Equipment Fault Diagnosis Method Based on Online Learning of Fault Mechanism and Statistical Model

Abstract

A mechanical equipment fault diagnosis method based on online learning of fault mechanism and statistical model belongs to the field of mechanical equipment fault diagnosis. The specific steps are: (1) construct a health data generation model based on working condition characteristics and real-time status monitoring data; The fault mechanism and health data generation model are used to build a diagnostic model; (3) to identify the possible fault types of equipment; (4) to determine the probability of various types of equipment failures. This method combines the equipment failure mechanism with real-time operation data to construct a failure diagnosis model for a specific piece of equipment, which effectively solves the problems of insufficient model learning failure case data and poor model generalization ability in existing methods.

Description

technical field [0001] This patent belongs to the field of mechanical equipment fault diagnosis and relates to a mechanical equipment fault diagnosis method based on online learning of equipment fault mechanism and generative statistical model. Background technique [0002] After long-term development of mechanical equipment fault diagnosis, some progress has been made in the identification of equipment status. It has realized the distinction between the healthy state and the fault state of mechanical equipment, but there are still difficulties in the identification of fault types. The mathematical essence of fault identification is the problem of pattern classification, and its development mainly includes two stages: the stage of traditional diagnosis technology based on signal processing technology and the stage of intelligent diagnosis technology with artificial intelligence technology such as expert system and neural network as the core. At present, traditional diagnosti...

AI Technical Summary

Problems solved by technology

The invention combines the equipment failure mechanism with real-time state monitoring data, effectively solving the problems of less failure case data in the statistical model training process, easy over-fitting in training, and difficult generalization of the model

Images