Design And Analysis On Accelerated Life Test Of The Gyroscope Motor

The research around design and inference of the accelerated life test has been carried out. The failure models and mechanisms of the gyroscope motors has been analyses, and the reliability model has been built. Based on above, the plan of the accelerated life test under multi-type stresses has been make and the data from the test has been statistical inference. Generally speaking, such five aspects of work have been done in this dissertation:1. From the analysis of the main failure models and mechanisms of the gyroscope motors, the environment stresses which affect the motors reliability deadly has been found out and used as accelerated stress. After referring to lots of literatures on motor reliability and the data of congener products, the reliability model of the gyroscope motor is built.2. The accelerated model under multi-type stresses is built after referring to literatures and experts. And the mathematic model of the optimum design of the accelerated life test under Weibull distribution is also induced.3. As the stress in the test is tress kinds, according to the national standards, at least three levels should be selected for each kind of stress. If the full factorial design scheme was carried out, too many tests should be experimentalized. Even if the test scheme based on orthogonal design theory was selected, there were too many tests to be accepted in engineering. Facing this problem, the tests plans based on Latin hypercube test design theory is selected in this dissertation. Five tests in all are carried out and a very good result is get.4. To carry out an accelerated life test of the gyroscope motor is an expensive project. In order to get the expected results, a simulation test to evaluate the test plans is necessary. Monte Carlo simulation tests are proposed to evaluate the tests plans. Monte Carlo simulation methods are proposed to calculate the elements in the Fisher information matrix. The precision and efficient are taken into account and the best test plan is found.5. The plot analysis method and the MLE are associated to analysis the data from the tests. The plot analysis method is easy and intuitional. The MLE has many statistical advantages. Deal with the complicated likelihood function, artificial immune algorithm is introduced to optimize the likelihood function. And the plot analysis method is used to evaluate the fitness of the model estimated by the MLE. The results inferred from the test data are consistent with the estimation given by the experts...