Research On Radar Mixing-Circuit Fault Diagnosis

In high-tech modern war, over-the-horizon combat in campaign especially in air battle has become a mainstream mode of war. It has become a cherished thing that did always "first step enemy" between the two sides. Therefore, as a high-tech, all-weather detection equipment, the radar is getting more and more recognition by military experts, the accuracy, reliability and rapidity of radar has been most concerned by scientific and technological personnel. If the two sides of war are two boxers, then the radar systems are their "eyes". The reliability, maintainability and mensurability of radar systems determine the outcome of a war, and are getting more and more concern by researchers, and gradually become one of the hot researches. The problems that radar system failure diagnoses have been investigated by using artificial intelligence control theory and signal processing methods. Aiming at some problems of radar system such as complex structure, difficult to establish precise mathematical model, high frequency signals and too strong noise, the new research methods of radar system failures diagnosis is brought forward based on signal processing, neural networks and expert systems. The author studies the radar system, and respectively establish a radar failure diagnostic expert system, radar board level fault diagnosis system and radar components level fault diagnosis system, which prove that diagnostic methods is proper and the same with engineering application.Establishing radar failure diagnostic methods on the basis of a certain type of radar systems expert system, and a radar system failures diagnostic fault tree model is established, which not only avoid failure as a result of complex systems by building models to the problem, but also significantly improve the efficiency of the diagnostic system.Artificial neural network theory is applied to radar failure diagnosis. A radar components level failure diagnosis model of a certain type of radar is established based on BP neural network. Making circuit simulation use pspise software, training data use real data, and complete the setting of network parameters.A board-level fault diagnosis model of the radar is established based on signal processing. Input and output signal of radar circuit boards is made spectrum analysis, and then made signal relativity analysis to determine whether the failure of the circuit boards, in which units to malfunction.The author uses virtual instrument technology. Based on the study results of radar system malfunctions diagnostic, the author develops an intelligence failure diagnosis system of the radar. The software use mixed programming techniques, matlab software is embedded in c++builder software, and using sql server2000 database. An intelligence failure diagnosis system of the radar is developed successfully, and achieves the radar system malfunctions diagnosis. The diagnostic systems have been delivered, and achieved good results.