PGRS-FaultNet: A Fusion Model for Fault Detection and Prediction in Intelligent Manufacturing

Fault detection and prediction are the key tasks in the field of intelligent manufacturing. This paper proposes a fusion model called PGRS-FaultNet, based on PCA, GRU and SVM, for fault detection and prediction of high-dimensional time-series data. The model utilizes PCA for dimensionality reduction to decrease data redundancy; GRU is employed to capture temporal features, enhancing the model’s ability to handle complex time-series data; and SVM serves as the classifier to achieve accurate fault state recognition. We conducted experiments on the PGRS-FaultNet using the C-MAPSS dataset and the PHM Data Challenge dataset and the results indicate that the model outperforms existing mainstream models in key performance metrics such as accuracy, recall, F1-score and AUC. Particularly in AUC, PGRS-FaultNet achieved 0.95 and 0.93 on the two datasets, respectively, demonstrating exceptional fault detection capabilities. Ablation experiments further confirmed the effectiveness and rationality of the PCA, GRU and SVM modules. This study suggests that PGRS-FaultNet can provide an efficient solution for fault detection and prediction in intelligent manufacturing, offering potential application value for predictive maintenance of intelligent industrial equipment. Future research directions include data augmentation, real-time optimization and multimodal data fusion to further enhance the model’s robustness and applicability.