LAUSR

Mechanical equipment in the process of operation appears a variety of faults due to different reasons. These faults affect the operation of machinery, causing economic losses, moreover, it may cause accidents, or even casualties. Predicting the nature, degree, development trend and position of mechanical faults is of great significance for making fault early warning, changing scheduling scheme and determining optimal maintenance time. This study aims to propose a generalized mechanical fault prediction method under the condition of short data validity. In the aspect of application, this paper hopes to combine the fault prediction with the shop dynamic scheduling, and constructs the mode of forecasting and optimizing the scheduling plan. The results show that the DGM (1,1) model based on amplitude compression is effective on the prediction of oscillation sequence. Markov chain modification could reduce the error greatly. The feasibility of fault prediction by using Grey-Markov chain has been proven by an illustrative example.