Volume 11 Issue 3 ( September 2024)

The purpose of this study is to develop a neural network model for optimising the coagulant dosing process. To achieve this goal, the study used methods of analysis, systematisation, modelling, and comparison. Industrial wastewater treatment methods designed to remove pollutants and improve water quality before being poured into reservoirs or water intakes were considered. Conventional methods include coagulation and flocculation, which involve adding a coagulant to agglomerate polluting particles and remove them from the water. The paper investigated the use of neural networks in the coagulant dosing system during wastewater treatment. In particular, a direct propagation network, also known as a multilayer perceptron, was considered. A neural network model has been created that allows determining the optimal dose of coagulant based on input parameters, such as water pH, turbidity, electrical conductivity. It is noted that the use of neural networks for coagulant dosing can improve the accuracy and efficiency of the wastewater treatment process, as well as reduce the cost of chemicals and ensure the stable operation of wastewater treatment plants.

Keywords: multi-layer perceptron; water purification automation; automated system; water parameters