Some Research On Reliability Of Electronic Product

There are many ways on the research of the reliability,This paperresearchs the reliability of the system and product,it from two aspects:Accelerated Life Test and The Confidence Limit.The Accelerated Life tests isby changing the stress (temperature, voltage, etc) make the product more easyto failure,and we get the data we want in the relatively short time. then,according to the distribution of products subject to specific and acceleratemodel, get reliability of products in the normal stress.Then combined with theaverage life expectancy of the product, and found that both differ is not big,explain this test is feasible. According to the stress level is different,Accelerated life tests can be divided into three types:(1) the constant stressaccelerated life tests,;(2) step into stress accelerated life tests;(3) sequence intostress accelerated life tests. Because of the experimental equipment and testsequence demand is higher, so they were the first two ways of test. Underdifferent index distribution situations, stress was timely converted in step-stressaccelerated life tests to obtain failure data of products under different stresses,which were accordingly converted into the data under constant-stress. Then, reliabilities of products’ average life expectancy under normal stress wereachieved with these data by using constant-stress calculation method. At last,products could be inspected by the achieved reliability data. Their purpose areto get life in the normal stress, with the data in different stress tests, throughthe acceleration model and a series of into the process, finally get in normalstress of life under the formula, and the calculated results coincide with theexpectation that the feasibility of the experiment.For research system confidence limit, this system has a relativelyindependent components and a conversion switch composition, switch issuccess or failure model components, and the life of the components to obey thesame parameters of geometric distribution, based on the system has terminatedtesting experiment data, and the conversion switch, of the geometric distributionreserves the point estimation and the approximate system confidence limitconcluded. Find the data of the simulation results and practical consistent withthat the feasibility of the experiment.