Accelerated Degradation Test of ANN-Wiener Process Considering Measurement Uncertainty
Considering the measurement uncertainty of key performance parameters in the process of accelerated degradation test,the measurement uncertainty was treated as interval number,and a reliability evaluation meth-od of interval accelerated degradation data combining artificial neural network and Wiener process was estab-lished.Based on the ANN-Wiener process,the negative log-likelihood function of the accelerated degradation data was constructed.and the genetic algorithm was used to establish the solution method of the unknown pa-rameters of the negative log-likelihood function.Finally,the reliability evaluation of the interval accelerated deg-radation test was realized.The proposed method was verified by an example of accelerated degradation test of la-ser.The absolute errors of reliability and true reliability obtained by the proposed method and other methods were compared.The results showed that the evaluation results of reliability based on ANN-Wiener process were more accurate.Considering the influence of measurement uncertainty,the conservative estimation and optimistic estimation of laser reliability were obtained.Comparing the evaluation results when ignoring the measurement uncertainty with the evaluation results of the conservative estimation,under the same reliability,the evaluation time when ignoring the measurement uncertainty was later than the evaluation time of the conservative estima-tion,which would lead to the delay of the preventive maintenance time of the product,thus increasing the failure risk during the operation of the product and increasing the loss caused by the failure of the product.
万方数据
维普
