Predictive analytics

15 Nov 2018 By Russell Dupuy

Predictive analytics collaboration in Retail Petroleum

Environmental Monitoring Solution Pty Ltd [EMS] in collaboration with La Trobe University has solved the problem of identifying unexpected fuel system configuration changes, as part of a broader research engagement.

The joint research collaboration was made possible via client collaboration through deployment of the award-winning Fuelscan® and Fuelsuite® Industry 4.0 technologies that monitor a number of sensory inputs at retail petrol stations in real time.

Fuelscan® live fact data is untouched by human intervention allowing Fuelsuite® artificial intelligence algorithms to detect when a change occurs in dispenser hose to tank mapping. Often this is a valid maintenance or upgrade event, however, the resulting impact to fuel reconciliation and/or sales data can be significant and take time to resolve.

The outcome may lead to consumers either not receiving the fuel they expected or paying too much for a lower grade fuel, whilst also creating confounding results for the petroleum retailer. In some countries, this can be an intentional manipulation resulting in fraud to the consumer.

Traditional reactionary maintenance would respond with a contractor site visit to inspect the fuel system that often would lead to a second site visit from the POS/BOS technician to simply reconfigure the hose to tank mapping. Adding up the cost becomes expensive for what is often a simple correction when it can be pinpointed through smart data analytics.

Another key component of this research was to detect when an Automatic Tank Gauge [ATG] float is ‘stuck’, which can lead to confounding fuel reconciliation issues and the more ominous concern of loss of containment through overfill.

The next stage of the project which also engages with the Royal Melbourne Institute of Technology [RMIT] seeks to model, identify and quantify fuel transportation shrinkage and fraud at the dispenser, which in some countries is a significant problem that can’t be resolved until all other fuel system losses/gains are identified and quantified.

Headed up by Dr Roshan Kumar of the EMS Research and Development team and LaTrobe University, the research is now in the Fuelsuite® productisation stage for release early 2019.

To learn more about this project’s functionality or Fuelsuite contact us via email on [email protected] or get in touch with us here.