An Adaptive Multi-model Driven Device Fault Diagnosis Method Based on Edge-Cloud Collaboration

Abstract

The invention discloses an adaptive multi-model driving device fault diagnosis method based on edge-cloud collaboration. First of all, the present invention proposes a multi-model branch selection method based on the time tolerance factor, setting multiple diagnostic model branches, and selecting the branch model with the highest accuracy within the range of the time tolerance factor; further adopting a diagnosis model division method based on edge-cloud collaboration, The fault diagnosis model based on deep learning is divided between edge and cloud at the granularity of layers; finally, a cross-working condition diagnosis method based on edge-cloud collaboration is proposed, and the general working condition model is trained on the cloud and sent to the edge. The terminal diagnoses the data of its personalization status. The present invention aims at the problem that the cloud computing fault diagnosis mode is not real-time enough, and the computing resources and storage capacity of the edge equipment are limited so that the deep learning fault diagnosis model cannot be directly deployed, and the traditional cloud computing diagnosis mode is improved by means of edge-cloud collaboration. The time delay is effectively reduced and cross-working condition diagnosis is realized.

Description

technical field [0001] The invention relates to the technical field of industrial equipment fault diagnosis and edge-cloud collaboration, in particular to an adaptive multi-model driven device fault diagnosis method based on edge-cloud collaboration. Background technique [0002] With the development of artificial intelligence, Internet of Things and industrial Internet technologies, industrial manufacturing is moving towards digitalization and intelligence, and mechanical equipment is also developing in an increasingly complex and integrated direction. Once the precision and key mechanical parts with complex structures fail, it will seriously affect the normal operation of mechanical equipment and cause heavy losses. Therefore, the prediction, diagnosis and prevention of large-scale mechanical equipment failures are important issues in the development of industrial manufacturing. [0003] Traditional fault diagnosis methods mainly include artificial feature extraction, sig...

AI Technical Summary

Problems solved by technology

[0004] With the exponential increase of equipment status data, the centralized cloud computing mode will cause large end-to-end delay and energy consumption, and is not conducive to the privacy protection of data, and cannot meet the real-time performance and reliability in the field of industrial equipment fault diagnosis and security requirements; the edge computing model can well solve the above-mentioned problems of the cloud computing model, but due to the limited computing power and storage capacity of the edge, it is impossible to directly deploy the diagnosis model based on deep neural network

These problems have brought great challenges to industrial equipment fault diagnosis

Images