Complex electronic system fault diagnosis method based on fault propagation network

Abstract

The invention discloses a complex electronic system fault diagnosis method based on a fault propagation network, and the method comprises the steps: building a fault propagation directed graph of a complex electronic system through employing a graph theory, and achieving the storage and hierarchical optimization of the fault propagation directed graph through employing a matrix theory; simplifyingthe fault propagation network according to the fault phenomenon nodes; establishing a candidate fault source set by calculating the information output intensity and the working intensity of the nodes; and finally, calculating the comprehensive influence degree of the fault source by using an entropy weight method so as to determine a fault source therefrom. According to the invention, a fault propagation network is established based on a physical structure of a complex electronic system; the modeling problem of fault diagnosis of a complex electronic system is solved; design indexes such as the information output quantity and the fault-free working time of system components are utilized to solve the quantification problem of the system fault propagation influence relation, meanwhile, dependence on fault historical data is avoided, a fault source can be rapidly recognized and positioned, and the fault diagnosis capacity of a complex electronic system is improved.

Description

technical field [0001] The invention relates to the technical field of fault diagnosis of complex electronic systems, in particular to a fault diagnosis method for complex electronic systems based on a fault propagation network. Background technique [0002] A complex electronic system is a whole composed of multiple components (computers or electronic devices with computing power). When a component fails, the error information of the faulty component is often transmitted to one or more other components along the physical connection path. Information will be accumulated and amplified, causing other components to fail to work properly, which will cause a cascading effect of failures and ultimately affect the overall performance of the system. [0003] Due to the changing working environment and complex physical connection relationship, the main reason why a component in the system cannot work normally is caused by its own failure, such as hardware disconnection, short circuit...

AI Technical Summary

Problems solved by technology

[0010] The second is the problem of mutual influence of complex system test point signals

In order to improve the maintenance efficiency of complex systems, test points are usually set in limited positions of the system. The measurable signals (such as frequency, amplitude, etc.) of the test points come from multiple devices and are easily affected by factors such as the environment and electromagnetic interference. Difficult to handle

The method based on signal processing identifies a single fault based on the change of the measurable signal, and does not consider the interaction of signals when multiple faults occur, which limits the application of this method in complex systems

[0011] The third is the difficulty of obtaining complex system working condition data

However, the use of these methods requires the support of the actual operating status of the system and fault data: methods such as neural network and model logic require a large amount of data for training and testing, expert systems require working condition data and knowledge of domain experts, and graph theory methods need to use Fault Occurrence Probability Solving the Ranking Problem of Influence Degree of Multiple Fault Sources

The operating environment of complex systems is complex, some faults only occur under specific environmental conditions, and it is impossible to provide all the data required by the above method, so the artificial intelligence method needs to be further improved in the fault diagnosis of complex systems

[0012] At present, the system-level fault diagnosis method that adapts to the characteristics of large-scale complex systems, complex physical cross-linking, and meets the requirements of rapidity and accuracy has not yet been effectively solved.

Images