Research On Abnormal Prediction And Fuzzy Control Of Coal Storage Belt Conveyor Based On Big Data Analysis

With the introduction of industrial 4.0,China’s steel industry has taken the lead in the development of industrial intelligence.In order to ensure normal production of iron and steel plant,an on-line monitoring system for belt conveyor has been designed in a coal warehouse in Inner Mongolia,but the function of early warning and adaptive control is not enough to prevent the occurrence of faults.In view of this,this thesis optimizes LSTM neural network commonly used in data analysis,and uses the optimized algorithm to predict and warn the monitoring data of belt conveyor.A power adjustment system of belt conveyor based on fuzzy PID is designed to control the speed of the belt conveyor,which can ensure the stable operation of the belt conveyor.In this thesis,firstly,various data monitored by the belt machine system of a coal storage center in Inner Mongolia are merged,missing value filled and normalized to achieve data pre-processing.Next,the MPA-LSTM model and SMA-LSTM model are optimized using the marine predator algorithm and the sticky bacterium algorithm for the limitation of LSTM neural network in data analysis by the lack of local optimum and generalization ability.Then the LSTM model,MPA-LSTM model and SMA-LSTM model were used to analyze and compare the belt machine monitoring data,and the MPA-LSTM model with the best optimization effect was selected to predict the belt machine monitoring data and realize the early warning of belt machine working status.Finally,based on the results of early warning analysis and the relationship between motor power and speed,a fuzzy PID-based belt conveyor power regulation system is designed to control the motor speed by adjusting the belt conveyor power to achieve self-protection and recovery of the belt conveyor.