A predictive maintenance method for shearer key parts based on qualitative and quantitative analysis of monitoring data

Abstract

Appropriate predictive maintenance is critical to shearers under poor working conditions to prevent major safety accidents and casualties. A predictive maintenance method for shearer key parts based on qualitative and quantitative analysis of monitoring data is proposed. Firstly, a high-fidelity model and hyper-realistic behavior simulations of the shearer key parts are implemented based on offline monitored data and online monitoring data, the obtained position, posture, trajectory and other information can be used for qualitative analysis. Subsequently, an autoencoder (AE) combined with a deep bidirectional gated recurrent unit (bi-GRU) prediction model is constructed to predict the RUL based on online monitoring data for quantitative analysis. The adjusted_R2 value obtained for the deep bi-GRU model was 0.9916, determined in a model-accuracy verification test. Finally, taking the rocker arm of shearer as an example, its working status is visualized using digital mirror from physical space to virtual space, and then its RUL is predicted based on the AE bi-GRU prediction model. The decision-making that guides predictive maintenance can be obtained based on the synthesis of qualitative and quantitative analysis. The application test results demonstrate that the proposed method can be effective for improving the intelligence of predictive process and the accuracy of predictive results.