Fault diagnosis method, device and equipment for industrial system and storage medium

Abstract

The invention discloses a fault diagnosis method, device and equipment for an industrial system and a storage medium. The method comprises the following steps: obtaining historical original sequencesoutput by a plurality of sensors preset in an industrial system and corresponding industrial system fault types; splicing the data of different sensors in the same time slice in sequence, and normalizing the reconstructed sequence obtained by splicing to obtain a preprocessing sequence corresponding to the time slice as a training sample; training a deep learning model by using all the training samples to obtain an industrial system fault diagnosis model; and obtaining a diagnosis sample from the original sequence obtained in real time according to the same obtaining method as the training sample, and carrying out online diagnosis on the industrial system by using the industrial system fault diagnosis model and the diagnosis sample. The method has no requirements for the sampling rates ofdifferent sensors, can reserve most original data of the sensors, and improves the accuracy of fault diagnosis.

Description

technical field [0001] The invention relates to the field of fault diagnosis, in particular to a fault diagnosis method, device, equipment and storage medium of an industrial system. Background technique [0002] As the cost and complexity of industrial systems increase, fault diagnosis has received extensive attention. Accurate fault diagnosis can significantly reduce safety hazards, reduce performance degradation, and improve production efficiency. The vigorous development of intelligent manufacturing provides new opportunities for data-driven fault diagnosis methods, which use historical data to establish fault diagnosis models and make decisions based on online data collected by sensors. [0003] Data-driven fault diagnosis methods generally include four steps: data collection, feature extraction, model training, and model-based fault diagnosis. In the process of data acquisition, multiple sensor data signals, such as vibration, current, pressure, speed and temperature,...

AI Technical Summary

Problems solved by technology

[0006] The above methods for processing multi-sampling rate data are all machine learning methods, and most of them are only applicable to dual-sampling rate systems, and it is difficult to apply them to multi-sampling rate systems

Moreover, there is no research using deep learning methods to solve this problem.

Images