Modern heavy-duty trucking fleets generate massive volumes of real-time data through the SAE J1939 protocol on the vehicle CAN-Bus. At Strong Tower Consulting, we help motor carriers turn this raw ECM telemetry into actionable intelligence, particularly around fuel consumption anomalies that directly impact operating margins.
By performing edge ingestion of critical J1939 SPNs (Suspect Parameter Numbers) such as engine torque, fuel rate, idle percentage, vehicle speed, and transmission gear ratio, carriers can establish precise baseline fuel burn curves calibrated to payload weight, route topography, and weather conditions. Deviations from these baselines can signal fuel theft, aerodynamic issues, driver behavior that wastes fuel, or mechanical faults, and each flagged event should be confirmed against other evidence before action is taken.
- Fuel card data alone cannot separate legitimate consumption from theft. J1939 data shows second-by-second engine state.
- Core SPNs include 183 fuel rate, 512 and 513 engine torque, 84 vehicle speed, and 190 engine speed.
- The baseline is a regression on weight, grade, ambient conditions, engine duty cycle, and idle versus moving time.
- Edge scoring sends only exception events and summary statistics to the cloud, reducing cellular data costs.
- Baselines are built from 90+ days of historical telemetry correlated to load and route data.
Why J1939 Telematics Beats Traditional Fuel Card Reporting
Fuel card data alone cannot distinguish between legitimate consumption and theft. A driver can idle for hours at a rest stop or transfer fuel to a personal vehicle without triggering a fuel card exception. J1939 data provides second-by-second visibility into actual engine state.
J1939 engine data
- Second-by-second visibility into actual engine state
- Baselines calibrated to payload weight, route topography, and weather
- Objective, physics-based anomalies for coaching
Fuel card data alone
- Cannot distinguish legitimate consumption from theft
- Hours of idling or fuel transfer with no exception
- Simple idling thresholds that drive confrontational coaching
Key parameters we extract and analyze include:
- SPN 183 - Engine Fuel Rate (L/h or gal/h)
- SPN 512 - Driver's Demand Engine Percent Torque
- SPN 513 - Actual Engine Percent Torque
- SPN 84 - Wheel-Based Vehicle Speed
- SPN 91 - Accelerator Pedal Position
- SPN 110 - Engine Coolant Temperature
- SPN 171 - Ambient Air Temperature
- SPN 190 - Engine Speed
- SPN 247 - Engine Total Hours of Operation
- Transmission gear ratio and current gear (from TC1 or ETC2 messages)
Building a Fuel Burn Baseline Model
Using historical J1939 streams, we construct regression models that predict expected fuel consumption as a function of:
- Gross Combination Weight (GCW) derived from payload sensors or manifest data
- Route-specific rolling resistance and grade (integrated with digital elevation models)
- Ambient temperature and wind conditions
- Engine duty cycle (torque × RPM)
- Idle time versus moving time
expected_fuel = f(GCW, rolling resistance and grade, ambient temp and wind, torque x RPM, idle vs moving time)
anomaly when observed SPN 183 fuel rate deviates significantly from expected_fuel
theft signature: fuel card gallons > ECM fuel used (SPN 250) + tank level change, OR fuel level (SPN 96) drops while engine speed = 0 RPM
idle violation: 600 <= engine speed <= 900 RPM AND road speed = 0 AND coolant temp normal
Once the baseline is established, real-time anomaly detection algorithms running at the edge or in the cloud flag statistically significant deviations. These events are then correlated with GPS location, driver logs, and load information from your TMS (Truckmate, McLeod, or TMW).
