Iterative Learning Pressure Control in Water Distribution Networks

Proper management of pressure in water distribution networks is important, since leakages are more likely to occur if the pressure in the network is unnecessarily high with the consequence that the produced water is wasted. Furthermore, too low pressure will lead to reduced comfort of end-users and a potential increased risk of pollutants entering the network. In this paper, we investigate control actions for management of pressure in water distribution networks with multiple inlets. In particular, the problem of controlling the pressure at the network inlets to accommodate pressure requirements at designated network vertices, where the pressure is measured, is addressed. In this paper, we propose an Iterative Learning type control structure, as the behaviour of the consumers in water distribution networks is typically periodic in nature. Numerical and experimental results show that the proposed controller is able to fulfil the requirements.