Research And Application Of Alumina Dissolution Prediction Algorithm

In the process of alumina production,high pressure stripping is one of the key processes,and its effect will directly affect the quality and yield of alumina.There is a certain lag in obtaining the dissolution effect in actual production.If the dissolution rate can be predicted in advance,the ratio of raw pulp raw material and the dissolution temperature and time can be adjusted in time according to the predicted results,so as to improve the dissolution effect.In this paper,the improved grey Wolf algorithm is firstly improved,and then the improved grey Wolf algorithm is used to optimize XGBoost,support vector regression(SVR)and extreme learning machine(ELM)algorithms respectively.Then the optimized model is used to predict the dissolution rate respectively.The research and development contents are as follows:1.Preprocessing and visualization analysis of the production data collected from alumina enterprises during the stripping process were carried out.2.XGBoost algorithm was used in combination with alumina process to analyze the characteristics of alumina dissolution process parameters and determine the process parameters related to the prediction of dissolution rate.3.The gray Wolf algorithm is improved by changing the progressive long update mode,which improves the convergence speed and execution efficiency of the algorithm.4.On this basis,XGBoost,support vector regression and extreme learning machine prediction models based on the improved gray Wolf algorithm are built respectively.5.The production history data of alumina stripping process were used to train the above three prediction algorithm models,and the prediction effect was evaluated and analyzed.Finally,the prediction results were visualized.6.Finally,the alumina dissolution rate prediction system for alumina enterprises is developed.