A Review of Artificial Intelligence Integration in Electrocoagulation for Wastewater Treatment

ABSTRACT Electrocoagulation (EC) is a reliable technology for the treatment of industrial wastewaters with complex characteristics. It is, however, sensitive to many different factors including operating conditions, the dynamics of electrodes and the characteristics of the wastewater being treated. In addition, these factors are correlated; therefore, there is an increasing need for the development of AI‐based tools that can model non‐linear (NL) systems and provide a more accurate optimization process. In real life applications, sensitivity has been reported not only in the efficiency of EC treatment, but also in the energy required per unit of wastewater treated, as well as difficulties when scaling up EC systems from laboratory to industrial applications. This comprehensive study of the basic principles; key process parameters; and the key processes of reactors, electrodes; etc., and their limitations will provide an overview of how various types of artificial intelligence (AI) algorithms, including artificial neural networks (ANNs), adaptive neuro fuzzy inference systems (ANFISs), support vector machines (SVMs), and combinations thereof, can provide the necessary tools to predict, analyze and optimize the results of electrochemical processes. An assessment of how some of the main operational variables that can impact EC performance (e.g., pH, current density, electrode type and configuration, and length of treatment time) impacts the efficiency of pollutant removal, energy use, and electrode life, will also be included. These methods of analysis are able to accommodate complexity, multivariate nature and NL relationships of all the operational variables that have an impact on EC performance. Rather than a simple review of methods, this paper provides a holistic perspective that connects the mechanisms of electrochemical processes with current modelling trends being used to predict performance based on the use of data‐centric methods in conjunction with physicochemical theories.