重复使用运载火箭发动机疲劳载荷特征识别方法

Identification Method of Fatigue Load Characteristics for Reusable Launch Vehicle Engine Based on Gaussian Distribution

针对重复使用运载火箭发动机的原始疲劳载荷数据建模困难问题,选取均方根值作为重复使用运载火箭原始疲劳载荷数据工况划分的标准,通过对重复使用运载火箭原始疲劳载荷数据进行修正短时傅里叶变换滤波处理、雨流循环计数和疲劳载荷特征量高斯分布拟合,实现疲劳载荷数据特性量的识别与规律化处理。研究表明,重复使用运载火箭疲劳载荷数据能用高斯分布模型进行概率分布描述,异常疲劳载荷数据高斯分布参数为正常疲劳载荷数据的3倍以上。该方法可用于精确识别重复使用运载火箭发动机异常疲劳载荷数据,相比于传统异常数据识别方法,可提供异常程度量化指标,为重复使用运载火箭疲劳载荷设计与实时故障分析定位提供一种新的分析手段。

Reusable launch vehicle is important to reduce the cost of launch service. This paper focuses on the modeling difficulty on the original fatigue load data of reusable launch vehicle engine. In this paper, the root mean square value is selected as the division standard for the original fatigue load data of the reusable launch vehicle. Original data are processed by modified shorttime Fourier wave filtering, rain flow cycle counting and Gaussian distribution fitting for the identification and regularization of fatigue load data. Fatigue load data of reusable launch vehicle can be described by Gaussian distribution model. The Gaussian distribution parameter of abnormal fatigue load data is more than 3 times of normal fatigue load data. This method can be used to accurately identify the abnormal fatigue load data. Compared with traditional anomaly data identification methods, this method provides a quantitative index of abnormal data, which is a new analysis method for fatigue load design and real-time fault analysis and location of reusable launch vehicle.