Abstract: Since the data-driven life prediction methods were modeled by linear degradation model and the actual degradation data were mostly nonlinear,a new nonlinear degradation model was proposed.The residual life probability density function of the proposed new mode was derived by using the first passing time principle,the unknown parameters in the nonlinear model were estimated with maximum likelihood method,and the accuracy of the model was verified by AIC criterion.The simulation data and the performance degradation data of the laser generator were used to verify and analyze the model.The results showed the accuracy and feasibility of the new nonlinear degradation model.

Key words: nonlinear degradation modeling, parameter estimation, residual life prediction, laser generator

摘要： 鉴于数据驱动的寿命预测方法多采用线性退化模型建模,而实际退化数据多呈现为非线性形式,为此,提出一种新型非线性退化模型,利用首达时间原理推导出模型剩余寿命概率密度函数的解析表达式,采用极大似然法估计模型中的未知参数,并用AIC准则检验了模型的准确性。通过仿真数据与激光发生器的性能退化数据对模型进行了验证分析,结果表明了新型非线性退化模型的准确性和实用性。

关键词: 非线性退化建模, 参数估计, 剩余寿命预测, 激光发生器