基于重构相空间短时状态预测残差的直升机旋翼损伤跟踪方法

运用相空间重构技术与Volterra泛函级数提出基于短时状态预测残差的旋翼损伤跟踪方法。应用相空间重构方法将旋翼状态监测信号变换到不同维数的相空间内以不同时间尺度观测系统的动态行为特性,从新视角研究旋翼损伤的演化过程,据旋翼实测相轨迹与Volterra基准状态模型预测结果之间的残差生成新的损伤观测信号,采用Unscent-ed滤波器对Volterra基准模型预测残差进行滤波,可进一步优化估计旋翼的损伤状态。该方法可利用单个状态监测信号的非线性特性在相空间中重建旋翼系统动力学本质,可更有效地开展旋翼的损伤跟踪与使用监测。

Method to monitor helicopter rotor health based on short time state prediction error in reconstructed phase space

A new method to monitor helicopter rotor health in flight was developed based on short time state prediction error generated by the Volterra series applicable in the reconstructed phase space .Using the embedding theory, the evolving property of damaged rotor behavior can be seen and understood at appropriate time scales from particular state space. A new damage tracking metric was derived from the difference between measured trajectories and the reference trajectories predicted by the Volterra series. The Unscented filter was applied to filter the measurements and estimate the rotor damage state. The proposed method can reconstruct the underlying system dynamic nature in the phase space using the nonlinear property of simple rotor vibration signal time series, which holds much promise for the rotor health and usage monitoring system.