System Belief Reliability Evaluation Method Based on Uncertainty Theory

On the basis of probability theory, fitting data based on the law of large numbers can quickly obtain the distribution or variation patterns of data. However, in the system design phase, the system has not yet been put into use, and reliability cannot be analyzed through statistical fault data. The performance data of various functions and modules in the design phase is also relatively limited. In this case, directly using probability theory methods to fit the data would result in significant cognitive uncertainty. Therefore, this article proposes a method for analyzing the reliability of test data from the perspective of functional performance, based on uncertainty theory. This method collects data on key parameters of the system and fits their distribution and two factor degradation patterns. From the perspective of functional margin, it establishes disciplinary equations for performance parameters and injects the distribution and degradation patterns of key parameters to conduct reliable analysis. This article takes integrated operational amplification circuits as a case study to compare the analysis results based on uncertainty theory and probability theory. The results indicate that in the case of limited data, uncertainty theory can better describe cognitive uncertainty in fitting, and reliability analysis results are more conservative.