Prognostication of waste water treatment plant performance using efficient soft computing models: An environmental evaluation. (May 2019)

• Record Type: Journal Article
• Authors: Najafzadeh, Mohammad; Zeinolabedini, Maryam
• Is Part Of: Measurement. Volume 138(2019)
• Page Start: 690
• Page End: 701
• Publication Date: 2019-05
• Subjects: Artificial intelligent models -- Environmental assessment -- Flow rates -- Management -- Wastewater treatment plant; Weights and measures -- Periodicals; Measurement -- Periodicals; Measurement; Weights and measures; Periodicals; 530.8
• DOI: 10.1016/j.measurement.2019.02.014
• Languages: English
• ISSNs: 0263-2241

Abstract

Highlights: SVM, RBF-NN, FFBP-NN, and ANFIS were used to predict the flow rate for the WWTP. SVM and FFBP-NN had better performance than RBF-NN and ANFIS approaches. Soft computing techniques were applied to efficiently estimate the flow rate at the WWTP. Abstract: The chief purpose of designing the wastewater treatment plant (WWTP) is to provide a suitable system which is capable of eliminating the excessive impurities or pollutants found in the influent to the desired level. In this way, daily flow rates is one of the most crucial components to contribute in order to design wastewater treatment processes and plant units. So, in the present research work, various soft computing approaches including feed forward back propagation neural network (FFBP-NN), radial basis function neural network (RBF-NN), adaptive neuro-fuzzy inference system (ANFIS), and support vector machine (SVM) were employed to predict daily flow rates for the WWTP. To develop artificial intelligence models, flow rates datasets has been used over a five-year period. The performance of the proposed models were assessed for training and testing stages using statistical error indicators. Performance of techniques indicated that SVM (RMSE = 1435.4 and MAE = 1031.1) and FFBP-NN (RMSE = 1445.9 and MAE = 1036.7) techniques have provided more precise prediction of flow rates compared to the ANFIS (RMSE = 1515.6 and MAE = 1075.4) and RBF-NN (RMSE = 1501 and MAE = 1048.7). This study was proven that soft computing techniques, as robust tools, can be efficiently applied to design the flow rates in the WWTP with a persuasive degree of accuracy.