A review on experimental design for pollutants removal in water treatment with the aid of artificial intelligence

Water pollution occurs mainly due to inorganic and organic pollutants, such as nutrients, heavy metals and persistent organic pollutants. For the modeling and optimization of pollutants removal, artificial intelligence (Al) has been used as a major tool in the experimental design that can generate the optimal operational variables, since Al has recently gained a tremendous advance. The present review describes the fundamentals, advantages and limitations of Al tools. Artificial neural networks (ANNs) are the Al tools frequently adopted to predict the pollutants removal processes because of their capabilities of self learning and self-adapting, while genetic algorithm (GA) and particle swarm optimization (PSO) are also useful Al methodologies in efficient search for the global optima. This article summarizes the modeling and optimization of pollutants removal processes in water treatment by using multilayer perception, fuzzy neural, radial basis function and self-organizing map networks. Furthermore, the results conclude that the hybrid models of ANNs with GA and PSO can be successfully applied in water treatment with satisfactory accuracies. Finally, the limitations of current Al tools and their new developments are also highlighted for prospective applications in the environmental protection. (C) 2018 Elsevier Ltd. All rights reserved.