面向一类混合退化装备RUL预测的平行仿真技术

RUL Prediction Oriented Parallel Simulation Technology for Hybrid Degradation Equipment

针对一类带离散冲击的混合退化装备剩余寿命预测问题,研究了面向混合退化装备剩余寿命预测的平行仿真技术.提出以混合Wiener状态空间模型为基础仿真模型,以泊松冲击到达为模型选择判据,在实时退化数据驱动下,实现仿真模型在线选择,利用强跟踪滤波和期望最大化算法进行仿真模型数据同化和未知参数在线估计,从而实现仿真模型演化,提高仿真模型逼真度.在此基础上,实现了基于平行仿真的剩余寿命实时预测.利用某轴承性能退化数据对平行仿真方法进行了实例验证,仿真结果表明平行仿真方法能有效仿真轴承的性能退化过程,剩余寿命预测的不确定性小、精度高.

Aiming at the remaining useful life (RUL) prediction issue of hybrid degradation equipment with discrete shock, a parallel simulation technology was studied to predict the RUL of hybrid degradation equipments. Firstly, a simulation model was proposed based on a hybrid Wiener state space model, taking the arrival of Poisson shock as the selection criteria of the model. Driven by the real-time degradation data, the online simulation model selection was carried out, realizing the data assimilation of simulation model and online estimation of unknown parameter with a strong tracking filter and expectation maximum algorithm. And it was accomplished to evolve the model and improve the fidelity of simulation model. As a result, the parallel simulation based real-time prediction of RUL was realized. Then, utilizing the degradation data of a bearing, the proposed method was verified. The simulation results show that the parallel simulation method can simulate the performance degradation process effectively, and can predict the RUL with less uncertainty and high accuracy.