复杂噪声环境下运载火箭时变模态在线辨识方法研究

由于无法准确模拟飞行状态下运载火箭的真实工作环境，通过有限元分析或地面振动试验获得的模态参数往往与飞行状态下的真实值存在一定差异，天地一致性难以保证。然而，飞行状态下运载火箭的动力学特性会随着燃料的快速消耗而不断变化，所面临的噪声环境也极为复杂，现有辨识方法尚不具备对其时变模态进行高效鲁棒在线辨识的能力。面向运载火箭时变模态在线辨识的巨大需求，本项目针对运载火箭振动响应实测数据信噪比很低、常常包含由测量/传输故障引起的过失误差成分和由激励非白噪声引起的谐波成分的实际问题开展研究，旨在赋予现有时变模态仅输出递推辨识方法处理复杂噪声环境的能力，突破现阶段真实工作状态下运载火箭时变模态在线辨识的关键技术障碍，为满足我国航空航天领域日益增长的在线辨识需求提供理论与方法支持。

Due to the fact that the real operational environment of launch vehicles cannot be accurately simulated, the modal parameters obtained by finite element analysis or ground vibration test are not always consistent with the true values of launch vehicles in flight. Besides, the dynamic characteristics of launch vehicles are changing with the rapid fuel consumption, and the noise environment faced by them is also extremely complex. The existing identification methods are not yet efficient and robust to identify their time-varying modal parameters online. Based on the huge demand for online identification of time-varying modal parameters, this project focuses on three key issues concerning the response measurements of launch vehicles: lower signal-to-noise ratio, gross errors caused by the measurement/transmission failures and harmonic components caused by the non-white excitation. By making the existing output-only recursive time-varying modal identification methods to operate in complex noise environment, this project aims to solve the current bottleneck problem of the online identification of time-varying modal parameters for in-flight launch vehicles, and provide theoretical and methodological support to meet the increasing demand of online identification in China’s aerospace engineering applications.