Research On The Mechanical Reliability Growth Test And The Reliability Growth Model

Mechanical reliability growth technology is an important part in mechanical reliability theory. It exists in the whole lifespan of design, production and usage. Mechanical reliability growth technology aims to evaluate mechanical product reliability with the synthesized knowledge of different kinds, such as mechanical engineering, statistics and management. In the whole lifespan of the products, reliability growth test, data analysis and management are effective and economical methods to enhance the reliability stage.Since mechanical produce always has complex structure, many failure modes and poor commonality, the reliability growth test costs too much time as well as expenditure. So progress is relatively slow and has yet to form a unified methods or standards, which create difficulties for the practical application of engineering. Usually, the reliability growth method of mechanical products is similar with electronic products, and characters of mechanicals have to be taken into consideration.This paper carried out research on mechanical products, which life obeys the Weibull distribution or the log-normal distribution and the small-sample are required in the life test.With the help of classical reliability theory and practical research basement both domestic and foreign, the paper studied the issue of reliability growth test method, data analysis and the reliability growth modeling. Besides, examples are used to illustrate the effect of the method proposed in the paper. The main contents of the paper are stated as follows:(1) Compare with reliability growth theory and test method of electronic products, reliability characters of mechanical products are summarized in the paper. Research method of mechanical reliability growth is proposed as well.(2) Considering the life of mechanical products usually obeys the Weibull distribution and small sample data is obtained in reliability test, the paper chose Grey Estimation Method for three parameters estimation. Accordingly, Duane model is established for reliability growth evaluation. (3) Usually, fixed number test is common style in mechanical reliability test. Estimating method for small sample date is vital to find. Bayesian method for small data analysis is a quite effective way to solve this problem. Duane model is still used in this chapter to evaluate the situation of reliability growth.(4) In the reliability growth test, some of the date whose general obeys the Weibull distribution comes from small sample test, hence failure are hardly to observe. This non-failure situation is a vital problem in parameters estimation. Modified maximum likelihood estimation for parameters estimation of non-failure data analysis is initiated in the last but one chapter. Duane model which simple but efficient in evaluating the level of reliability growth.