Shearer Cutting Pattern Recognition Based on Multi-scale Fuzzy Entropy and Support Vector Machine

Abstract

Aiming at the problem of low intelligent level of shearer, a shearer cutting pattern recognition method is proposed based on the combination of multi-scale fuzzy entropy, Laplace score and support vector machine. By extracting the multi-scale fuzzy entropy of the vibration signal under different cutting modes, the feature vector representing the cutting pattern is mastered. At the same time, the Laplace score is used to select the feature vectors with possessing rich cutting pattern information. The selected features are produced as the learning samples of support vector machine. The experimental system of shearer cutting coal-rock is built, and the vibration signals of rocker arm under different cutting patterns are extracted. The experimental analysis is carried out and the results indicate that the cutting pattern recognition method proposed in this paper has high recognition accuracy, and the correct rate can reach to 98.86%. The research results provide technical support for the intelligent and rapid development of fully mechanized mining face.