论文部分内容阅读
飞行器结构的疲劳裂纹扩展预测对保障结构安全、实现视情维护具有重要意义。结合粒子滤波算法和结构健康监测方法进行在线的疲劳裂纹扩展预测是近年来刚刚开始研究的新方法,该方法通过状态空间模型表征疲劳裂纹扩展过程中的不确定性,同时通过贝叶斯方法将结构健康监测所获取的结构实际裂纹观测值用于修正裂纹扩展模型的预测误差,实现更准确的疲劳裂纹扩展在线预测。由于该方法的研究刚刚开展,已有研究中粒子滤波算法的重要性密度函数往往简单选取为先验转移概率密度,存在严重的粒子退化问题。另一方面出于简单考虑,仅采用表征裂纹稳定扩展区的Paris模型。针对上述问题,本文提出一种基于高斯权值-混合建议分布粒子滤波的疲劳裂纹在线预测方法,基于表征裂纹全扩展区域的NASGRO裂纹扩展模型建立疲劳裂纹扩展状态方程,以主动Lamb波监测方法实现结构裂纹的在线监测,借助在线结构健康监测的优势,在粒子滤波时选取重要性密度函数为观测概率密度和先验转移概率密度的混合分布,同时基于先验估计获取高斯权值进行权值更新。本文进一步进行了仿真研究,结果表明所提出的方法优化了疲劳裂纹扩展预测的准确性。
Prediction of fatigue crack growth in aircraft structures is of great importance to ensure structural safety and maintain conditions as appropriate. The combination of particle filter algorithm and structural health monitoring method for online fatigue crack growth prediction is a new method just started to study in recent years. This method characterizes the uncertainty of fatigue crack propagation through the state space model. At the same time, The actual structural crack observation obtained by structural health monitoring is used to correct the prediction error of the crack growth model and achieve more accurate online prediction of fatigue crack growth. Due to the research of this method has just been carried out, the importance density function of the particle filter algorithm in the existing research is often simply selected as the prior probability density of transfer, and there is a serious problem of particle degeneration. On the other hand, for simplicity, only the Paris model, which characterizes the stable propagation of cracks, is used. In order to solve the above problems, this paper proposes a method for on-line prediction of fatigue crack based on particle filter with Gaussian weights-mixing distribution. The fatigue crack growth equation of state is established based on the NASGRO crack propagation model that characterizes the full crack propagation region. Active Lamb wave monitoring With the help of on-line structural health monitoring, the importance density function is selected as the mixed distribution of observed probability density and prior probability density, and the weight of Gauss is obtained based on the a priori estimation to update the weights. . The simulation study is carried out in this paper. The results show that the proposed method optimizes the accuracy of fatigue crack growth prediction.