基于多应力可变参数的加速寿命试验与可靠性研究

For the reason of that traditional accelerated model cannot explicitly express the life-stress relationship and that there are rare failure samples under lower stress level, the multiple stresses accelerated life test (ALT) with multiple variable lifetime parameters is introduced in this project which will improve the accuracy of extrapolation and test efficiency. By using classical statistical theory and methods, Bayesian theory, and with the aid of modern numerical technology, this project will study the statistical theory and methods of multiple stresses ALT as well as products reliability assessment. By analyzing the effect mechanism and interaction relationship of different stresses, we construct the multiple stresses and multiple variable lifetime parameters accelerated model which is the key of ALT analysis. Then the statistical models are established by using sampling theory and the characteristic of failure samples, and mathematical models of multiple stresses ALT design and acceptance sampling are also constructed according to different testing design and sampling criterion under constraints and non-constraints cases. Finally, through theoretical analysis methods and numerical computational techniques, the related theory and methods of theory statistical inference, test design and acceptance sampling are studied for multiple stress ALT with multiple variable lifetime parameters. This project provides new ideas and methods for accelerated life test techniques, and this study can make an extending part of the reliability test theory and its application. Thus the project has important theoretical significance and application prospects.

针对传统加速模型应力寿命关系刻画不准确，低应力水平下失效样本量小等因素导致加速寿命试验统计精度差、试验效率低等问题，本项目引入多应力加速寿命试验，建立多应力与多可变寿命分布参数加速模型，应用经典统计理论、贝叶斯分析和现代数值技术，研究多应力可变参数加速寿命试验及产品可靠性评估。本项目从应力对寿命特征的影响机制及应力间交互影响关系出发，研究多应力多可变参数应力寿命关系，这是实现多应力加速寿命试验的关键；然后根据失效样本特征，利用统计抽样理论建立统计模型。同时，在试验条件有无限制场合下，根据不同的设计准则和抽样准则，建立加速寿命试验优化设计和抽样检验数学模型；最后，通过理论分析和数值计算技术等手段，研究多应力可变参数加速寿命试验统计推断、优化设计和抽样检验问题的理论与方法。本项目的研究将为加速寿命试验技术提供新的思路和方法，丰富可靠性试验理论和应用研究，具有重要的理论意义和应用价值。

针对传统加速模型应力寿命关系刻画不准确，低应力水平下失效样本量小等因素导致加速寿命试验统计精度差、试验效率低等问题，本项目引入多应力加速寿命试验，建立多应力与多可变寿命分布参数加速模型，应用经典统计理论、贝叶斯分析和现代数值技术，研究多应力可变参数加速寿命试验及产品可靠性评估。本项目从应力对寿命特征的影响机制及应力间交互影响关系出发，研究多应力多可变参数应力寿命关系，然后根据失效样本特征，利用概率统计抽样分布理论建立各类不完全数据情形下的寿命试验统计模型。最后，基于经典统计推断理论、贝叶斯抽样方法和现代数值计算技术等手段，研究多应力可变参数加速寿命试验统计推断、优化设计和抽样检验问题的理论与方法。在项目研究期间，本课题取得了一系列的结果，相应的研究成果发表在SCI索引国际期刊上14余篇，指导和培养可靠性统计方向研究生4人。本项目的研究为加速寿命试验技术提供新的思路和方法，丰富可靠性试验理论和应用研究，具有重要的理论意义和应用价值。