Research On Remote Real-time Fault Diagnosis Method Of Electronic Devices

New electronic devices play an important role in the development and use of weapons and equipment. By remotely monitoring the status of the electronic devices and real-time diagnosing faults, it can be realized that to ensure the weapons and equipment always in the good technical condition, to improve the life and reliability of electronic devices, to reduce the maintenance cost, and improve the maintenance efficiency as well. Also, it has a significant value in engineering applications.In this paper, aimed at the fault diagnosis algorithm of electronic devices and the hardware-software implementation of the remote real-time electronic fault diagnosis system, some studies are launched.1) Based on the fault diagnosis principle of the analog circuit in electronic devices, a new idea of the fault detection and diagnosis is proposed by adopting the circuit fault simulation. Some experiments of the fault simulation and detection are conducted, which lay the foundation for the subsequent fault diagnosis.2) The method to extract fault feature vectors is studied, by applying the wavelet theory. By extracting the sum of the absolute values of the wavelet decomposition coefficients of each layer, the wavelet transform generates fault feature vectors of the analog circuit. And by reconstructing the decomposition coefficients of the wavelet packet and extracting the signals from each frequency range, the wavelet packet decomposition method can calculate the total energy of each frequency signal as the fault feature vectors of the analog circuit. It is proved that the wavelet packet decomposition method possesses the effectiveness and the superiority via the experimental comparison and analysis.3) The fault diagnosis principle with the neural network and BP algorithm are elaborated. Aimed at the limitations of BP algorithm, a modified BP neural network is proposed to be adopted for the fault diagnosis method and its design procedure. In addition, the performance test of the designed neural network is successfully completed. And it is verified that this method can be effectively adopted for the fault diagnosis of electronic devices.4) Combining with the development background of one product, a remote real-time fault diagnosis system based on FPGA is designed for electronic devices. And this fault diagnosis method is realized with hardware and software. What’s more, it is verified that the proposed theoretical method possesses the effectiveness with the diagnosis applications, and it is suitable for engineering application and possesses certain practical significance.