Abstract:
Accurate identification of the wear and failure of shearer components can provide the necessary support for the prevention and early warning of shearer failures and related accidents, and shearer performance degradation evaluation is an effective way to identify the wear and failure of shearer. For the non-linearity of shearer performance degradation process, an artificial intelligence-based shearer performance degradation evaluation method is proposed. The working condition monitoring parameters and performance monitoring parameters of the shearer are achieved, and the working condition of the shearer is identified by applying the extreme learning machine method. The performance monitoring parameters are dimension reduced by using the principal component analysis method, and the benchmark Gaussian mixture model under each working condition is established. The relative entropy is used to measure the difference between the Gaussian mixture model and the benchmark Gaussian mixture model at a certain moment, so as to measure the performance degradation trend of each component of the shearer. It is proposed that the performance monitoring parameters can be obtained from geological conditions, environmental factors, vibration and load, shearer tilt, etc. The parameters can be obtained according to the availability of data and changes in practical applications. The principles of selecting shearer performance monitoring parameters are proposed, and the performance monitoring parameters can be chosen based on the classification of common electromechanical equipment monitoring parameters and the actual assembly condition of the shearer sensor. A case study is carried out by analyzing the performance of shearer cutting part, which is the part with the highest fault rate. The working conditions of the cutting part of the shearer are divided into four types: high-speed straight cutting, high-speed oblique cutting, low-speed straight cutting and low-speed oblique cutting. The traction speed is used as the working condition monitoring parameter, and the left cutting motor current is used as the performance monitoring parameter. The rationality of the parameters is verified by correlation analysis. The analysis results show that the performance degradation status of the shearer can be obtained by comparing Gaussian mixture model, and the performance degradation trend of the shearer cutting part at each monitoring point can be measured through the relative entropy.
