A Novel Fault Diagnosis Method Based on SWT and VGG-LSTM Model for Hydraulic Axial Piston Pump

Since the hydraulic axial piston pump is engine that drives transmission systems, it widely utilized in aerospace, marine equipment, civil engineering, and mechanical engineering. Operating safely dependably crucial, failure poses a major risk. Hydraulic malfunctions are characterized by internal concealment, challenging self-adaptive feature extraction, blatant timing of fault signals. By completely integrating time-frequency conversion capability synchrosqueezing wavelet transform (SWT), extraction VGG11, as well memory long short-term (LSTM) model, novel intelligent identification method proposed this paper. First, status data transformed into two dimensions terms time frequency using SWT. Second, depth features time–frequency map obtained dimensionality reduction carried out deep mining VGG11. Third, LSTM added to provide damage model for long-term capabilities. The Softmax layer evaluation various patterns health state. identify diagnose five typical states, including normal state, swash plate wear, sliding slipper loose slipper, center spring failure, based on externally observed vibration signals pump. results indicate average test accuracy state reaches 99.43%, standard deviation 0.0011, duration 2.675 s. integrated exhibits improved all-around performance when compared LSTM, LeNet-5, AlexNet, other models. validated be efficient accurate common defects pumps.