Model predictive control for water management and energy security in arid/semiarid regions

This paper aims to develop a realistic operational optimal management of a water supply system in an arid/semiarid region under climate change conditions. The developed model considers the dynamic variation of water demand, rainfall, weather, and seasonal change in electricity price. It is mathematically developed as a multi-constraint non-linear programming model based on model predictive control principles. The model optimises the quantities of water supplied by each source every month and improves the energy efficiency in a water supply system with multiple types of sources. The effectiveness of the developed MPC model is verified by applying it to a case study and comparing the results with those obtained with an open loop model. Results showed that using the MPC model leads to a 4.16% increase in the water supply cost compared to the open loop model. However, when considering uncertainties in predicting water demands, aquifer recharges, rainfall, and evaporation rate, the MPC model was better than the open loop model. Indeed, the MPC model could meet the water demand at any period due to its predictability of variations, which was not the case with the open loop model. Moreover, a sensitivity analysis is conducted to verify the capacity of the developed model to deal with some phenomena due to climatic changes, such as in rainfall.