The Reliability Growth Analysis Of The Chemical Battery

Chemical battery has a wide range of application in communication, electronics,automobile, military, aerospace, etc. As the increasing depth of its application scope anddegree, the reliability of the battery is getting more and more attention. This thesis is basedon the background, researching the reliability growth theory in the application of thechemical battery.Firstly, this thesis introduces the basic theory of the reliability growth and the reliabilitymodel---Duane model and AMSAA model, focusing on the AMSAA model of a singleproduct, AMSAA-BISE and AMSAA model of the multiple products. Then, we alsointroduce the structure of the Li-ion battery and the common analysis methods in thereliability growth process, which can improve the reliability of the battery effectively.For that it is difficult to obtain the reliability growth data of the chemical battery, thethesis proposes a method that combines the inverse transformation technology with theWeibull process to simulate the reliability growth data. Then, we use the MATLABsoftware to simulate the data based on the Fixed Time Censoring and Fixed NumberTesting, respectively. Also, we calculate and analyze the data of the single product andmultiple products, discussing the influence of the model parameter variation.However, because both Duane model and AMSAA model are the model which is basedon the product failure time, so they are only suitable for dealing with the continuous timereliability growth based on TAAF mode. While for the phased reliability growth mode withdelayed correction, only using the model based on failure time cannot fully use the existingtest information in all stages, which is a flaw in the application.Aiming at the above problem, the thesis comes up with a phased Bayes reliabilitygrowth model of the product combining the Bayes method. The traditional reliabilitygrowth assessment method makes the assessment result more conservative, while the Bayesreliability growth model considers both the prior information and the posterior information,which can dynamically evaluate the reliability growth process of the battery and predict the reliability level of the battery, and the evaluating results are more realistic.