Online modelling of water distribution system using Data Assimilation

Abstract

This paper applies Data Assimilation (DA) methods to a Water Distribution System Model to improve the realtime estimation of water demand, and hydraulic system states. A time series model is used to forecast water demands which are used to drive the hydraulic model to predict the future system state. Both water demands and water demand model parameters are corrected via DA methods to update the system state. The results indicate that DA methods improved offline hydraulic modelling predictions. Of the DA methods, the Ensemble Kalman Filter outperformed the Kalman Filter in term of updating demands and water demand model parameters. © 2013 The Authors.