Automatic Dosing System for Slime Water Sedimentation Process Based on Intelligent Optimization Algorithm

Coal slime water treatment is a non-linear, strong coupling, large lag process, and it is difficult to establish an accurate model. The model established by the existing research adopts a simple linear relationship or empirical formula, and the dosage of the drug is fixed or inaccurate, resulting in the inaccuracy of the drug. Improper use and waste will affect the processing rate of the slime water system. Therefore, according to the characteristics of the coal slime water treatment process, this paper studies the automatic dosing system in the coal slime water settlement process based on the intelligent optimization algorithm. In this paper, the Lssvm prediction model of the concentration process is established, and the multi-objective particle swarm algorithm (MOPSO) is used to optimize the dosage of the chemical agent. On the basis of ensuring the treatment effect and treatment rate of carbon-dyed water, the consumption of chemical agents is reduced. Experiments show that the LSTM model has the highest overall accuracy for flocculant prediction, and the accuracy rate P reaches 89.88%. Therefore, compared with BP neural network and RNN, LSTM is more suitable for the prediction of flocculation amount in the process of coal slime water settlement in this paper.