面向可靠性评估和改进的试验设计和分析

发布时间：2018-03-03 18:01

本文选题：加速寿命试验　切入点：试验设计　出处：《天津大学》2014年博士论文　论文类型：学位论文

【摘要】：在全球化竞争和提升用户满意度的大背景下,产品要尽可能的可靠,尤其是是对于高科技和有安全风险的产品。企业强调降低产品全生命周期费用并且尽可能快地将产品推向市场。这样,如何经济有效地评估和改进产品的可靠性是迫切需要解决的问题。本博士论文在实践基础上,针对具体的问题,利用无偏因子法、两阶段法、自助法等,系统地研究了分组数据的分析方法和技术。这些研究对制造企业评估和改进产品可靠性具有很重要的意义。本论文的主要研究内容包括:1.证明了当各组寿命数据服从Weibull分布且数据类型为无截尾或type II截尾时,存在两个关于形状参数和尺度参数的枢轴量,并采用无偏因子法修正极大似然统计量。仿真结果表明,无偏因子法得到的统计量相对偏差和均方误差在绝大多数情况更小。2.改进了现有的加速寿命数据统计分析方法。首先,用改进的两阶段法和自助法分析加速寿命数据;其次,考虑了形状参数不固定的情况,并提出了简单有效的分析方法;最后,考虑了带有子抽样的情况,并提出用两阶段自助法获得置信区间。论文通过实例详细介绍推荐的方法,并通过仿真与其它方法进行比较。3.提出了面向可靠性改进的试验设计数据分析方法。首先,建立了响应和试验因子之间的关系模型;其次,考虑分位数或是平均故障前时间的望大,识别显著性因子;最后,用两个实例阐述推荐的方法:其中一个实例考虑了截尾的情况,另一个实例没有考虑截尾的情况。
[Abstract]:In the context of global competition and customer satisfaction, products should be as reliable as possible. Especially for high-tech and safety-risk products. Companies emphasize reducing product lifecycle costs and bringing products to market as quickly as possible. How to evaluate and improve the reliability of products economically and effectively is an urgent problem to be solved. On the basis of practice, this doctoral thesis uses unbiased factor method, two-stage method, self-help method and so on. The analytical methods and techniques of grouping data are systematically studied. These studies are of great significance for manufacturing enterprises to evaluate and improve product reliability. The main research contents of this paper include: 1. When distributed from Weibull and the data type is untruncated or type II truncated, There are two pivots about shape parameters and scale parameters, and the unbiased factor method is used to modify the maximum likelihood statistics. The simulation results show that, The statistical relative deviation and mean square error obtained by unbiased factor method are smaller in most cases. The existing statistical analysis methods of accelerated life data are improved. Firstly, the improved two-stage method and the self-help method are used to analyze the accelerated life data; secondly, the improved two-stage method and the self-help method are used to analyze the accelerated life data. Considering the uncertainty of shape parameters, a simple and effective analysis method is proposed. Finally, considering the case with sub-sampling, a two-stage self-help method is proposed to obtain the confidence interval. And compared with other methods through simulation. 3. A reliability improvement oriented test design data analysis method is proposed. Firstly, the relationship between response and test factors is established. Secondly, the relationship between the response and the test factors is established. Finally, two examples are used to illustrate the recommended method: one case considers the truncation case, the other example does not consider the truncation case.
【学位授予单位】：天津大学
【学位级别】：博士
【学位授予年份】：2014
【分类号】：TB497

【相似文献】