Cutting performance evaluation of a roadheader machine by PCA and RBF

AbstractAs coal mining technology has continuously evolved, gradually the industry has moved toward fully mechanized mining. A roadheader machine is important mechanical equipment for roadway drivage through mechanical crushing. Through analysis and research to discover the key parameters relating to the cutting performance of the roadheader machine, the performance of the roadheader machine must be optimized and costs reduced, as well as productivity increased. As one of the most important statistical methods, principal component analysis PCA could not only reduce many factors into fewer overall targets, it could also provide the comparative item weighting, improving the computational efficiency and error precision of the radial basis function neural network, eliminating the correlation of each input variable, and increasing the stability of the network model. The principal variables are determined and a cutting performance evaluation model developed that allow both performance prediction and cutting performance evaluation. From the analysis it is concluded that the primary indicators of the roadheader cutting performance were the unidirectional compressive strength, the cutting resistance fluctuation, the weaving speed of the cutting head, the cutting power fluctuation, and the traction resistance fluctuation. The model is consistent with the practical test results and contributes to discovery of future optimization procedures.