摘要
为了弥补最优奇偶向量法在冗余陀螺故障诊断中存在的不足,提出了基于支持向量机的故障诊断方法。在最优奇偶向量法的基础上,将冗余陀螺测量单元中所有陀螺的奇偶残差看做整体,作为多故障支持向量分类机进行训练数据,并从训练数据预处理、核函数选择和参数寻优等方面进行了研究。将训练后的支持向量分类机用于对九陀螺冗余测量单元的故障诊断试验,试验结果表明,该方法具有良好的故障识别率、较低的漏检率和虚警率。
For the on-line fault diagnosis method of redundant gyroscopes, the optimal parity test （OPT） is widely used. By taking the limitation of the OPT into consideration, the fault diagnosis method based on support vector machine （SVM） is proposed. Firstly, the parity residuals for all gyroscopes in the nine-re- dundant-gyroscope-unit are taken as a whole to be the training data. Then the training data preprocessing, the kernel selection and the parameter optimization are researched. Finally, the trained SVM is used for the nine-redundant-gyroscope-unit fault diagnostic test. The test results show that the SVM method has high recognition rate of failure, low missing rate and false alarm rate.
出处
《航天控制》
CSCD
北大核心
2014年第5期77-83,共7页
Aerospace Control