| Anomaly Type | Typical J1939 Indicators | Operational Impact | Corroborating Evidence |
|---|---|---|---|
| Fuel Theft at Rest Stop | Fuel card gallons exceed ECM-reported fuel used plus tank level change, or fuel level drops while the engine is off | Direct loss of diesel inventory | Fuel card transaction, GPS location, tank capacity |
| Aerodynamic Drag (Damaged Fairing or Open Doors) | Elevated fuel rate at constant speed and torque with no change in grade or wind data | Increased fuel per mile, sized against the unit's own baseline | Equipment inspection, driver vehicle inspection report |
| Driver Coasting or Speeding | Excessive time in non-optimal gear ratios, frequent high accelerator pedal position | Increased fuel per mile and potential ELD HOS violations | GPS speed, posted limits, ELD logs |
| Mechanical Fault (Clogged Filter, Injector Issues) | High torque with disproportionately high fuel rate and elevated exhaust temperatures | Progressive efficiency loss and potential roadside breakdown | Diagnostic trouble codes, maintenance records |
| Idling Policy Violation | Extended periods with engine speed 600-900 RPM, zero road speed, coolant temp in normal range | Excess fuel burn and increased engine wear | ELD duty status, ambient temperature, idle policy exceptions |
Integration With Carrier Operations
The true power of J1939 fuel anomaly detection emerges when fused with your core business systems. At Strong Tower Consulting we design architectures that correlate CAN-Bus events with:
- EDI 204 load tenders and 214 shipment status updates
- Driver settlement calculations including fuel surcharge recovery
- IFTA fuel tax reporting and state mileage audits
- CSA SMS scores through integration with ELD telematics
- SQL-based data mining against your McLeod, TMW, or Truckmate database schemas
For example, a fuel theft event detected at a known rest stop can automatically trigger a review of the associated load's 210 invoice and driver pay calculation, ensuring the carrier is not absorbing the loss through incorrect fuel surcharge application or detention billing.
Edge vs Cloud Processing Considerations
Modern telematics devices can perform initial anomaly scoring at the edge using lightweight models, reducing cellular data costs. Only exception events plus summary statistics are transmitted to the cloud for deeper analysis, driver coaching workflows, and integration with your transportation management system. This approach is designed to scale for private fleets and for-hire carriers running 50 to 5,000 power units.
Implementation Roadmap
Strong Tower Consulting guides carriers through a structured four-phase approach:
- J1939 data validation and SPN mapping from your specific engine and transmission combinations
- Baseline model development using 90+ days of historical telemetry correlated to load and route data
- Real-time anomaly detection rules and exception workflow design integrated with your existing dispatch and safety platforms
- Continuous model refinement using SQL data mining techniques against your operational data warehouse
Frequently Asked Questions
How accurate is J1939-based fuel theft detection compared to tank level sensors?
ECM fuel-used data combined with engine state avoids some failure modes of aftermarket tank level sensors, such as calibration drift, sloshing, and tampering, while tank sensors show fuel leaving the tank when the engine is off. When both are available, we fuse them with fuel card transactions. Precision should be measured during a pilot by confirming each flagged event against fuel card records, GPS location, and the outcome of the investigation.
Can this solution work with our existing ELD and TMS platforms?
Yes. Our architecture is deliberately hardware and software agnostic. The design works with ELDs such as Geotab, Samsara, PeopleNet, and Omnitracs that expose J1939 data through their APIs, as well as McLeod, TMW, and Truckmate TMS platforms. The key is establishing clean, normalized data pipelines between the telematics stream, your SQL database, and operational workflows.
What is the typical ROI timeline for implementing CAN-Bus fuel analytics?
The timeline depends on data quality, prior fuel controls, and how quickly flagged events are acted on, so it should be measured rather than assumed. Record fuel cost per mile, gallons per mile by unit and lane, idle percentage, and confirmed theft events for a baseline period, then track the same figures after the baseline model and exception workflow go live. Because the baseline itself needs 90+ days of history, plan the comparison period accordingly.
How does this approach support driver coaching without being punitive?
By focusing on objective, physics-based anomalies rather than simple idling thresholds, coaching conversations become data-driven and less confrontational. Drivers can see exactly how their gear selection, speed choices, or idle habits affect fuel burn relative to the established baseline for that specific load and route.
Strong Tower Consulting, LLC was founded in Minnesota in 2008 by Scott Elliott. In business for more than 18 years, we provide independent systems architecture and technology advisory services exclusively to motor carriers, freight brokers, and private fleets.
Strong Tower Consulting, LLC and Exspeedite LLC are separately organized sister companies under common ownership. Carriers never have to buy our software. Our only product is better decisions.
Turn J1939 data into fuel decisions.
Strong Tower Consulting provides independent systems architecture, feasibility audits, and engineering roadmaps for motor carriers and freight brokers. Contact our advisory team or call 763-244-1373 to discuss your operational data and evaluate practical implementation paths for your fleet.