Prediction of Compressive Strength of Pellet Ore Based on Feature Selection and PSO-XGBoost
The chain grate rotary kiln process is a set of processes for producing pellet ore.Due to the complexity of the chain grate rotary kiln production process,the quality of the finished pellets is coupled with the production parameters and process parameters of each process,making it difficult to establish an accurate mechanism model.There is a significant lag in the detection of the quality of finished pellet ore in the actual production process,and it is not possible to provide timely feedback and adjust the process production parameters.In response to this issue,a prediction model for compressive strength of ball ore was established based on Spearman correlation coefficient and Pearson correlation coefficient for parameter feature processing,combined with XGBoost algorithm and particle swarm optimization algorithm.The model was trained using actual production data.The results show that the model has a high prediction performance,with a hit rate of up to 94％,which meets the actual production needs on site and can achieve fast and accurate prediction of the quality of ball ore,providing reference for on-site production and adjustment of process parameters.
NETL
NSTL
万方数据
