Chapter 3 - Review of Soft Computing Techniques for Modeling, Design, and Prediction of Wastewater Removal Performance

Abstract

For preventing disruption of environmental balance, it is essential to maintain stable operating conditions and appropriate decision-making. Control, prediction, and optimization of wastewater treatment processes over a wide range of operating conditions are important goals for achieving the reduction of water pollution all across the globe. Of all other wastewater treatment processes, adsorption has received significant importance. Recent studies have reported that the application of models developed using soft computing techniques has enhanced the performance efficiency of this process. This study reviews core soft computing techniques like adaptive neuro-fuzzy inference systems, response surface methodology (RSM), artificial neural networks (ANNs), fuzzy logic, support vector machines, etc. for optimization and prediction of adsorption of emerging aqueous pollutants. Recent applications of these techniques for modeling adsorption of various aqueous pollutants have also been discussed herein. Review of existing literature reveals the bio-inspired population-based metaheuristic models to be the most efficient algorithms for simulation and optimization. Results indicate that the integrated RSM–ANN approach has been the most widely implemented method for simulation and optimization of adsorptive removal of aqueous pollutants. However, further research is required for developing simpler and user-friendly soft computing techniques capable of optimizing several process parameters in a cost-effective manner. Moreover, adsorptive removal of different types of aqueous pollutants like dyes, heavy metals, pesticides, and pharmaceuticals should be carried out in mixtures bearing multiple pollutants for better simulation of real-time scenarios.