Wiener process reliability analyzing method considering self-correlated measurement errors

Abstract

The invention provides a Wiener process reliability analyzing method considering self-correlated measurement errors. The problem that an existing performance degradation analyzing method severely influences reliability evaluating precision is solved. The method includes the steps of randomly extracting a sample from a batch of products to be put in a test, collecting product performance degradation data, describing the real performance degradation process of the products according to the Wiener process, establishing a Wiener process performance degradation model considering the self-correlatedmeasurement errors, estimating unknown parameters of the performance degradation model through a likelihood function, and analyzing the product reliability according to the estimated unknown parameters. The method is a degradation reliability analyzing method of a linear Wiener process of the AR (1) measurement errors, the evaluating precision is effectively improved, universality is achieved, adaptability is higher, a reliability function analyzing expression under a given degradation failure threshold is given, and the basis is provided for subsequent reliability evaluation of the products.

Description

technical field [0001] The invention relates to the technical field of small sample reliability analysis, relates to a reliability analysis method based on performance degradation data, in particular to a Wiener process reliability analysis method considering autocorrelation measurement error. Background technique [0002] In aerospace, weaponry and other fields, the characteristics of small samples, high reliability and long life of products have become increasingly prominent, which has brought great challenges to traditional reliability analysis methods based on failure time. Performance degradation modeling is one of the key technologies to solve the reliability analysis of such products, and it is also a hot and difficult point in the field of reliability engineering research. [0003] Considering that the performance degradation process of products is often affected by many random factors at the same time, degradation modeling based on stochastic process theory is the f...

AI Technical Summary

Problems solved by technology

[0004] However, in all current degradation models considering measurement errors, it is assumed that the measurement errors at different times are independent of each other and obey the same normal distribution

However, this is inconsistent with many actual situations, especially when the test time interval is short, the autocorrelation between continuous test measurement errors is often not negligible, and the independence assumption will seriously affect the accuracy of reliability analysis

In engineering practice, the measurement errors in performance degradation data usually have autocorrelation, and sometimes cannot be ignored. However, the existing performance degradation analysis methods all assume that the measurement errors between different test times are independent of each other and obey the same normal distribution, which can seriously affect the reliability assessment accuracy, especially when the test interval is short

Images