Circuit health state prediction method and system based on integrated deep neural network

Abstract

The invention discloses a circuit health state prediction method and system based on an integrated deep neural network, belongs to the field of power electronic circuit fault prediction, and aims to realize health state identification and diagnosis of an analog circuit through an integrated deep neural network based on historical data. The method comprises the steps of: carrying out parameter aging simulation experiments of different devices; extracting a series of time domain features of output signals by using a time sequence transformation method, and establishing a health index of the devices based on the improved angular similarity; carrying out health state prediction on an analog circuit in degradation by combining a convolutional auto-encoder and a long-short-term memory cycle network; and evaluating the effectiveness of the circuit health state prediction method by referring to related evaluation indexes. The method can effectively predict the health state of the analog circuit, and has the characteristics of high precision and easiness in implementation.

Description

technical field [0001] The invention belongs to the field of fault prediction of power electronic circuits, and more specifically relates to a method and system for predicting circuit health status based on an integrated deep neural network. Background technique [0002] In today's information age, the degree of informatization of manufacturing, automobiles, and power grids has become unprecedentedly complex with the development of integrated circuits. This is due to the fact that the mutual coupling of electronic components in the system brings reliable and efficient operation of equipment. challenge. Moreover, the aging and degradation of components in electronic circuits may cause very serious disasters, which need to be given sufficient attention. [0003] In analog circuits, various components: capacitors, resistors, inductors, power switches, etc., may experience aging and performance degradation. Various degradation modes have different impacts on circuit performanc...

AI Technical Summary

Problems solved by technology

Finally, although the hybrid forecasting method combines the advantages of the two methods, it is still largely affected by the model-based method, such as: high computational complexity, parameter sensitivity, and modeling difficulties.

Images