Special Cases in Failure Analysis: Research on the Optimal Confidence Limit Evaluation by Zero-Failure Data
- PDF / 609,420 Bytes
- 6 Pages / 595.276 x 790.866 pts Page_size
- 43 Downloads / 211 Views
TECHNICAL ARTICLE—PEER-REVIEWED
Special Cases in Failure Analysis: Research on the Optimal Confidence Limit Evaluation by Zero-Failure Data Haiyang Li . Zeyu Zheng . Zihao Yang . Chen Xu
Submitted: 21 April 2020 ASM International 2020
Abstract For the failure analysis of products with high reliability requirements, it is often encountered that the failure data cannot be obtained for a long period of time. Based on this situation, the reliability evaluation method of zero-failure data has been developed and achieved certain results in the actual engineering application process. In this paper, when the product life is subject to Weibull distribution, the reliability evaluation of the bearings is obtained according to the optimal confidence limit evaluation method. A method to extend the application of optimal confidence limits is proposed. Combined with the results of numerical simulation case studies, the results indicate that the type of zero-failure data grouping has a direct impact on the reliability evaluation results, and the reliability estimate is closer to the real situation in a certain period of mission time. Keywords Reliability evaluation Zero-failure data Weibull distribution Confidence limit assessment method
Introduction With the improvement in manufacturing technology and the application of high-quality materials, the reliability level of bearing products has been significantly improved [1]. For the failure analysis of products with high reliability requirements, it is often encountered that failure data cannot be obtained for a long period of time [2]. Although the product has not yet failed, there is a need for reliability H. Li (&) Z. Zheng Z. Yang C. Xu Digital Factory Department, Shenyang Institute of Automation, Chinese Academy of Sciences, Shenyang 110169, China e-mail: [email protected]
evaluation and life prediction in practical engineering applications. The evaluation based on zero-failure data is a special case in the failure analysis [3]. According to the test samples (complete sample data) in the case of failure in life test, there is a set of mature methods for the point estimation and confidence limit estimation of reliability parameters which obey various common distribution types [4]. Martz and Waller [5] first put forward the reliability control test method of zero-failure data under the condition of exponential distribution, and selected the priori distribution of failure probability according to the posterior distribution risk criterion, and then got the estimation of the failure probability. In order to cope with the limitation of insufficient data samples, the method of numerical simulation can increase the sample size of the research object during the reliability evaluation of zero-failure data [6]. The current commonly used methods for reliability analysis of zero-failure data include the sampling inspection methods [7], the life distribution parameter estimation methods [8, 9], and the optimal confidence limit evaluation theory [10]. Under the condition that
Data Loading...
