Dissertation > Excellent graduate degree dissertation topics show

Bayesian Approach to a Constant Stress Accelerated Life Test Based on Gibbs Sampling under Log-Normal Distrbution

Author: SunYang
Tutor: ZhangKuiYuan
School: Jilin University
Course: Applied Mathematics
Keywords: Lognormal distribution Constant stress accelerated life testing Bayes estimation
CLC: O29
Type: Master's thesis
Year: 2004
Downloads: 294
Quote: 4
Read: Download Dissertation

Abstract


Accelerated life test is to shorten the test cycle with increasing test stress life test if you use it, the general reliability of the product can be quickly assessed out quickly identify the cause of failure, high reliability and long life reliability assessment made possible .70, accelerated life testing entering the country, immediately aroused the interest of the engineering and statisticians side edge research, electronic, mechanical, instrumentation and other industries for 20 years, has accumulated a lot of practical experience in the accelerated life test , made a number of new accelerated life test model and data processing methods, these have greatly enriched the theory and method of accelerated life testing. accelerated life testing type, commonly used is the following three categories .1. constant stress accelerated life testing, referred to as an accelerated life test it is to elect a group of accelerated stress level, for example,, ...,, they are higher than normal stress level, generally there lt; lt; ... lt; lt;, then divided into a number of samples groups, each group in an accelerated stress level life test, until the groups have a certain number of products lapses .2 step stress accelerated life test, referred to as step - stress, it is also the first selected a group of accelerated stress horizontal, ..., they are higher than the normal stress level, the start of the test, life test is carried out to a certain number of samples are placed under stress level after a period of time, such as hours, the stress is increased to the unexpired sample to proceed under stress life test so continue until a certain number of the sample lapsed .3. progressive stress accelerated life test, referred to as stress life tests plus with step-by-step test is basically the same, the difference is only that they applied accelerated stress level will rise continuously over time the most simple, straight up. stress life tests stress change, failure comes, it is the implementation of the Accelerated test requires specialized equipment control stress levels and recording equipment tracking of product failure, there is no such device, Accelerated test difficult to implement the above three applications, there are a number of successful examples of accelerated life test, but an accelerated life test instance. an accelerated life test method is simple, data processing method is also more successful. although it required test time is not the shortest, but still double than the average life test to shorten the time, it is often the actual use of an accelerated life test has also developed a national standards. No. GB2689.1 ~ 4-81. Bayesian statistical analysis methods of statistical inference difficulty lies in the posterior of the calculation of the marginal distributions, in many cases it is difficult, or even impossible to get the posterior marginal distributions analytical expressions sometimes may be resolved but the result is quite complex and not easy to application and promotion. obey the lognormal distribution conditions, the life of the product, constant stress accelerated life test in this case of the non-censored data of Bayesian statistical analysis of the data, and the use of Gibbs sampling iterative algorithm to solve extremely complex after inspection marginal distributed computing text is divided into six parts:

Related Dissertations

  1. Based on statistics of the lognormal distribution heteroscedasticity model inferred,O212.1
  2. Exponential distribution made ??up several Censored Accelerated test of quadratic estimates,O211.3
  3. Mixed Exponential Distribution under Censored Accelerated test of quadratic estimates,O211.3
  4. Quasi-brittle material strength size effect of statistical and numerical simulation approach,TU502
  5. Jump-diffusion model set of financial products broker pricing,F832.39
  6. Research on Methods to Avert and Prevent Malicious Bidding in Tenders,F284
  7. Statistical Analysis for the Generalized Exponential Distribution Based on Accelerated Life Test Data,O212.6
  8. Bayes Estimators of the Three-parameter Weibull Distribution Based on Type I Censoring Sample,O212.8
  9. Several distribution parameters and application of Bayes statistical inference,O212
  10. Parameter Estimation of Commonly Used Survival Distribution Based on Randomly Censored Data,O212
  11. The Statistical Analysis of Zero-Failure Data,O212
  12. An Empirical Study on Deciding Stuffing Numbers in a Call Center,C934
  13. Conclusion of Censored Data of Exponential Distribution under a Class of Loss,O211.3
  14. The Estimation about a General Class of Distributions in the Case of Censored Data,O212
  15. Fuzzy reliability study,TB114.3
  16. Statistical analysis of incomplete data,O211
  17. Step increases and statistical analysis of the sequence plus composite test,O211
  18. Parameter Estimation for Weibull Distribution Based on Incomplete Samples,O212
  19. Research of the Statistical Data Quality Inspection and Evaluation,C829.2
  20. The Research on New Terrain Elevation Matching Approaches and Their Applicability,TN966

CLC: > Mathematical sciences and chemical > Mathematics > Applied Mathematics
© 2012 www.DissertationTopic.Net  Mobile