Research On Vibration Signal Of Shearer Rocker Arm And Its Cutting Pattern Recognition Method

Shearer is one of the key equipments to implement safety and high-activity yield of coal mine,the intelligence level of shearer is the key factor to implement “less humanizing” or “unmanned” of fully mechanized coal face and the cutting pattern accurate recognition is premise of shearer mining intelligently.Therefore,it's necessary to research on vibration signal of shearer rocker arm and cutting pattern recognition for implementing shearer cutting automatically and self-adaptive control.Therefore,it is necessary to research on vibration signal of shearer rocker arm and its cutting pattern recognition to establish foundation for automatic cutting and self-adaptive control.The environment of actual working conditions is extremely bad in coal mining,and shearer rocker arm is interfered by cutting coal wall,the machine body saltation and traction speed fluctuation,the vibration appear a nonlinear complex signal with noise.In this paper,the complex vibration signal of shearer rocker arm is taken as subjects,research on feature vector extraction method under different time scale,establish classification models of shearer cutting pattern and implement different cutting pattern recognition based on improved SVM.The main results could be expounded as follows:(1)By analyzing the basic structure and working mechanism of shearer,research on change mechanisms of shearer rocker arm vibration signal,prove the feasibility through decomposition different acceleration signal to recognize cutting pattern,and define the different cutting pattern based on roof,floor and coal characteristics.(2)Aiming at signal-noise ratio,false component and feature dimension of shearer rocker arm complex vibration signal,de-noising different frequency bands signal based on multi-threshold criteria wavelet packet.Eliminate the false component which occurs in empirical mode decomposition based on K.L,extract vibration signal feature vector based on multiscale fuzzy entropy and Laplasian score.(3)A shearer cutting pattern recognition method based on improved SVM is proposed to improve recognition accurate,research on an optimization algorithm through artificial fish swarm algorithm and particle swarm optimization to implement kernel parameter and penalty factor of SVM optimization.Design the system framework and flow of shearer cutting pattern recognition.(4)The capture system of shearer rocker arm vibration signal was built and the ground experiment was tested at Zhangjiakou Coal Mining Machinery Co.Ltd.The results showed that the recognition accurate of improved SVM based AFSA-PSO is higher than improved SVM based on AFSA and PSO,the proposed method is illustrated validity and effectiveness.