基于小波变换与等价空间的无人机作动器故障检测

基于等价空间的故障检测可实现残差对初始状态与输入全解耦,在实际工程系统中取得了广泛应用.然而,为了确保故障诊断系统性能,高阶的等价空间导致残差在线计算量大,且制约了可检测故障信号的频率范围.本文提出一种基于小波变换与等价空间结合的无人机作动器故障检测方法,将故障检测问题归结为小波基函数选取和等价空间向量的优化设计,使产生的残差信号满足了干扰鲁棒性与故障灵敏性比率型性能指标的最小化,并通过引入平稳小波变换对残差进行多尺度滤波,利用小波变换的时频局部化特性和快速算法,在满足故障检测系统性能的前提下有效降低了等价空间阶数,一定程度上实现了较宽频率范围内作动器故障信号的检测,克服了传统等价空间方法的不足.最后,以无人机作动器故障检测为例,通过仿真实验验证了本文方法的有效性.

Wavelet transform and parity space based actuator fault detection for unmanned aerial vehicle

The parity space based fault detection (FD) is one of the most common approaches by using the system input and output data over a time window and attracts considerable attention in practical applications. This approach is characterized by its straight forward manner in design and the generated residual is independent of the system initial state and input. However, high parity order has to make sure for improving the FD performance, which may lead to heavy computational task and very narrow frequency range of fault to be detected. To overcome this, the paper proposes to deal with the problem of actuator FD for unmanned aerial vehicle (UAV) by applying parity space approach and wavelet transform. Then the residual can be generated based on the parity relation between the inputs and outputs such that a ratio criterion of robustness to unknown input and sensitivity to fault is minimized. Moreover, with the selection of wavelet basis and residual evaluation for a group of residuals, the faults within broader frequency range can be detected. As a result, the parity space order can be efficiently reduced for some acceptable FD performance, hence, both the computational cost of residual evaluation function and the frequency range of fault can be improved. Finally, an example of UAV longitudinal system is used to demonstrate the effectiveness of the proposed method.