The Fuel Volume NEW report provides information on the fuel volume for the specified objects for the selected time period.
The report consists of several sections:
- Graph of change in fuel volume
- Table with information on fillings
- Details by dates table with fuel and mileage calculations
- Assessing the quality of the fuel level sensor
The information in the generated report refers to the Refuelling event and includes such attributes as:
- Date and time
- Fuel volume
- The address where the filling event took place
You can also build a report using the platform API. Use plugin 95.
What affects the readings in the report
In order for the platform to calculate and display the information correctly, the fuel level sensor must be set up in advance:
- If you use a single fuel level sensor, make sure that the calibration table is filled out correctly according to the manufacturer's recommendations. In addition filling the calibration table, it is strongly recommended to add thresholds for emission filtering. To do this, go to the "Advanced" section (under the calibration table), enter the data in the "Ignore values" fields.
- If your vehicle has multiple tanks, each of the senors must be calibrated separately. After calibrating and specifying the thresholds, create an aggregate sensor by specifying the volume units in its settings: litres or gallons. (It is not recommended to use "Custom Option" in the sensor's Unit of Measure setting. This setting may result in incorrect display of information in the report).
For fuel level sensors that send data as a percentage of the total fuel volume, the calibration process must also be performed. For this, at least two values must be specified: the minimum and maximum fuel volume (100%) in litres or gallons.
After changing the sensor settings, you can build the report for the same period again with different settings and the changes will be applied to the recorded data.
How the fuel reading and report works
To provide information in the report, the platform does the following:
- Collects and stores raw data from devices and fuel sensors.
- Depending on the ignored values set in the sensor settings, it may filter out sensor readings with displaying the data. On its own, the platform does not make up changes, and records data as is
- The raw data is converted to user-specified values in the calibration table to generate the report. If the calibration table is not filled in, the platform will not perform any conversions and will display the stored values as received or the system will yield an error
- Based on the report parameters selected by users, the platform builds a graph of stored values. Fuel sensor settings, including threshold values, are applied to this graph
- Next, the platform calculates fillings for the table and adds the filling information to the graph
- The process of smoothing the graph is fully automated - all smoothing and filtering settings are applied automatically
Fuel volume change graph by sensor
Presents fuel level sensor readings in graphical form. Displays readings from the built-in sensor (if just one fuel level sensor is used) or from the aggregate sensor (if two sensors are used). If an aggregate sensor is used, the readings of the sensors in the aggregate sensor will not be displayed.
The graph will also show the fillings in the form of numbers. The numbers are in the order from the Filling table. The filling points is shown on the graph in the filling peaks.
Fillings table
This table will display information about the recorded fillings for the period.
Details by dates table
The table presents daily data, including details of fueling, fuel volume, consumption, and fuel consumption per 100 km.
- Date - the calendar day for which the calculations are made.
- Mileage, km/mi - the distance traveled by the vehicle on the given day.
- Count of refuelings - the number of refueling operations per day.
- Volume, l/gal - the total volume of fuel filled during the day.
- Consumed, l/gal – the actual fuel level consumption for the day. It is calculated using the formula: "initial fuel level + volume of refuelings - final fuel level".
- Consumption l/100km, gal/100mi – the actual fuel consumption per 100 km. It is calculated using the formula "(initial fuel level + volume of refuelings - final fuel level) / mileage * 100".
Quality of fuel sensor readings (LAB)
"LAB" refers to the experimental feature.
This section of the report provides information on the fuel level sensor reading quality assessment, which is calculated by analysing the noise of raw data from fuel level sensors, including OEM and separately installed fuel level sensors. This information will be useful to users who analyse fuel calculations and graphs, and also would like to receive more insights on the nature of the fuel data. For instance, common questions are: why the graph is uneven, or why it does looks like a saw.
In case, there is enough data to calculate the quality score of the sensor metrics, the algorithm will provide a quantitative and qualitative result:
The quantitative result is represented as a score from 1.0 to 10.0, where 1.0 is low quality of sensor performance, 10.0 is high quality of fuel sensor performance.
The quality score is presented in one of three quality levels: low, for sensors with scores between 1.0 and 3.9, medium for sensors with scores between 4.0 and 7.9, or high, with scores between 8.0 and 10.0.
If there is insufficient data for the calculation, for example, if the sensor has been installed recently, the system will display the message: "Insufficient data for quality assessment of fuel sensor readings. The most likely reasons may include: an insufficient number of vehicle trips for the selected period, too infrequent intervals of fuel data collection, or poor calibration of the sensor on the platform. The minimum recommended period for assessing the quality of readings is 14 days of active vehicle operation."
Also, the section "Quality of fuel level sensor readings" contains recommendations. The information provided can be used to make decisions, especially if the results of the calculations are not as expected. An example of what the Fuel Sensor Quality section looks like is shown in the figure below.
Insights from the Report
For instance, if we discovered that dishonest company employees involved in fuel fraud are increasingly abandoning "classic" siphoning methods, where fuel is pumped from the vehicle's fuel tank into a canister. Instead, most fuel fraud occurs directly at the fuel station. Thus, fueling data, including date and time, location, and volume of fuel, are becoming increasingly relevant for monitoring and controlling fuel usage.
The new report helps to obtain detailed information on fueling. It employs an adaptive algorithm that eliminates almost all false fueling events from the report, which mainly occur due to natural reasons (for example, fuel fluctuations when the vehicle moves on uneven roads).
The platform also implements the new algorithm for smoothing and filtering noise and outliers in the graph, allowing the platform to produce a user-friendly, readable charts without artifacts. This can be used to analyze changes in fuel level relative to mileage and identify patterns and anomalies in consumption.