The Fault Diagnosis Of Millimeter Wave Radiation Characteristic Testing System

The paper is built on our research group’s actual project that is the test system for millimeter-wave radiation characteristics. Based on the research of the system’s structure and working process, the requirements of the fault diagnosis and some fault parameters that can be detected are proposed. Combined with the knowledge-based fault diagnosis method, the design of the target system’s fault diagnosis has been completed.Firstly, the system structure of millimeter-wave radiation characteristics of the test system is researched. In view of its overall structure’s hierarchical modular design features and some analog circuits’requirements for fault diagnosis, the diagnosis scheme is proposed, that is using the fault tree analysis to complete the system-level fault diagnosis and BP neural network method to achieve the fault diagnosis of the specified analog circuits.Secondly, according to the established scheme, a proper diagnosis of the fault tree is built by layer-by-layer analysis to find the failure’s cause of selected top fault event and the fault tree’s qualitative and quantitative analysis is completed. After that, the prioritization of end-events’fault diagnosis is given in accordance with their degree of importance of the structure.Thirdly, a proper diagnosis of BP neural network aims at radiometer’s intermediate frequency amplifier and video amplifier circuit is created. By ascertaining the suitable network parameters which are based on the diagnosis theory and the analysis of experiment results, the network training process is accomplished. Then, a large number of sample data for different fault types is tested to verify the network’s diagnosis effect due to the tolerance effect of the circuit components, and the result shows that the network has a high diagnostic accuracy.Finally, experimental methods to detect the test system’s fault diagnosis parameters are given, such as the way to gets sensitivity, signal-to-noise ratio, integration time and so on, which provide the support for system fault diagnosis to makes its failure events can be detected or monitored. In order to facilitate the operation of the user’s actual diagnosis, the UI(User Interface) of the fault diagnosis is designed, which actually can improve the diagnosis system’s practicality and integration.