火箭发动机故障检测的快速增量单分类支持向量机算法
2024,46(2):115-122
张万旋
北京航天动力研究所, 北京 100076，wanxuanz@163.com
张箭
北京航天动力研究所, 北京 100076
卢哲
北京航天动力研究所, 北京 100076
薛薇
北京航天动力研究所, 北京 100076
张楠
北京航天动力研究所, 北京 100076
北京航天动力研究所, 北京 100076，wanxuanz@163.com
张箭
北京航天动力研究所, 北京 100076
卢哲
北京航天动力研究所, 北京 100076
薛薇
北京航天动力研究所, 北京 100076
张楠
北京航天动力研究所, 北京 100076
摘要：
为解决液体火箭发动机故障诊断正负样本不平均问题,以及实现发动机稳态工作段自适应故障检测,建立了基于快速增量单分类支持向量机的异常检测模型。采取特征工程方法,对传感器获得的多变量时间序列进行特征提取。通过增量学习方法,对单分类支持向量机模型进行改进,并应用于液体火箭发动机异常检测,使单分类支持向量机检测模型具备对不同台次、不同工况的自适应性,提高了模型的计算速度。对多台次热试车数据的分析结果表明,该模型十分有效,训练速度快,具备实用价值。
为解决液体火箭发动机故障诊断正负样本不平均问题,以及实现发动机稳态工作段自适应故障检测,建立了基于快速增量单分类支持向量机的异常检测模型。采取特征工程方法,对传感器获得的多变量时间序列进行特征提取。通过增量学习方法,对单分类支持向量机模型进行改进,并应用于液体火箭发动机异常检测,使单分类支持向量机检测模型具备对不同台次、不同工况的自适应性,提高了模型的计算速度。对多台次热试车数据的分析结果表明,该模型十分有效,训练速度快,具备实用价值。
基金项目：
国家自然科学基金资助项目(52232014)
国家自然科学基金资助项目(52232014)
Fast incremental one-class support vector machine algorithm for rocket engine fault detection
ZHANG Wanxuan
Beijing Aerospace Propulsion Institute, Beijing 100076, China,wanxuanz@163.com
ZHANG Jian
Beijing Aerospace Propulsion Institute, Beijing 100076, China
LU Zhe
Beijing Aerospace Propulsion Institute, Beijing 100076, China
XUE Wei
Beijing Aerospace Propulsion Institute, Beijing 100076, China
张楠
Beijing Aerospace Propulsion Institute, Beijing 100076, China
Beijing Aerospace Propulsion Institute, Beijing 100076, China,wanxuanz@163.com
ZHANG Jian
Beijing Aerospace Propulsion Institute, Beijing 100076, China
LU Zhe
Beijing Aerospace Propulsion Institute, Beijing 100076, China
XUE Wei
Beijing Aerospace Propulsion Institute, Beijing 100076, China
张楠
Beijing Aerospace Propulsion Institute, Beijing 100076, China
Abstract：
In order to solve the problem of imbalance between positive and negative samples in liquid rocket engine fault diagnosis, and to enable adaptive fault detection during engine steady working state, a anomaly detection model based on fast incremental one-class support vector machine was established. Feature engineering method was adopted to extract features from sensor-obtained multivariate time series. Through incremental leaning, the one-class support vector machine model was improved and applied to liquid rocket engine anomaly detection. The one-class support vector machine detection model was endowed with adaptability for various engine individuals and multiple working conditions, while increasing computing speed. The analysis results of multiple hot test data show that the model is effective, fast-training and practically valuable.
In order to solve the problem of imbalance between positive and negative samples in liquid rocket engine fault diagnosis, and to enable adaptive fault detection during engine steady working state, a anomaly detection model based on fast incremental one-class support vector machine was established. Feature engineering method was adopted to extract features from sensor-obtained multivariate time series. Through incremental leaning, the one-class support vector machine model was improved and applied to liquid rocket engine anomaly detection. The one-class support vector machine detection model was endowed with adaptability for various engine individuals and multiple working conditions, while increasing computing speed. The analysis results of multiple hot test data show that the model is effective, fast-training and practically valuable.
收稿日期：
2021-11-20
2021-11-20
