基于级联多模块的永磁同步电机驱动系统传导EMI建模与预测

Cascaded Multi-module Based Conducted EMI Modeling and Prediction Method for Permanent Magnet Synchronous Motor Drive System

随着半导体器件开关频率的不断提高,采用脉宽调制(pulse width modulation,PWM)技术的电机驱动系统面临着严重的传导电磁干扰问题。在实际应用中,诸如滤波器、电缆、电机及驱动器单元通常处于级联状态,同时系统级的电磁干扰(electromagnetic interference,EMI)预测无需获得系统内部每个支路/节点的EMI信息。针对上述特点,文中提出一种基于级联T参数的模块化EMI建模方法。所提方法采用基于替换定理的干扰源建模技术以考虑交流电机驱动系统中的PWM开关动作行为。同时,该方法将系统中的各个阻抗单元视为一个二端口网络“黑盒”,并采用T参数进行描述以避免Z参数不能直接应用于级联单元运算的问题。此外,对EMI标准未限制的级联器件单元进行模块化划分,并采用T矩阵运算对其进行简化,该简化过程不存在理论近似。理论分析表明,因EMI优化而增加或改变滤波器等级联模块时,基于级联T矩阵的EMI预测方法只需对涉及的级联单元重新进行划分和阻抗计算即可。最后,以一台永磁电机驱动系统为例进行分析及验证,实验对比结果表明,所提模型具有模块化、复杂度低的特点。

With the increase of the switching frequency of semiconductor devices,conducted electromagnetic interference(EMI)problems in pulse width modulation(PWM)based motor drive system become more serious.The components of this system,such as filters,cables,motors,and motor drive units,are typically cascaded in practical applications.Meanwhile,in system-level EMI modeling,it is not necessary to acquire complete EMI information for each branch/node.Taking these factors into account,this work proposed a modular EMI modeling method based on the cascaded matrix T.The interference source was established by using the substitution theorem to consider the PWM switching behavior in ac motor drive system.Besides,the impedance of each part was represented by a two-port network“black box”,and the matrix T was utilized to describe these units to avoid the problem that the matrix Z cannot be directly applied to the cascade system.Furthermore,the cascaded units can be modularly split without any theoretical approximations since they were not bound by EMI standards,and matrix T based calculation was employed to simplify it.Theoretical analysis shows that when filter units are added or changed in the EMI optimization processing,only the relevant cascaded matrix T has to be divided and recalculated in the EMI prediction.Finally,a permanent magnet synchronous motor(PMSM)drive system was employed to analyze and verify the proposal,and the experimental comparison results suggest that the model proposed has the advantages of modularity and low complexity.