| 摘要: | Following my project in August 1, 2010 to July 31, 2011, this research investigates the statistical inference of lifetimes of high-reliability products under the accelerated degradation test model. Considering the physical conditions and requirements in practical test situations for high-reliability products, this research will develop an inference procedure with three phases for accelerated degradation data based on the stochastic process methods. In Phase I, a training program was conducted for personnel. We set up experimental conditions for the constant-stress accelerated degradation test and collaborated with the Electronics and Opto-Electronics Research Laboratories, Industrial Technology Research Institute in Taiwan for collecting degradation information of high power light emitting diode lamps. The statistical model and inference methods will be developed, too. The Phase I project was granted by the National Science Council, Taiwan from August 1, 2010 to July 31, 2011. Till November 30, 2011, we implemented the accelerated degradation test over 3000 hours for light emitting diode lamps with six combinations of stress levels of ambient temperature and current. In Phase II, we will evaluate the quality of gathered degradation information of light emitting diode lamps and make a decision to develop an extra accelerated degradation test or not. Then start to conduct an intensive simulation study to evaluate the performance of the proposed method. Moreover, the proposed method will be illustrated by the degradation data of light emitting diode lamps. In Phase III, the performance of the estimated mean life and percentiles of the lifetimes of light emitting diode lamps using the proposed method will be compared with those using different existing methods in the literature. Moreover, a study will be conducted to determine the optimal design conditions including the sample size, the measurement frequency and the stopping time of test.
Following my project in August 1, 2010 to July 31, 2011, this research investigates the statistical inference of lifetimes of high-reliability products under the accelerated degradation test model. Considering the physical conditions and requirements in practical test situations for high-reliability products, this research will develop an inference procedure with three phases for accelerated degradation data based on the stochastic process methods. In Phase I, a training program was conducted for personnel. We set up experimental conditions for the constant-stress accelerated degradation test and collaborated with the Electronics and Opto-Electronics Research Laboratories, Industrial Technology Research Institute in Taiwan for collecting degradation information of high power light emitting diode lamps. The statistical model and inference methods will be developed, too. The Phase I project was granted by the National Science Council, Taiwan from August 1, 2010 to July 31, 2011. Till November 30, 2011, we implemented the accelerated degradation test over 3000 hours for light emitting diode lamps with six combinations of stress levels of ambient temperature and current. In Phase II, we will evaluate the quality of gathered degradation information of light emitting diode lamps and make a decision to develop an extra accelerated degradation test or not. Then start to conduct an intensive simulation study to evaluate the performance of the proposed method. Moreover, the proposed method will be illustrated by the degradation data of light emitting diode lamps. In Phase III, the performance of the estimated mean life and percentiles of the lifetimes of light emitting diode lamps using the proposed method will be compared with those using different existing methods in the literature. Moreover, a study will be conducted to determine the optimal design conditions including the sample size, the measurement frequency and the stopping time of test. |
