Optimal sizing of stand-alone wind-powered seawater reverse osmosis plants without use of massive energy storage

A method, which involves genetic algorithms, is presented for the optimal sizing of a system comprising a medium-scale modular seawater reverse osmosis desalination plant powered exclusively by off-grid wind energy. The system uses a water storage reservoir that allows coverage of a particular hourly freshwater demand. The use of massive energy storage devices is discarded, although flywheels are used as a dynamic regulation subsystem as well as an uninterrupted power device to supply energy to the control subsystem. The method considers the interannual variation of wind energy, for which it uses machine learning techniques, and introduces randomness in the daily freshwater demand profile. The control strategy is based on ensuring that the energy consumption of the desalination modules remains in synchrony with wind generation throughout the system's useful life, either operating under constant pressure and flow conditions or varying these parameters within an acceptable range. The proposed method is applied to a case study, aiming to cover a freshwater demand of 1825 x 103 m3/year, which is equivalent to the water production of a desalination plant with a 5000 m3/day capacity. As the proposed method evaluates the influence of diverse economic and technical parameters, it constitutes a useful tool in the design and implementation of such systems. The results obtained with the optimal system of the case study are compared with those obtained on the basis of a configuration that uses backup batteries to ensure continuous operation. It is shown that the variable operating strategy provides the optimal economic system.