基于粒子滤波的疲劳裂纹扩展和剩余寿命预测实验研究

本文对应用粒子滤波在线预测疲劳裂纹扩展和剩余寿命进行了实验研究。根据ASTM标准制作了紧凑拉伸(CT)试件,结合MTS疲劳测试系统进行了Q235钢的疲劳裂纹扩展实验。通过将传统的Paris模型嵌入到贝叶斯状态估计的框架中,采用粒子滤波方法,结合裂纹张开位移(COD)引伸计不断获得的裂纹监测信息,在线预测未来时刻疲劳裂纹扩展情况并更新剩余寿命。实验结果验证了粒子滤波对疲劳裂纹扩展和剩余寿命在线预测的可行性和有效性,表明粒子滤波不仅能进行在线预测,还能给出预测结果的置信区间;随着监测信息的增多,剩余寿命预测结果的不确定性逐渐减小。

Experimental study on prediction of fatigue crack growth and residual life by using particle filter

This paper presents an experimental study on online prediction of fatigue crack growth (FCG) and residual life (RL) by using particle filter (PF). According to ASTM standard, compact tension (CT) specimens are manufactured and FCG experiments for Q235 steel are performed by MTS fatigue test system. By embedding the conventional Paris model into the Bayesian framework, PF is employed to online predict FCG in the future and update the estimate of RL with continuously monitoring data from a crack open displacement (COD) gauge. Experimental results have validated the feasibility and effectiveness of the PF-based online prediction method, demonstrating that PF can not only give prediction results, but also their confidence intervals;with the increase of monitoring data, the uncertainties of RL predictions are progressively reduced.