Time-variant solar-powered electrodialysis reversal desalination for affordable off-grid clean water supply

Thesis: S.M., Massachusetts Institute of Technology, Department of Mechanical Engineering, September, 2019Cataloged from the PDF version of thesis.Includes bibliographical references (pages 91-96).The goal of this research is to design affordable photovoltaic-powered electrodialysis reversal systems capable of desalinating brackish water in remote locations of developing countries, and thereby increase the availability of freshwater in water-stressed areas such as rural India. At the village scale, electrodialysis technology for brackish groundwater desalination has the potential to substantially reduce water wastage and energy consumption compared to on-grid reverse osmosis (RO) plants currently commercialized. Moreover, PV-powered systems can supply water in off-grid locations where on-grid systems cannot be installed, at no environmental cost. However, the levelized cost of water achieved by state-of-the-art PV-EDR systems is still prohibitively high for cost-constrained communities, due to the large battery capacity required to reshape the solar power profile and accommodate the inflexible power profile of the desalination load in constant operation. To reduce water cost, a novel, flexible operational strategy for PV-EDR systems is presented and experimentally validated on a full-scale pilot. By actively controlling voltage and flow rate with a dedicated hardware and software architecture, the ED power consumption is matched to the available solar power at any time. The experimental fraction of solar energy directly used reached 76%, which is 91% higher than in the constant operation case, where the PV-EDR system runs at constant voltage and flow rate. As a result, the experimental system dynamically adapted its desalination rate to the solar irradiance profile, producing freshwater in synchronization with the sun while reducing the need for batteries by 92% on average compared to the conventional operation. Because desalination efficiency decreases as a function of operating power, it is suggested that a small battery capacity would allow reshaping the direct solar power profile into a more suitable profile for water production. If optimally managed, a 3 kWh battery addition on the experimental setup is predicted to increase water production by 25%. A machine-learning-based algorithm was designed to predict the optimal battery management strategy online and is demonstrated in simulation to achieve over 99% of the ideal water production. Shifting from constant to flexible operation is expected to reduce the levelized cost of water by 22% compared to current state-of-the-art PV-EDR systems. This number was obtained by optimizing the flexible PV-EDR system design to minimize levelized cost of water (LCOW) while answering daily demand for one year for a case study village location in Chelluru, India. Most importantly, the optimal flexible PV-EDR system is shown to be cost-competitive with current on-grid community-scale RO desalination solutions in India. Cost projections for ED membrane and brine disposal show that in the future, PV-EDR could produce water at 60% of the cost of water produced with on-grid RO.by Anne-Claire Le Hénaff.S.M.S.M. Massachusetts Institute of Technology, Department of Mechanical Engineerin