Weighted Feature Fusion of Convolutional Neural Network and Graph Convolutional Network for Mechanical Fault Diagnosis under Time-varying Speeds

In real engineering environments, the time-varying speed condition is very common. However, most fault diagnosis methods take the constant speed into account, thus ignoring the fault signal changes under time-varying speeds. In this paper, we propose a weighted feature fusion framework based on convolutional neural network (CNN) and graph convolutional network (GCN) to achieve mechanical fault diagnosis. First, CNNs and GCNs are adopted to extract graph and long-range features. Then, a weighted fusion strategy is utilized to integrate the output of the two networks to obtain more diagnostic results. Finally, extensive experiments conducted on a time-varying speed dataset are used to validate the superiority and effectiveness of the proposed method compared to the other methods. Copyright (C) 2024 The Authors. This is an open access article under the CC BY-NC-ND license (https://creativecommons.org/licenses/by-nc-nd/4.0/)