Digital twin to optimise the cost of electricity consumed by a water desalination plant under a time-based tariff

ABSTRACT Variable tariffs for the price of electricity are typically used to encourage consumption at times when there is less demand. Considering these tariffs, optimising the operation of the consuming equipment can result in significant economic benefits. In this work, a digital twin of a reverse osmosis desalination plant is designed and implemented to forecast the permeated water flowrate and electricity consumed, depending on the main operational parameter: the desalination pressure. The models relating these variables are fitted based on data measured from the process and are periodically updated. These models are utilised to optimise the plant's operation by relying on genetic algorithms based on the variable electricity tariff and using the pressure for each hour of the day to minimise the electricity cost. The proposed approach was tested by considering different permeated water production levels. Finally, an analysis was carried out by comparing the cost, during an entire month, when using a fixed operating pressure and when using the proposed digital twin-based optimisation. The resulting outcomes justify the convenience of optimising the desalination plant operation.