Agent Based Modelling of city-wide Water Demand Seminar

Short overview of the seminar

Current methods to simulate household water demand, whether statistical, machine learning or agent-based modelling, are unable to capture the micro-components of water use at the granularity of appliance use and human behaviour at high temporal solution and scale these to the district or city scale.

In this research, a custom developed Agent-Based Model together with 2200 smart meter, high resolution water demand were used to derive a large per household parameter dataset. Combining it with household-level census data, city-scale models were generated and verified with demand measurements in Newcastle upon Tyne.

It has been shown that such a model can provide useful simulations of water demand at an individual appliance and person level with a high temporal resolution for city-scale simulations. The improved understanding of water demand enables more granular planning for both future housing estate construction work; understanding and predicting changing behaviours and water use; and the impacts of climate change on water availability and the benefits of adaptation actions. However, it is clear that the quality and completeness of data is limited and a step change in the accuracy and resolution of water meter data is essential to reduce uncertainties in these kind of models sufficiently to produce a complete Digital Twin at city-scale.

Speaker Information

Sven Berendsen

 

Link to join in