Photoelectric Identification Mechanism And Experimental Study Of Coal And Gangue Based On X-Ray Detection Technology

Coal and gangue identification is an important link in clean and efficient utilization of coal.Traditional coal and gangue identification methods,such as moving sieve jig and heavy medium,are criticized for environmental pollution and high energy consumption.Therefore,more efficient,intelligent and environmentally friendly coal and gangue identification technology is urgently needed.X-ray intelligent coal and gangue identification technology has the advantages of high precision,low energy consumption,easy operation and maintenance,green environmental protection,and has broad development prospects in coal and gangue identification.However,in the identification process,the identification accuracy of coal and gangue with small particle size is low,and the coal classification effect is poor.In order to solve the above problems,a method for identifying coal and gangue based on X-ray detection technology,that is,a combination of dual-energy X-ray transmission(DEXRT)and energy dispersive X-ray diffraction(EDXRD),was put forward.The pretreatment algorithm of coal and gangue identification data,the photoelectric identification mechanism and method of coal and gangue were studied,and the corresponding coal and gangue identification system was designed.The main research contents are as follows:(1)Collection and pretreatment of coal and gangue identification data.The particle size of coal and gangue can be divided into 5 ～ 13 mm,13 ～ 30 mm and 30 ～ 100 mm through coal screening test,which lays a foundation for subsequent collection of dualenergy X-ray transmission images and energy dispersion X-ray diffraction patterns of coal and gangue.In order to solve the problem of missing detection of small targets in dual-energy X-ray transmission images of coal and gangue,a target segmentation method based on maximum inter-class variance method(Otsu method)with the tree branch structure and block threshold is proposed.In order to eliminate the irrelevant information and noise contained in the energy dispersive X-ray diffraction patterns of coal and gangue,the denoising effect of moving average smoothing method and Savitzky-Golay convolution smoothing method on the pattern was studied,which laid a foundation for the extraction of coal and gangue identification features and the implementation of coal and gangue identification experiments.(2)Theoretical research on coal and gangue identification based on DEXRT and EDXRD.The basis for identifying coal and gangue based on the DEXRT is illustrated.Elemental analyzer and industrial analyzer are used to test the organic matter content and ash content of coal,and the mineral content of coal is calculated,and then the theoretical R value of coal is calculated.X-ray fluorescence scanner was used to measure the oxide content of gangue and calculate the theoretical R value of gangue.The theoretical calculation method is compared with the actual calculation method,and then the problem of identifying coal and gangue based on DEXRT principle is analyzed.The main mineral composition of coal and gangue is tested by Angular dispersive X-ray diffraction(ADXRD)instrument.Based on the analysis results of ADXRD,the feasibility and advantage of identifying coal and gangue based on the EDXRD principle is discussed by combining simulation and experiment.Thus,it lays a theoretical foundation for DEXRT and EDXRD to identify coal and gangue.(3)Research on the identification of coal and gangue with different particle sizes based on DEXRT and EDXRD.Using Support Vector Machine(SVM)as the basic model for coal and gangue identification,the DEXRT and EDXRD are used to study the identification effect of coal and gangue with particle size of 5 ～ 13 mm,13 ～ 30 mm and30 ～ 100 mm.Meanwhile,in order to optimize the performance of SVM coal-gangue recognition model,the effects of Particle Swarm Optimization(PSO),Genetic Algorithm(GA),Grid Search(GS)on the coal and gangue recognition model were studied,the best coal and gangue recognition model is obtained,and the coal and gangue recognition method based on DEXRT and EDXRD principle is verified.It is found that GS-SVM is used as the coal and gangue recognition model.When GS-SVM is used to identify coal and gangue with the particle sizes of 5 ～ 100 mm,the comprehensive recognition accuracy of coal and gangue is 97%,which is better than that of DEXRT principle alone.(4)Research on identification of coal types and gangue based on EDXRD.Taking the coal classification and efficient utilization as the goal,taking the coal mining situation of 1613 A working face of Zhangji Mine and 360802 working face of Xinji No.1 Mine in Huainan as the engineering background,taking 1/3 coking coal,gas coal and gangue as the research object,the difference of EDXRD patterns of 1/3 coking coal,gas coal and gangue is analyzed,which lays a theoretical foundation for the identification of coal types and gangue by EDXRD.Meanwhile,in order to eliminate redundant identification features,Principal Component Analysis(PCA),Partial Least Square(PLS)and Nonnegative Matrix Factorization(NMF)based on stoichiometry were used to reduce the dimensionality of EDXRD patterns,and GS-SVM coal and gangue recognition model was used to evaluate the three dimensionality reduction methods.The study found that the PLS dimension reduction effect is the best,at this time,the identification accuracy of1/3 coking coal,gas coal,gangue is 98.3 %.(5)Design of coal and gangue identification system based on X-ray detection technology and experimental verification.The framework of coal and gangue identification system based on X-ray detection technology is deployed.Aiming at the problems of raw coal particle size classification,adhesion,regular arrangement and coal and gangue separation,a corresponding mechanical system is developed,which mainly includes particle size screening mechanism,material conveying mechanism,material arrangement mechanism and coal coal and gangue separation mechanism,laying a foundation for subsequent identification data collection and separation of coal and gangue.A feature acquisition system based on dual-energy X-ray transmission and energy dispersive X-ray diffraction is designed to collect the identification data of coal and gangue.In order to control and monitor the working status of each module in real time,an integrated industrial computer is built and the corresponding visual interface is designed.Finally,the effectiveness and feasibility of the coal and gangue identification method in this paper was verified.Figure [101] Table [21] Reference [162]...