Variable operation of a renewable energy-driven reverse osmosis system using model predictive control and variable recovery: Towards large-scale implementation

Abstract

Powering Reverse Osmosis (RO) systems with Renewable Energy (RE) is essential for decarbonising water production. Integration of RE requires large-scale RO plants to operate efficiently using variable power. Nevertheless, variable operation (involving matching the RO load to available power without battery back-up) has only been implemented for small-scale systems. This paper presents a variable-speed operation technique suitable for large-scale RO systems using an optimised operational strategy and a Model Predictive Controller (MPC). The technique was validated using a laboratory test rig having comparable performance to large-scale systems. A dynamic plant model was used to design the operational strategy and control system. Several operational strategies were explored for varying the operating parameters according to power available from a RE source. An advanced control system based on MPC was designed and compared to a conventional Proportional-Integral-Differential controller. The results showed that operation at variable recovery with constant brine flowrate delivered the lowest specific energy consumption and widest operation range for a system with an isobaric pressure exchanger. The MPC controller improved the settling time for a 10% step-change in permeate flowrate by 47%. Moreover, it improved energy utilisation, giving a 2.35% increase in hourly permeate production for a defined power input time-series.