Sign in

Fuel event analysis

Last updated on 

The Rangr system performs complex data analysis in order to detect fuel events such as fills and drains. Here, we give you an overview of the analysis process, why Rangr is far more reliable than any other fuel monitoring systems available, factors affecting the analysis process and what to do if you think detected events may not be true or if true events did not get detected.

Introduction

In the Rangr system, the fuel sensor installed on the vehicle sends the fuel level readings to our servers via the GPS unit. This data is generally collected at an interval of every few seconds. Along with the fuel readings, the GPS sends readings of many other parameters with respect to the state of the vehicle - for example speed, ignition, battery voltage, spatial orientation etc. 

Data analysis, fuel events and state maintenance

As part of data analysis, the system first removes noise (i.e. abrupt/sudden changes) from the data it receives. Then, the fuel level data is combined with other vehicle state data, and analysed in realtime to evaluate the fuel consumption state of the vehicle. Fuel events  e.g. "fuel is being filled", "fuel is being drained" etc., are reported on the basis of this analysis.

As a simple example: a fuel event is reported when there is change in the fuel volume, greater than the set threshold, and that change of volume is consistently maintained for a given period of time (called “state maintenance”). 

Rangr's reliability - going beyond "sudden changes"

It is important to note the difference in Rangr’s analysis process compared to other systems:  most of the other systems only detect a “sudden change” (spike) and report a fuel event without any other context, which makes them unreliable to begin with. 

Whereas, Rangr waits to see what happens after the event, and takes into consideration the overall state of the vehicle in order to make the best possible decision on whether the event is true of false. This makes Rangr much more reliable than the other systems.

Factors affecting fuel event detection and unavoidable tradeoffs

There are many factors that could result in change of fuel level in the tank, which may or may not be known to the system via the data it receives. For example, fuel consumption is affected due to several factors like change in vehicle load, driving behaviour, vehicle maintenance status, vehicle standing on an inclination for a long period of time, etc.

Additionally, the fuel level sensor has an accuracy of 99% of the full tank volume. Therefore, for a tank of e.g. 400 litres, there will be an inaccuracy of 3-4 litres. Additional inaccuracy may be introduced due to human factors involved during the installation process, e.g. during calibration process of the tank.

Keeping all of the above in mind, it is necessary to understand that the system cannot be 100 percent accurate - there will be instances when false events are generated by the system as well as some true events being missed. This becomes more pronounced due to the following factors:

  1. The threshold volume for fuel event detection is lowered 
  2. The time window for data noise removal is changed 
  3. The time window for event state maintenance is changed

Hence, there are unavoidable tradeoff between accuracy of events detected and the factors listed above. For example:

  1. If the threshold is lowered, you must expect false events to increase. 
  2. If the time window for data noise removal is lowered, you must expect false events to increase. 
  3. If the time window for event state maintenance is lowered (e.g. if you need events to be reported as soon as they occur), you must expect false events to increase. 

Handling false and missed events: working around the above factors and tradeoffs

Unlike any other system in the market, Rangr evaluates fuel events in a very different, and much more reliable way.

Even though the system is much more reliable compared to others, we recognise that there are still a small number of instances when evaluation of events is inaccurate, as discussed above. To cover for this situation, Rangr provides you:

    1. The ability to control the above mentioned factors, depending on your requirements and keeping all the tradeoffs in mind. 
    2. All the necessary tools to analyse the events detected, or in rare cases missed by the system

The following articles go into the details of when and how to perform Fuel Event Analysis from the Rangr dashboard, and all the related tools and configurations provided:

NOTE: The following are applicable only with the Rangr Premium Fuel plan
  1. Initiating Fuel Event Analysis
  2. Understanding the fuel analysis graph
  3. Understanding analysis configuration
  4. Evaluation of a fuel event
  5. Changing an event incorrectly detected as true to false
  6. Finding missed events