Blade fouling fault detection based on shaft orbit generative adversarial network

Abstract

The precision and stability of anomaly detection methods are vital for the secure and efficient operation of machinery. In this paper, finite element model is firstly used to analyze the shaft orbit of cantilever rotor from the perspective of fault mechanism. Then shaft orbit generative adversarial network is proposed and applied to detect the blade fouling fault. Variational autoencoder is used as the foundational network architecture for extracting high-dimensional latent features from shaft orbit images. Concurrently, the seventh-order moment of shaft orbit images is extracted and embedded into a bypass within the generator, thereby enhancing the accuracy of fault detection. Two sets of real-world blade fouling fault data are collected and meticulously analyzed. The results demonstrate that the proposed method exhibits higher accuracy and more robust generalization capability in anomaly detection. Additionally, the utilization of gradient information for the localization and visual analysis of anomalies dynamically tracks the spatial evolution of the rotor shaft orbit throughout its entire lifecycle. The data generation capability and interpretability of the proposed model can effectively support the digital twin modeling and health management of rotating machinery.