Circuit failure diagnosis method based on neural network

Abstract

The invention discloses a circuit fault diagnosis method based on a neural network, wherein current excitation or voltage excitation is exerted on a to-be-tested circuit; the voltage of a testable junction is measured and is compared with that of a normal condition to obtain voltage increment; according to the Kirchhoff theorem of circuit, the relation equation of circuit device parameters and excitation and response is obtained; then, element parameter increment is taken as an optimization variable; the relation equation is solved through adopting a constraint nonlinearity noncontinuous optimization method to obtain element parameter increment; finally, the increment is compared with a normal parameter to determine a fault element. The circuit fault diagnosis method has the characteristics of strong anti-interference capability, fast location and high accuracy, etc.; moreover, the method has wide application range and is suitable for the test and the diagnosis of a linear circuit, a nonlinear circuit and a mixed signal circuit.

Description

technical field [0001] The invention relates to a circuit fault diagnosis method, in particular to a neural network-based circuit fault diagnosis method. Background technique [0002] The parameter identification method is an important method for electronic circuit fault diagnosis. Generally, the component parameters are obtained by solving the circuit characteristic equation. However, since the scale of the circuit is getting larger and larger, and its circuit characteristic equation is generally a large-scale nonlinear equation, the solution is often time-consuming and has the disadvantages of large errors. The emergence of neural network has improved a new and important method for circuit fault diagnosis, but the existing neural network still has great shortcomings in solving circuit characteristic equations. For example, the approximate processing of the absolute value function will produce a large error, and in order to improve the accuracy of the solution, it is neces...

AI Technical Summary

Problems solved by technology

However, since the scale of the circuit is getting larger and larger, and its circuit characteristic equation is generally a large-scale nonlinear equation, it is often time-consuming to solve and has the disadvantages of large errors.

The emergence of neural network has improved a new and important method for circuit fault diagnosis, but the existing neural network still has great shortcomings in solving circuit characteristic equations.

For example, the approximate processing of the absolute value function will produce a large error, in order to improve the accuracy of the solution, it is necessary to increase the penalty parameter infinitely, etc.

Images