Research Of The Loading Ways And Statistical Analysis Method For Progressive Stress Accelerated Life Test

Accelerated Life Testing(ALT)s are widely used to study the reliability of products. ForProgressive Stress ALT the testing stress increases continually and thus has a high testingefficiency. At present, the Progressive Stress ALT is an important technology appied to assessthe reliability of product. The research of Progressive Stress ALT at home and abroad is focus onthe optimal design of testing plan and statistical modeling. However, the loading ways have notbeen systematically studied with the hope of obtaining a precise test results. Also, using StepStress ALT to approximate Progressive Stress ALT to avoid the limitation of Progressive StressALT’s high demanding for test equipment and achieve the same effect have not been studied.Besides, there are no works on how to gain a higher estimation accuracy than the optimumConstant Stress ALT by using Progressive Stress ALT. Therefore, in order to solve the problemin how to load progressive stress and analyze the test data, this paper takes aerospace electricconnector as engineering background, and the products whose life could be described bytwo-parameter Weibull distribution and Arrhenius equation or inverse power law as example,has systematically studied the loading ways and statistical analysis method for Progressive StressALT to obtain an efficient loading way which is easy to implement and has a high estimationaccuracy, and give a more accurate estimation method for parameters. This work provides atechnical guidance for the implemention of the Progressive Stress ALT for electric connector andother products with high reliability or long life.The main content in every chapter is:In Chapter1, the research background and significance are stated, then the main researchstatus of ALT is showed from the perspective of statistic analysis, optimal design andengineering application. And the current studies and its defects about Constant Stress ALT, StepStress ALT and Progressive Stress ALT are also introduced. Finally the main research content inthis paper is presented.In Chapter2, the basic model assumptions are given firstly, then, for the ease of theconstruction of Progressive Stress ALT’s statistical model, linearly changing transformed stressis adopted, and three common stress loading ways used in the Progressive Stress ALT are given,and the life distribution functions under the three kinds of progressive stresses are derived,finally the parameters’ estimation method is presented.In Chapter3, the optimum test plans of the three kinds of Progressive Stress ALT are foundto minimize the asymptotic variance of the MLE of the natural logarithm of a specified quantileat design stress. The effects of the pre-estimates of design parameters are studied. Then, the threekinds of loading ways are compared from the perspective of estimation accuracy, robustness, complexity of data analysis and so on. It can be concluded that each loading way has its merits,however, the second loading way whose transformed stress increases linearly and has transitionis the best comprehensively, for its high estimation accuracy, good effect of acceleration andease of implemention and control.In Chapter4, the Progressive Stress ALT simulation methods are presented based on MonteCarlo method, and to the problem that maximum likelihood estimation method has larger error insmall sample situation, the small sample errors of maximum likelihood estimation method areinvestigated utilizing the simulation method and the corresponding correction methods areprovided.In Chapter5, the Step Stress ALT and Progressive Stress ALT are compared from theperspective of accelerated effect and estimation accuracy, and it can be concluded that a StepStress ALT with a dozen of steps is almost equivalent to Progressive Stress ALT, whicheffectively solves the problem of Progressive Stress ALT’s loading difficulty.In Chapter6, the relationship between Constant Stress ALT and Progressive Stress ALT isstudied, which shows that Constant Stress ALT could be viewed as a special case of theProgressive Stress ALT when its stress rates is zero. As the optimum Constant Stress ALT isconsist of two tests with different stress levels, it can be considered as a special case of aProgressive Stress ALT with two different Progressive Stress, then, by optimizing theProgressive Stress ALT, a new ALT which contains both constant stress and progressive stressand has a higher estimation accuracy than the optimum Constant Stress ALT is obtained, and theProgressive Stress could be replaced by a Step Stress with several steps.In Chapter7, the main content is summarized, and the further research directions are raised.