摘要
针对目前支持向量机参数选择时人为选择的盲目性,将具有良好优化性能的蚁群优化技术应用到支持向量机惩罚函数和核函数参数的优化,提出了蚁群优化支持向量机方法。根据内燃机气门振动信号实测数据,建立了基于蚁群优化支持向量机的内燃机气门间隙故障诊断模型,并与基于遗传支持向量机和反向传播神经网络算法的模型比较。结果表明:应用蚁群优化支持向量机建立的内燃机气门间隙故障诊断模型无论从学习效率还是故障识别准确性上都优于应用另外两种算法建立的模型,能够有效地进行内燃机的故障诊断。
Due to blindness of man-made choice for parameters of a support vector machine(SVM),an ant colony optimization was used to select parameters of SVM;and a novel algorithm 'ant colony optimization support vector machine'(ACO-SVM) was put forward.According to the measured data of the vibration signal of an engine valve,an engine valve fault diagnosis model based on ACO-SVM was established,and compared with the models based on GA-SVM and BPNN.Results showed that the engine valve fault diagnosis model based on ACO-SVM outperforms the models based on the other two algorithms in learning efficiency and diagnosis accuracy,and is effective for engine fault diagnosis.
出处
《振动与冲击》
EI
CSCD
北大核心
2009年第3期83-86,共4页
Journal of Vibration and Shock
基金
863计划资助(编号:2006AA04Z408)
关键词
蚁群算法
支持向量机
BP神经网络
故障诊断
ant colony optimization(ACO)
support vector machine(SVM)
BP neural networks(BPNN)
fault diagnosis
genetic algorithm(GA)
