A plug-and-play data processing module for complex faults diagnosis

Intelligent fault diagnosis technology is a vital approach to detect and sustain contemporary equipment. Although it is widely utilized for equipment monitoring, the limited number of samples available in the industry and complex faults restrict its further development. This paper presents a plug-and-play data processing module to expand the number of samples and improve the internal relationship between them to improve the overall performance of the deep learning model. an Amplitude-phase composite data enhancement (APCUP) method for expanding training samples is designed in the module, which uses Discrete Fourier Transform to randomly fuse multiple fault types in proportion to generate new fault samples, achieving the goal of expanding the training dataset. Based on the interdependence of faults in real scenes, the module also designs an adjacent label smoothing (ALS) strategy to process the original label data, which smoothen the label over, prevents the excessive absolute classification and improves the robustness of the model. The efficacy and viability of the module have been demonstrated through comprehensive experimentation on the cutting-edge industrial robot arm platform and two public datasets. The incorporation of a data processing module has markedly enhanced the model's capacity to diagnose intricate faults. As the module processes the input, there is a significant increase in accuracy and a reduction in errors.