改进的离散小波-优化极限学习机在倾转旋翼机故障诊断中的应用

针对倾转旋翼机飞控系统的故障诊断问题，提出一种改进的离散小波-优化极限学习机（OMELM）的故障诊断算法。提出自适应启发式小波去噪方法对采集的信号进行消噪，定义了帕塞瓦尔能量用来提取测量信号经离散小波变换分解后的特征，并对OMELM进行了改进。将提取的故障能量特征进行归一化后输入到改进的OMELM多分类器中进行分类，以美国XV-15倾转旋翼机为例进行仿真验证。结果表明文中方法平均辨识率高，诊断时间短，对未来我国进行倾转旋翼机故障诊断的研究有一定参考价值。

Fault Diagnosis of Tiltrotor Aircraft via Improved Discrete Wavelet-OMELM

An improved discrete wavelet-optimization method-based extreme learning machine (OMELM) algorithm is presented for the fault diagnosis of flight control system in tiltrotor aircraft. An adaptive heuristic wavelet denoising method is used to denoise the sampled signal. Feature vector of each layer is extracted using Parseval energy after the discrete wavelet decomposition of fault signal. The energy feature is normalized as the improved OMELM network input, and then the actuator fault models is classified using the improved OMELM network. Finally, an XV-15 tiltrotor aircraft mode is validated by simulation. The results show that the method has a higher average recognition rate, and needs a short diagnosis time.