Optimising the Operation of Renewable Energy-Driven Reverse Osmosis Desalination

  • Mohamed Tarek Mito

Student thesis: Doctoral ThesisDoctor of Philosophy

Abstract

The integration of Renewable Energy (RE) and Reverse Osmosis (RO) is essential for sustainable water production. However, it requires large-scale RO plants to accommodate fluctuating power inputs. Variable operation of RO plants by matching their load to available power, without battery back-up, has only been implemented for small-scale systems. This thesis presents a variable operation control procedure suitable for operating large-scale RO systems using RE. The procedure consists of two techniques, i.e., variable-speed operation and modular operation, for matching the RO load to varying degrees of RE fluctuation. The solutions presented were developed using a pilot RO plant that delivers similar performance to large-scale systems to allow implementation to such scale. Wind energy was used as a representation of an intermittent and fluctuating RE source. For variable-speed operation, multiple strategies were explored for varying the operating parameters according to available power. An advanced control system based on Model Predictive Control was designed and compared to a conventional Proportional-Integral-Differential controller. For modular operation, neural networks were developed to provide long- and short-term wind speed prediction for scheduling the RO units operation. 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. For a 10% step-change in permeate flowrate, the MPC controller improved the settling time by 47%. The long-term wind speed prediction was used to estimate the number of operational RO units for a day ahead for three random days, reaching a correlation of R2 0.78, 0.64, and 0.79 with the actual wind speed. This allowed scheduling the RO units to operate with a smooth operation profile that avoids unexpected shutdowns. By combining the optimised variable-speed and modular operations techniques, 90.9%, 91.5% and 91.4% of the available wind energy was utilised for Days 1, 2 and 3, which led to a high cumulative daily permeate production of 78 m3, 91.5 m3 and 123.4 m3, respectively. The solutions developed in this thesis showed that RO systems can be powered efficiently by RE using variable operation. This is fundamental for implementing this technology on a large-scale and decarbonising water production.
Date of AwardSept 2021
Original languageEnglish
Awarding Institution
  • Aston University
SupervisorXianghong Ma (Supervisor), Ahmed Rezk (Supervisor), Philip Davies (Supervisor) & Hanan Albuflasa (Supervisor)

Keywords

  • Desalination
  • reverse osmosis
  • renewable energy
  • variable operation
  • model predictive control
  • wind speed prediction

Cite this

'