Decentralized model reference adaptive control of water quality in water distribution networks

This paper presents a partially decentralized adaptive control approach for water distribution networks. The approach is based on approximating the input/output dynamic behavior of chlorine concentration in the distribution system as a time-varying, discrete-time linear model with uncertain or unknown coefficients. Each local adaptive control algorithm requires knowledge of the remaining control actions but feedback information from only one sensor output. The closed-loop stability properties of the control algorithm are investigated and a simulation example is presented to illustrate the approach.