摘要
针对数/模混合电路的故障诊断,将遗传算法与最小二乘支持向量机相结合,提出一种二值分类故障诊断方法,在保证故障诊断的准确率和可靠性的基础上,实现多类故障的快速诊断。采用遗传算法优化支持向量机的结构参数和核函数参数,建立基于支持向量机的故障分类模型;提出适用于数/模混合电路的故障诊断的二值分类策略;在Pspice环境下进行仿真验证,实验结果表明,该方法有效提高了的故障诊断的精度和效率,在类别较多故障模式中具有明显的优势和实用价值。
Arithmetic of mixed circuit fault diagnoses based on LS-SVM and GA is proposed, and verifies it is verified by simula- tion in Pspice. The simulation result reflects that, the SVM Binary classifier based on GA and binary classification strategy, in which structural parameters and kernel function parameters optimized by GA, has more obvious advantages and practical value in the case of more fault mode category , and it is quite suitable for gaining high accuracy and efficiency of diagnosis.
出处
《计算机工程与设计》
CSCD
北大核心
2012年第10期3984-3989,共6页
Computer Engineering and Design
