Predictive models for price forecasting in the electrical power industry

Our partner

  • Texas Utilities Ltd.

Project industry

  • Analytics and Business Intelligence

Project overview

We’re investigating the use of data mining and intelligent techniques to predict the energy consumption profiles of industry customers for an electricity distribution company in Melbourne, Australia.

This will also allow the company to customise their pricing offers to individual customers and optimise their electric power costs.

The project’s objectives include predicting the quarterly, seasonal and yearly energy consumption of electricity customers and integrating this predicted consumption with pricing offers. Future work will involve applying this method to commercial as well as seasonal customers.

Outcome

  • prediction error rate of less than 2 percent achieved.
  • improved regression techniques
  • a bespoke system used for forecasting as well as for pricing.