Autonomous water quality management in an electrochemical desalination process DOI Creative Commons
Zahid Ullah,

Nakyeong Yun,

Ruggero Rossi

et al.

Water Research, Journal Year: 2025, Volume and Issue: 280, P. 123521 - 123521

Published: March 20, 2025

This research explores advanced control strategies to enhance water quality in membrane capacitive deionization (MCDI) systems, employing a validated modified Donnan model. Three types of artificial neural network (ANN) controllers were developed and evaluated, namely, ANN-proportional-integral-derivative, ANN-Integral, Multiple Parallel ANN-Integral (MPAI) controllers. Among these, the MPAI controller demonstrated best performance was selected for further optimization. It then compared with an offline reinforcement learning using Conservative Q-Learning (CQL) algorithm. To optimize CQL controller, various reward functions tested, including quadratic penalty, exponential Gaussian function, function ultimately its effectiveness, achieving at approximately one. Both maintained effluent concentration 17 mM, despite variations inlet fouling dynamics, absolute errors under 0.4 mM. Notably, showed highest precision, error margin approaching nearly zero controller. study underscores potential AI-driven enhancing efficiency reliability MCDI contributing advancements treatment technologies.

Language: Английский

Multi-Modal Machine Learning Modeling of Electrosorption: Predicting Cr(Vi) Removal and Cr(Iii) Regeneration DOI

Yong-Uk Shin,

Sung Il Yu,

Hyokwan Bae

et al.

Published: Jan. 1, 2025

Language: Английский

Citations

0
Autonomous water quality management in an electrochemical desalination process DOI Creative Commons
Zahid Ullah,

Nakyeong Yun,

Ruggero Rossi

et al.

Water Research, Journal Year: 2025, Volume and Issue: 280, P. 123521 - 123521

Published: March 20, 2025

This research explores advanced control strategies to enhance water quality in membrane capacitive deionization (MCDI) systems, employing a validated modified Donnan model. Three types of artificial neural network (ANN) controllers were developed and evaluated, namely, ANN-proportional-integral-derivative, ANN-Integral, Multiple Parallel ANN-Integral (MPAI) controllers. Among these, the MPAI controller demonstrated best performance was selected for further optimization. It then compared with an offline reinforcement learning using Conservative Q-Learning (CQL) algorithm. To optimize CQL controller, various reward functions tested, including quadratic penalty, exponential Gaussian function, function ultimately its effectiveness, achieving at approximately one. Both maintained effluent concentration 17 mM, despite variations inlet fouling dynamics, absolute errors under 0.4 mM. Notably, showed highest precision, error margin approaching nearly zero controller. study underscores potential AI-driven enhancing efficiency reliability MCDI contributing advancements treatment technologies.

Language: Английский

Citations

0