Research on Fault Diagnosis Method for Ship Construction Equipment Based on Hidden Markov Model

In response to the low efficiency of welding and cutting machines in shipbuilding, which limits production capacity and efficiency improvement, this paper studies a fault diagnosis method for shipbuilding equipment based on hidden Markov models (HMM). Firstly, 11 types of fault features were extracted based on one-dimensional convolutional wavelet neural network method (1D-CWNN), and the fault features can be accurately extracted, providing accurate input for fault diagnosis algorithms. Then, a fault diagnosis model for welding and cutting machines based on HMM was established, and the model was well trained and tested through a series of algorithms, such as Baum-Welch algorithm and Viterbi algorithm. The results indicate that the proposed method can classify equipment faults accurately with an accuracy rate of 96.8%, providing a good technical means for further improving the efficiency of welding and cutting equipment.