Research on optimal arrangement strategy of top coal caving support sensors based on vibration characteristics of coal and gangue
Aiming at the research on intelligent identification of caving coal and gangue,in order to provide a complete and effective vibration signal acquisition scheme of coal and gangue,an optimal layout strategy of tail beam sensors of caving coal hydraulic supports based on vibration characteristics of coal and gangue was proposed.Firstly,the modal analysis of the tail beam model was carried out,extracting the vibration mode matrix,and the effective independent method was used to select the measuring points.Secondly,the vibration signals of coal falling and gangue falling at the corresponding primary measuring points from the tail beam test bench were obtained,and the feature extraction was carried out.Thirdly,the extracted features were visualized by t-distributed stochastic neighbor embedding(t-SNE)dimensionality reduction,and five features which were sensitive to the distinction between coal and gangue signals were selected as target features.Finally,the probability density functions of target features were estimated by the kernel density estimation method.The K-L(Kullback-Leibler)divergence was used to evaluate the approximation between combined signal of each measuring point and the complete signal and the difference between characteristics of coal and gangue.The evaluation indexes of coal and gangue vibration signals were constructed.Combined with Fisher information matrix criterion,a comprehensive evaluation index was formed to determine the optimal scheme of tail beam sensor arrangement.The results show that the sensor arrangement scheme determined by this method not only reduces the number of sensors on the basis of satisfying modal observability,but also makes the measured vibration signals have better coal gangue difference and information integrity.
Top coal cavingVibration signalOptimal sensor placementHydraulic support tail beamEffective independence methodK-L divergence
万方数据
维普
