Institute for Water and Wastewater Technology, Durban University of Technology, Durban, PO Box, 1334, South Africa.
Department of Civil Engineering, University of Malaya, Kuala Lumpur, Malaysia.
J Environ Manage. 2021 Sep 1;293:112862. doi: 10.1016/j.jenvman.2021.112862. Epub 2021 May 25.
To ensure the safe discharge of treated wastewater to the environment, continuous efforts are vital to enhance the modelling accuracy of wastewater treatment plants (WWTPs) through utilizing state-of-art techniques and algorithms. The integration of metaheuristic modern optimization algorithms that are natlurally inspired with the Fussy Inference Systems (FIS) to improve the modelling performance is a promising and mathematically suitable approach. This study integrates four population-based algorithms, namely: Particle swarm optimization (PSO), Genetic algorithm (GA), Hybrid GA-PSO, and Mutating invasive weed optimization (M-IWO) with FIS system. A full-scale WWTP in South Africa (SA) was selected to assess the validity of the proposed algorithms, where six wastewater effluent parameters were modeled, i.e., Alkalinity (ALK), Sulphate (SLP), Phosphate (PHS), Total Kjeldahl Nitrogen (TKN), Total Suspended Solids (TSS), and Chemical Oxygen Demand (COD). The results from this study showed that the hybrid PSO-GA algorithm outperforms the PSO and GA algorithms when used individually, in modelling all wastewater effluent parameters. PSO performed better for SLP and TKN compared to GA, while the M-IWO algorithm failed to provide an acceptable modelling convergence for all the studied parameters. However, three out of four algorithms applied in this study proven beneficial to be optimized in enhancing the modelling accuracy of wastewater quality parameters.
为确保处理后的废水安全排放到环境中，必须不断努力，通过利用最先进的技术和算法来提高废水处理厂（WWTP）的建模准确性。将受自然启发的元启发式现代优化算法与模糊推理系统（FIS）集成，以提高建模性能是一种很有前途且数学上合适的方法。本研究将四种基于种群的算法（粒子群优化算法（PSO）、遗传算法（GA）、GA-PSO 混合算法和Mutating invasive weed optimization 算法（M-IWO）与 FIS 系统集成。选择南非（SA）的一个全规模 WWTP 来评估所提出算法的有效性，其中对六种废水出水参数进行建模，即碱度（ALK）、硫酸盐（SLP）、磷酸盐（PHS）、总凯氏氮（TKN）、总悬浮固体（TSS）和化学需氧量（COD）。研究结果表明，与单独使用 PSO 和 GA 算法相比，PSO-GA 混合算法在建模所有废水出水参数时表现更好。PSO 在 SLP 和 TKN 方面的性能优于 GA，而 M-IWO 算法未能为所有研究参数提供可接受的建模收敛性。然而，本研究应用的四种算法中的三种都有助于优化，以提高废水水质参数的建模准确性。
