Control, design, and field validation of photovoltaic
electrodialysis desalination systems for decentralized
applications
Name
Bessette-jbessett-phd-meche-thesis-2026.pdf
Size
78.43 MB
Format
Adobe PDF
Checksum (MD5)
1c19f6f905a4367956aba90f3d0d4e08
Author(s)
Bessette, Jonathan
Advisor(s)
Winter V, Amos G.
Date Issued
February 2026
Publisher
Massachusetts Institute of Technology
Abstract
This thesis develops new control, power-electronics, and system-design theory to advance low-cost, renewable, brackish-water desalination using electrodialysis (ED). Decentralized desalination is increasingly critical as groundwater salinity rises due to climate change, aridification, and coastal intrusion, disproportionately affecting low- and middle-income communities. In these settings, water and energy infrastructure are often unreliable or absent, making photovoltaic (PV) electrodialysis a promising approach. However, current renewable ED systems remain costly and complex due to their reliance on energy storage, limited understanding of architectural tradeoffs, and a lack of practical system implementation knowledge at scale. This thesis addresses these challenges through four contributions spanning control, power electronics, thermodynamic and technoeconomic modeling, and an experimental campaign at commercial scale. The first contribution introduces and experimentally validates a simple, robust direct-drive control scheme: flow-commanded current control (FCCC) for variable-powered electrodialysis systems. By combining a cascade feedback controller with a subsystem-level power allocation strategy, this method maintains desalination at the instantaneous maximum allowable rate under highly intermittent solar conditions. This control scheme was implemented on 6- month long, community-scale field pilot in New Mexico and demonstrated 94% extracted solar energy utilization and a 99% reduction in battery capacity for a system of equivalent productivity while maintaining stable operation across varying weather and feedwater. This work demonstrated high-productivity, minimal-storage PV-ED desalination in real-world conditions. The second contribution presents a new power-electronics architecture: an integrated multiconverter maximum power point tracker (MPPT) that simultaneously performs maximum power extraction and optimal ED subsystem regulation. Unlike conventional off-grid systems requiring a centralized MPPT converter and a battery buffer, this approach coordinates downstream converters through a weighted supervisory MPPT signal. The system dynamically allocates PV power between the ED stack and pump to maximize water production while tracking the solar maximum power point. Laboratory demonstrations with brackish ED loads achieved power tracking within 14.3 ± 2.3% of the true MPP and subsystem regulation within 10.3 ± 0.3% of optimal setpoints. The architecture eliminates an entire converter and removes the need for energy storage, reducing cost and improving overall system efficiency. The third contribution establishes a unified thermodynamic and technoeconomic framework to evaluate multi-pass electrodialysis architectures—feed & bleed, batch, and a new architecture introduced here, “yo-yo.” Using thermodynamic modeling, the framework quantifies minimum separation energy and achievable productivity across salinity regimes. Batch and yo-yo configurations are shown to outperform feed & bleed systems in most scenarios due to reduced remixing losses, with the yo-yo architecture achieving the lowest theoretical energy consumption at high removals. Technoeconomic modeling across industrial polishing, brackish groundwater, and seawater scenarios reveals that batch operation provides robust, low-cost performance across all cases; feed & bleed is advantageous only in near-polishing regimes; and yo-yo is preferred at extreme salinities despite greater control complexity. This contribution establishes a thermodynamic and technoeconomic modeling framework that reveals the energetic penalties, productivity limits, and cost tradeoffs for multi-pass ED, creating a basis for selecting optimal and low-cost architectures across diverse operating conditions. The fourth contribution presents the development of a pilot-scale, multi-stack ED desalination platform and conducted comprehensive experimental evaluations of multi-pass architectures under varied feed salinities, recoveries, and hydraulic conditions, characterizing their energy, productivity, and cost performance. In our initial (1-2 stack) batch testing campaign, we find an all-inclusive, Pareto-optimal system SEC bandwidth of 2 to >14 kWh/m³ with a time-averaged production rate of roughly 1 to 8.5 m³/hr, respectively, in operation on 6000 and 12000 µS/cm NaCl solution. We find the pareto optimal (minimum SEC, maximum productivity) points are typically at the upper range of our applicable constant current operation; constant voltage operation does not ever appear in these experiments, on the pareto frontier—indicating constant current as desirable for batch operation at scale. Together, these contributions form design improvements for low-cost, high-productivity photovoltaic electrodialysis. This work advances fundamental understanding of ED architecture tradeoffs, introduces new control and power-electronics theory for renewable-powered operation, and demonstrates practical field viability at scale. Collectively, this work provides improvements towards resilient, decentralized desalination systems in hopes of expanding water access in resource-constrained and infrastructure-limited regions around the world.
MIT Department
Massachusetts Institute of Technology. Department of Mechanical Engineering
Terms of Use
Attribution-NonCommercial-NoDerivatives 4.0 International (CC BY-NC-ND 4.0)
Copyright retained by author(s)
Persistent DSpace Link
