一种Buck变换器输出电容故障在线预测方法

提出一种基于特征参数退化的Buck变换器输出电容故障在线预测方法。首先,分析输出电容的失效机理,选取反映电容退化规律的等效串联电阻(ESR)RESR和电容值(C)为退化特征参数。然后,根据Buck连续导通模式,提出一种新的利用输出电压纹波分量计算特征参数ESR值RESR和C的在线提取方法,该方法可高效无电流传感器实现ESR值RESR和电容C在线监测。最后,利用特征参数ESR值RESR和C的时间序列,采用最小二乘支持向量机、BP神经网络和灰色模型实现Buck变换器输出电容故障值预测。仿真实例及试验验证该方法的有效性和准确性。

AN ONLINE METHOD FOR FAILURE PREDICTION OF OUTPUT CAPACITOR IN BUCK CONVERTER

A online fault prediction method based on degradation characteristic parameters for output capacitor in Buck converter is proposed. Firstly,the failure mechanisms of output capacitor are analyzed,and the Equivalent series resistance(ESR) RESR and the capacitance(C)are selected as characteristic parameters to reflect the degradation of output capacitor. Then,a new on-line extracting method is presented for calculating RESR and C by using ripple component of output voltage in Buck converter with continuous conduction mode,which can monitoring RESR and C online with high efficiency and no current sensor. Finally,LS-SVM,BP neural network and gray model are applied to realize capacitor’s failure prediction according to the time series of RESR and C. The simulation and example prediction results validate the effectiveness and the accuracy of the proposed method.