Research On Key Technology Of Prognostics And Health Management For Digital Chips

The deterioration of digital chips function seriously affects the normal operation of electronic equipment. Digital chips failure caused great losses in some important applications. In order to improve the stability of the electronic equipment and reduce electronic equipment maintenance costs and extend the service life of electronic equipment, the key technology of prognostics and health management for digital chips researches on the influence of impact factors on the characteristics factors, researches and summarizes the variation law of characteristics, improves the stability of digital chips. This paper studies the influence of temperature parameters on the performance parameters of digital chips and the changing tendency of frequency with temperature and operating time, and prognostics the health of digital chips.This paper mainly uses data-driven method study failure prediction and health management technology for digital chips. This paper studies thorough analysis on the digital chips fault type and fault feature, selects the output frequency as characteristic parameter of digital chips and designs core algorithms to set up a model for mining characteristic parameters of hidden information, explores the rules of digital chips function deterioration and performance degradation. In this paper, we use the fault prediction and health management model, try to judge the working state of the digital chips and predict the working state of the digital chips. According to the results of the model and relevant experience, this paper puts forward maintenance suggestions. To prolong the service life of the digital chips, reduce the maintenance cost of the electronic equipment, and avoid the impact and loss caused by the digital chips failure.In this paper, the neural network algorithm in the field of machine learning artificial intelligence is used as the core algorithm to approximate the characteristics of the nonlinear function, and find out the hidden information of the characteristic parameters. In order to complete the experiment, we design the software and hardware platform for the measurement of characteristic parameters. The hardware research platform is composed of temperature sensor, digital chips, computer and other electronic components. Software research platform includes independent data receiving software and data processing program.The experimental results show that the error of the temperature fault model is about 3%, and the error of fault prediction model is about 7%. Experimental results show that frequency is inversely proportional to the surface temperature of the chip, the ambient temperature, and the length of operating time. The research platform can effectively measuring and recording data, and the research platform can effectively accomplish the function of prognostics and health management for digital chips. The two models can effectively realize the function of calculating and predicting the delay parameters. Because the fault prediction and health management for digital chips invloves the knowledge of many different fields, the researchers need to constantly in-depth study and strengthen the related design in this article.