| The problems of high power consumption,high steel consumption and low efficiency of grinding machine have puzzled many mineral processing enterprises and scientific researchers for a long time.With the rapid development of computer and artificial intelligence technology in recent years,the research on Optimization of grinding machine control driven by large data has highlighted a broad application prospect.This topic combines industrial big data,a section of abrasive class closed circuit operation process,and artificial intelligence.The offline predictive model of the power of the ball mill is established through industrial big data analysis,achieving the abrasive controller control optimization goals,and enhanced the system to respond to the nature of the ore.Ability.In order to combine industrial big data with a section of abrasive class closed circuit operation process,this project firstly analyzes the process and parameters of a period of abrasive class closed circuit operations,and proposes data selection methods that combine time scale and TLCC.This method combines the process through analysis of the particle size and time lag of the abrasive industry,and selects the most correlated variable sequence with the target parameter to provide high-quality data support for subsequent work.This topic also proposes a lightweight gradient improvement model based on flatting input.The model is flattened by high-dimensional data,and input is in the form of windows.Under the premise of ensuring the complete and low time complexity of the data structure,the long data prediction is transformed into a single data continuous prediction,which can effectively improve the predictive performance of the model.In order to reduce the adverse effects of dynamic and complex changes in the nature of the ore,this topic is proposed to stabilize the amount of ore by predicting and controlling the power of the ball mill.Related offline experiments prove that this ball mill control optimization strategy is effective. |
