Research On The Coal-rock Property Recognition Method During Drilling Process Of The Drilling Robot For Rockburst Prevention

Rockburst is one of the most serious dynamic disasters in deep coal mining in China.Drilling for pressure relief can effectively prevent rockburst.The drilling robot for rockburst prevention is the core equipment to realize the unmanned drilling for pressure relief.The recognition of coal-rock property during drilling process is the prerequisite for efficient drilling for rockburst prevention.In order to address the problem of complex and changing drilling conditions,noisy sensing data and low recognition rate of individual sensing information during drilling process,this thesis integrates the main drilling unit’s drilling sensing information such as vibration,propulsion pressure and torque to realize the recognition of coal-rock property during drilling process of the drilling robot for rockburst prevention.The signal denoising algorithm based on VMD-WT algorithm and the feature extraction algorithm based on SVD are designed.The SVM is optimized by ALO algorithm,which improves the core parameters of coal-rock property recognition model with single sensing information.The multi-sensor decision-level fusion based on improved D-S evidence theory is proposed to improve the recognition rate of coal-rock property during drilling process of the drilling robot for rockburst prevention.And the experimental verification was conducted.The main work and research results of the thesis are as follows:(1)The basic structure and the work characteristics in drilling process of the drilling robot for rockburst prevention are analyzed.Combined with the basic principle of drilling into the coal-rock,the main drilling unit’s sensing information such as vibration,pressure and torque are selected as the basis for the recognition of coal-rock property.The framework of the coal-rock property recognition system during drilling process of the drilling robot for rockburst prevention based on multi-sensor information fusion is established,and then the coal-rock property recognition process is designed.(2)The signal denoising algorithm based on variational modal decomposition and wavelet threshold is studied.The high-dimensional eigenvectors of the decomposed signals are extracted based on the kurtosis and the rms of the frequency.The singular value decomposition is used to reduce the dimension of the high-dimensional eigenvectors,so as to establish the key characteristic set which represents the coal-rock property.(3)The multi-sensor information fusion model of coal-rock property during drilling process is established.The SVM is optimized by ALO algorithm and the optimal parameters of the penalty factor and kernel function in the coal-rock property recognition model are given to improve the coal-rock property rate under single sensing information.And the multi-sensor information fusion recognition method based on the improved D-S evidence is proposed for decision-level fusion of recognition results of single sensing information,which improves the problem of low recognition rate of single sensing information.(4)The experimental platform for intelligent drilling of the drilling robot for rockburst prevention was built and experiments on the recognition of coal-rock property during drilling process were carried out.The experimental results show that: the sensing noise can be effectively eliminated by the pre-processing of sensing information based on VMD-WT algorithm.The key characteristic set can be established by feature extraction model based on SVD.The coal-rock property during drilling process can be effectively recognized by the coal-rock property recognition model based on the multisensor decision-level fusion.The accuracy rate of recognition results can reach 94.38%,which verifies the feasibility of the proposed method.