the raw time series and compare it with GNSS position, ignition, motion, refuelling documentation, access records and the next stable level. Record the reviewer, conclusion and reason. If the evidence is incomplete, classify the event as unresolved rather than forcing a definitive label.
6. Measure the monitoring process
Useful process measures include the proportion of sensors reporting, calibrations with complete records, alerts reviewed within the agreed window, confirmed false positives and unresolved events. A claimed reduction in loss requires a comparable baseline, consistent inclusion rules and independent operational records—not only the monitoring platform’s alert count.