A Data-Driven Diagnosis Scheme Based on Deep Learning toward Fault Identification of the Hydraulic Piston Pump

The piston pump is the significant source of motive force in a hydraulic transmission system. Owing to changeable working conditions and complex structural characteristics, multiple friction pairs are prone wear failure. An accurate fault diagnosis method crucial guarantee for system reliability. Deep learning provides great insight into intelligent exploration machinery diagnosis. Hyperparameters very important construct an effective deep model with good performance. This research fully mines feature component from vibration signals, converts failure recognition classification issue via establishing model. Furthermore, Bayesian algorithm introduced hyperparameter optimization as it considers prior information. adaptive convolutional neural network established typical pattern axial pump. proposed can automatically complete represents higher accuracy by experimental verification. Typical failures intelligently diagnosed reduced subjectivity preprocessing knowledge. achieves identification more than 98% five