Study On Jigging Intelligent Control Based On Data-driven

Jigging coal separation is one of the main washing processes used in coal preparation plants in China.However,it is difficult to establish an accurate mathematical model due to the numerous influencing factors and complex mechanism in the separation process.The process has the characteristics of large time variation and strong nonlinear,which seriously affects the improvement of production efficiency and product quality control in the separation process.In this thesis,by collecting the historical production data of jigging separation process in coal preparation plant and combining with the analysis of advanced control theory and algorithm,a data-driven intelligent control model was established,and an intelligent control system of jigging separation process with cleaned coal ash as the control target was constructed,which realized the control of cleaned coal product quality.The reliability of data source is the root of intelligent control.This thesis firstly studies the detection and acquisition of jigging parameters.Among them,through the analysis of jigging process flow and BATAC jigging mechanism,the key variables affecting the control of jigging process,such as feeding property,feeding amount,buoy weight,cleaned coal ash and air valve parameters,were determined,and the detection of key variables was studied.Based on the existing communication equipment of coal preparation plant,the data acquisition and storage methods were studied.According to the characteristics of various data changes in the sorting process,various data acquisition strategies are formulated,and the real-time database structure and data table storage form are designed.According to the historical data of the production process,the layered process control of jig machine was studied.Firstly,the influencing factors of bed stratification state were analyzed,the bed looseness was determined to represent the bed stratification effect,the online detection method of the looseness was studied,and then the fuzzy control scheme of the looseness was proposed based on the expert experience and the field practice.A fuzzy controller was designed which took the looseness deviation and the change rate of the deviation as the input and the adjustment amount in the intake and exhaust period as the output.The looseness fuzzy control model is established.Field application results show that the established control model can control the gangue section looseness within 0.20-0.40,and the middle coal section looseness within 0.35-0.50,which provides a guarantee for controlling the quality of cleaned coal products.When the ash content of washed raw coal changes abruptly,the key to ensure the quality stability of cleaned coal products is to control the bed in a good loose state and adjust the buoy weight in the middle coal section quickly and accurately after the fuzzy control model of loose degree is established.Through further analysis of the historical production data,the structure of the buoy weight prediction model in the middle coal section is determined.And through the data processing methods such as outlier detection,missing value interpolation and data noise reduction,abnormal data processing and time lag processing are carried out on the collected historical data,providing data preparation for the establishment of the model.On the basis of the above,the sample data is trained by using LS-SVM algorithm,and the optimal super parameters of the model are determined by using GS-CV algorithm,and then the LSSVM prediction model of buoy weight in middle coal section based on GS-CV optimization is obtained.Compared with the LS-SVM model and ELM model based on GS optimization,the results show that the LS-SVM model based on GS-CV optimization has the best forecasting ability.The MAE of the model is 57.05 and RMSE of the model is 75.19,both of which are smaller than the minimum value of manually adjusted weight,indicating that the model has good forecasting ability.It can satisfy the prediction and adjustment of coal buoy weight in the production process.Based on the looseness control model and the weight prediction model of the buoy in the middle coal section,the jigging intelligent control system frame with cleaned coal ash content as the control target is built.In order to further improve the control accuracy of the system,the scheme of adjusting the weight of the buoy in the middle coal section is optimized.A fuzzy compensator is designed,which takes the real-time cleaned coal ash deviation and the change rate of the deviation as the input and the compensation amount of the buoy weight in the middle coal section as the output.Through the real-time feedback of cleaned coal ash,the compensation of the predicted results of the weight model is realized,so as to control the quality of cleaned coal products.Then,King View and AB 1756 PLC are used to design the visual monitoring interface of the production process and the logic control program of the lower machine,and MATLAB and PLC data communication method is used to build a data-driven jigging process intelligent control system with cleaned coal ash as the control target.It provides guarantee for realizing visualization and intelligent control of production process.