You’ve reached your limit!

To continue enjoying Utility Week Innovate, brought to you in association with Utility Week Live or gain unlimited Utility Week site access choose the option that applies to you below:

Register to access Utility Week Innovate

  • Get the latest insight on frontline business challenges
  • Receive specialist sector newsletters to keep you informed
  • Access our Utility Week Innovate content for free
  • Join us in bringing collaborative innovation to life at Utility Week Live

Login Register

Affinity Water has said it will share a “pioneering” AI-driven system that predicts water consumption with the wider sector.

The company called the open data model a “major breakthrough” for the industry to better understand network-wide usage and manage impacts of fluctuating consumption.

The AI system can be used to monitor the impacts of weather events and unforeseen circumstances such as coronavirus on consumption in real time.

It measures water balance at district metered area (DMA) level and analysis the difference between behaviour in periods of hot weather and, for example, high demand in lockdown.

Andrew Morris, chief information officer at Affinity, said: “What is new is the idea to use fast logging to infer domestic consumption as well as adopting the powerful machine learning technique known as extreme gradient boosting that has recently been dominating applied machine learning and Kaggle competitions (the world’s largest data science community) for structured or tabular data.

The model was completed in four weeks, validated against existing data and then a per capita consumption (PCC) data dashboard was developed to access real-time data on daily usage. The system can model scenarios and compare usage in a normal year to exceptional events.

“In the future this will allow us to plan better for other abnormal events including weather patterns and to be able to monitor the impact much more closely of our work to reduce PCC through our other campaigns as well as it will give us more real-time insight,” Morris said.

The team has developed a customer level model to predict the water usage movement for each customer at any time. Affinity will release the model openly for all water companies to use freely to benefit the industry.

Morris said he hoped there will be more open data sharing in the future and sharing of proven data models.

PCC rates across Affinity’s region in 2019/20 were above average for England and Wales at 177 litres per person per day for non-metered customers and 136 litres per person daily for metered billpayers.