Research On Online Recognition Technology Of Coal-rock Interface Based On Near Infrared Spectroscopy

Coal-rock interface identification has always been a world-class problem,and researchers in various countries have used different methods to study it.Different molecular structures and compositions of coal and rock lead to different absorption and reflection characteristics in the near infrared spectrum region.According to this basic principle,the differences of coal and rock in terms of material essential composition are studied and distinguished.Based on the coal rock interface recognition need high real time,on the one hand can improve the ability of computer processing data,on the other hand,you can try to get a small amount of data but will consist of a sufficient amount of data in a timely manner in order to facilitate processing,at the same time,the coal rock interface needs associated with shearer position,make the interface data can provide a reference for coal winning machine cutting adjusted higher.According to the above conclusions,this paper provides a research idea of coalrock interface identification based on the qualitative and location of detection points: A single collimator probe is used to cooperate with the steering gear to collect the reflectance spectrum of the sampling points and the position information of the steering gear.The recognition algorithm is used to qualitatively identify the spectral data collected by the diameter probe,and to locate the detection points by analyzing the relative geometric relationship among the prototype,shearer,coal wall and the rotation Angle of the steering gear.The cyclic scanning method is used to obtain the information of a large number of scanning points,and the real-time coal-rock interface is obtained through data processing.The prototype has been tested in the State Key Laboratory of Heavy Coal Machinery,which lays a solid foundation for the further application of coalrock interface identification technology in underground coal mine and the subsequent control of sheaver height adjustment.The following studies were done(1)According to the application scene,the identification principle prototype is designed,which can be installed on the shearer body and can adjust the attitude according to the need to adapt to different working face geometric parameters,including spectral data acquisition system,Angle detection system of collimating mirror probe,real-time position acquisition system of shearer,computer program processing system,etc.(2)Spectral curves of different coal and rock samples at different scanning speeds were collected to analyze the spectral generation mechanism and the variation law of the scanning spectra at different speeds.By adopting smoothing algorithm,the recognition success rate of experimental samples could be effectively improved.Through the comparison of recognition rate and real-time performance,it is found that spectral Angle matching(SAM)is better than Pearson correlation coefficient as coal and rock recognition algorithm(3)Set up to simulate coal wall,conduct static coal rock interface recognition test,posture adjustment prototype set different prototype scanning speed,to obtain spectral curve under different prototype scanning speed,scanning and point location information,draw the space distribution of coal and rock,coal and rock by screening for differences,and six kinds of coal and rock interface prediction model is set up,With average relative error and the variance as evaluation parameters,by comparing the average relative error and the variance of each model,and contrast on the velocity section parameters,by comparing the selected the most suitable prediction model of coal and rock interface,found in the low speed scan-coal rock interface effect is best,choose the geometric distribution for coal and rock on a high speed scan mean more appropriate.(4)To carry out different walking speed of shearer and different prototype combination test between scanning speed,scanning speed and walking speed of shearer get prototype of regularity,found based on real-time computation of coal rock interface model is not suitable for dynamic collection,easy to be a single sample point,the mean to the geometric distribution of coal rock interface,According to the distribution of the experimental coal wall and the spatial distribution of the coal and rock samples at the scanning point,a more accurate coal and rock interface can be obtained by choosing the appropriate step distance to solve the coal and rock interface piecewide.Based on this,a solid foundation can be laid for the practical application in the mine.There are 71 figures,9 tables and 96 references in this paper.