Research on the evaluation and prediction of coal mining machine status based on multivariate analysis

The purpose of this paper is to solve the problem of poor accuracy, low efficiency and weak anti-interference capability of condition assessment of coal mining machines when coal mines undergo intelligentization, to propose an intelligent health assessment method based on principal component analysis and whale optimization algorithm optimized backpropagation（BP） neural network model. According to the structure of coal mining machine’s key components and working principle as well as practical conditions of mining enterprises, 14 state monitoring parameters are chosen according to the principle of measurability, independence and comprehensiveness. They include traction section, cutting section, hydraulic system and other related parameters in coal mining machine state maintenance. Using 1 000 sets of state data of coal mining machines provided by a certain coal mine as experimental data, PCA is used for dimension reduction. The first four principal components can reach cumulative contribution rate of 98.15%, which serves as input features for the model. Then it uses WOA to optimize the standard BP and built PCA-WOA-BP model. Matlab simulation experiments show that the established model has average assessment accuracy of 96.01%, more than that of standard BP and WOA-BP models. Also needed fewer input nodes and spent less training time than training time of two model, provides references for coal mining machine condition assessment and maintenance. More improvements are needed to improve the availability of the model to a great extent under the large sample data conditions.