Development of an Intelligent Coagulant Dosing System for the Water Purification Process Based on an Artificial Neural Network

Abstract

The article reveals the essence and features of an artificial neural network, which is used to regulate water purification processes. Features, principles and main stages of water purification are determined. The stages of learning artificial neural networks are disclosed. The approach to the use of artificial neural networks during dosing of the mixture for water purification is substantiated. The process of dosing the mixture for water purification and the related indicators, which are important for the implementation of the water purification process, are analyzed. A number of factors that directly affect the coagulation process and, as a result, the structure of the neural network include turbidity and flow speed. It is shown that determining the dose of coagulant is necessary to minimize time, implement a continuous process, stabilize variations in the operator’s observations, and improve the quality of water treatment. A coagulant dose adjustment mode is proposed, as well as a water purification process control scheme based on the developed artificial neural network with unsupervised learning, which is used to optimize the coagulant dosage in the water purification process.