Research On Recognition Technology Of Shearer Coal-rock Cutting Status Based On Multi-sensor Information Fusion

The shearer is one of core equipments in the fully mechanized mining face,accurate recognition of shearer coal-rock cutting status is the key to realize the intelligent fully mechanized mining face.Therefore,it is necessary to study the recognition technology of shearer coal-rock cutting status.This can improve the intelligent control level of the shearer.This thesis take the cutting sound signal and rocker arm vibration signal as the signal sources to identify the shearer coal-rock cutting status.To realize the accurate recognition of the shearer coal-rock cutting status,it is necessary to design feature extraction algorithm of signals,and study the multi-sensor information fusion recognition method which is based on the improved RBF neural network and D-S evidence.The main work and research results of the thesis are as follows:(1)The structure and working mechainism of the shearer are analyzed.Combined with the specific coal-rock cutting process,the cutting sound signal and the rocker arm vibration signal are selected as the basis to identify coal-rock cutting status.By using the cutting sound signal and two-axis vibration signal of the rocker arm,the structure of the shearer coal-rock cutting status recognition system based on multi-sensor information fusion technology is established,and designed the recognition flow.(2)The wavelet threshold denoising method for the cutting sound signal and the two-axis vibration signal of the rocker arm is studied.A variational mode decomposition method for signal based on improved gravity search algorithm is proposed.The high-dimensional eigenvectors of the decomposed signals are extracted based on envelope entropy and kurtosis.The principal component analysis is used to reduce the dimension of the high-dimensional eigenvectors,and then extracted the key feature information to represent the coal-rock cutting status.(3)The recognition model of sheaer coal-rock cutting status based on multi-sensor fusion is designed.To realize the recognition of coal-rock cutting status under single sensing signal,the pattern recognition method based on RBF neural network which is optimized by improved fruit fly optimization algorithm is proposed.The D-S evidence theory which is based on the evidence correlation coefficient,is used to make the decision-level fusion of recognition results obtained by three neural networks.The poor effect of the conflict evidence fusion is solved.In order to verify the research results of this thesis,the related experiments are carried out based on the shearer coal-rock cutting experimental platform,which is constructed by the research group in the Collaborative Innovation Center of Mine Intelligent Mining Equipment in Jiangsu Province.The experimental results obtained by the method proposed in the thesis show that: the key characteristics of the coal-rock cutting status in the cutting sound signal and the two-axis vibration signal of the rocker arm can be effectively extracted,the recognition of shearer coal-rock cutting status based on multi-sensor information fusion is realized,and the accuracy rate of recognition results is 96.5%,this verified the correctness and effectiveness of the proposed method.The research of this subject have great significance for improving the intelligent level of shearer,and promoting the development of "less people" or "unmanned" in fully mechanized mining face.