Electromagnetic interference prediction technology of new energy motor drive system

With the increasingly urgent situation of environmental protection and the instability of the world situation, fossil energy, mainly oil and natural gas, is increasingly exhausted, as many regions and countries have introduced environmental protection policies to reduce their dependence on fossil energy. New energy, also known as unconventional energy, is generally the renewable energy developed and utilised based on new technologies [1–4]. With the gradual implementation of the dual carbon policy, countries around the world are increasing the scale of new energy use to reduce carbon emissions. One of the important policies is to promote the marketisation of new energy vehicles, as well as the implementation of transformation programmes such as replacing oil with electricity, which is of great significance in reducing carbon emissions and achieving a carbon peak.

New energy vehicle technology as an emerging technology, after nearly 20 years of development, has integrated vehicle power control and helps to drive the advanced technology. The formation of advanced technology principles, with the new technology, has emerged as the new structure of the car. New energy vehicles include four types of HEV, and BEV, including solar vehicles, FCEV, and other new energy vehicles [5–7]. No matter which technology direction is adopted, the driving mode of new energy vehicles is essentially different from that of traditional oil vehicles. Traditional oil vehicles use gasoline combustion in an internal combustion engine to generate power, while new energy vehicles need to use current through the motor to generate force to achieve drive, belonging to the integrated high-power electronic and electrical device system. The drive system mainly includes high-power DC/DC converter or DC/AC inverter, power battery, drive motor and high-voltage cables connected between them [8, 9]. Due to the weight of new energy vehicles and the power demand in driving, the motor will produce a complex electromagnetic environment and send out strong electromagnetic interference (EMI) during operation. In the operation process of vehicle startup and flash-off, acceleration and deceleration are mainly realised by controlling the current size and direction of the motor. However, in the actual driving process, it is inevitable to conduct frequent acceleration and deceleration operations, which will also aggravate the complexity of the electromagnetic environment generated by the motor [10–13]. On the one hand, under the frequent on-off operation of the switch, the voltage and current jump in a short time, thus forming a strong electromagnetic radiation and interference; on the other hand, the high-voltage cable driving the components may produce a large current of hundreds of amperes during operation, which also generates strong electromagnetic radiation. Thus, new energy vehicles are in the process of running the complexity of EMI of motor drive system; Regarding motor drive system for new energy vehicle core parts, its safe operation is related to life and property safety and other major problems [14–16], and therefore very necessary for new energy vehicles’ EMI generated by the driving system which is the system theory and experimental research.

In view of the EMI phenomenon, many scholars put forward methods to predict it. As shown in Figure 1, the literature on EMI prediction reports generally shows an increasing trend. Xiang et al. [17] constructed precise mathematical models of common mode and differential mode interference sources, and systematically studied interference sources of a 10 kVA three-phase bridge converter. Zhou, [18] took the widely used switching power supply as an example and proposed a new EMI conduction modelling method. The impedance analysis method was used to model high-frequency parasitic parameters, and a model for accurately predicting the conducted EMI of switching power supply was established. Duan and Wen, [19] modelled detailed circuit parameters such as fast switching transients and parasitic parameters of power modules of SiC devices. Compared with traditional methods, this method improves the prediction accuracy of conducting EMI in the high-frequency range of the SiC motor drive system. Zhang et al., [20] proposed an EMI prediction method for UAV dynamic data link based on Gaussian Process regression (GPR). The injection test was conducted on the training subjects. Wu et al., [21] analysed and modelled the DM and CM EMI noises from the AC-DC power supply in LED TV. Jin et al., [22] proposed a new method based on the deep neural network (DNN) model, which can accurately and quickly predict the maximum radiation electric field at 3 m of the weld-wire ball grid array (WB-BGA) package. The radiation predicted by the DNN model is in good agreement with that by full-wave simulation.

Fig. 1

Regarding the hot new energy electric vehicles in recent years, this paper analyses the mechanism of EMI, find the EMI sources and transmission ways, interference components modelling analysis of motor, and EMI test and motor drive system. To accurately predict the EMI of the motor drive system of new energy vehicles, we had better avoid the impact of EMI.

In the motor drive system, when the switching components switch, the EMI is caused by dv/dt, di/dt. Non-negligible oscillations occur. Therefore, it is necessary to accurately model the parasitic inductance of power modules. In compact inverters with motor drive systems, three-phase full-bridge modules containing six power switches are commonly used between the two modules or between the die and directly bonded copper and aluminium bonding lines. In addition, there is a positive pole, and a negative pole connected to the external circuit. Therefore, there is parasitic inductance in the power module. After extracting the parasitic inductance of the power circuit, the equivalent circuit is drawn, and the two are coupled with each other. The inductance can be simplified that the same phase of upper or lower branch of the parasitic inductance, namely the LPm(U) and LNm(U) or LPm(L) and LNm(L), (m1/4a,b,c), according to the type (1) given in the concept of loop inductance, can derive the simplified parasitic inductance of each phase on or under.

1 LXm=LXmU+LXmL−2LXmUYmL−LXmYm \begin{equation} {L_{Xm}} = {L_{XmU}} + {L_{XmL}} - 2{L_{XmUYmL}} - {L_{XmYm}} \label{eq1} %(1) \end{equation}

Where: 2 LXmYm=LXmUYmU+LXmLYmL−2LXmUYmL−LXmLYmU \begin{equation} {L_{XmYm}} = {L_{XmUYmU}} + {L_{XmLYmL}} - 2{L_{XmUYmL}} - {L_{XmLYmU}} \label{eq2} %(2) \end{equation}

Where X1/4P, N, Y|P, N, m1/4a, b, c, X, Y are different, the equivalent large parasitic inductor is named LPm and the small parasitic inductor is named LNm. The mutual inductance between the two phases is called Lmn (m 1/4 a, b, c, n 1/4 a, b, c, m is different from n).

Similarly, it can be obtained through Eq. (3): 3 Lmn=LXmXn+LYmYn−LXmYn−LYmXn \begin{equation} {L_{mn}} = {L_{XmXn}} + {L_{YmYn}} - {L_{XmYn}} - {L_{YmXn}} \label{eq3} %(3) \end{equation}

Where: 4 LXmXn=LXmUXnU+LXmLXnL−LXmUYnL−LXmLXnU \begin{equation} {L_{XmXn}} = {L_{XmUXnU}} + {L_{XmLXnL}} - {L_{XmUYnL}} - {L_{XmLXnU}} \label{eq4} %(4) \end{equation} 5 LXmYn=LXmU+LXmLXnU−LXmUYnL−LYmLXnL \begin{equation} {L_{XmYn}} = {L_{XmU}} + {L_{XmLXnU}} - {L_{XmUYnL}} - {L_{YmLXnL}} \label{eq5} %(5) \end{equation}

For L_YmYn and L_YmXn, equations similar to Eqs (4) and (5) can be obtained, where there is a ground stray capacitance C_p between the drain and the copper substrate of each device. To accurately predict CM EMI, the analytical calculation can be carried out through Eq. (6).

6 Cp=εrε0Spdp \begin{equation} {C_p} = \frac{{{\varepsilon _r}{\varepsilon _0}{S_p}}}{{{d_p}}} \label{eq6} %(6) \end{equation}

In the equation, S_p is the copper surface area of the insulating substrate, d_p is the thickness of the insulator, ${\varepsilon _0}$ε0 is the dielectric constant of the free space, and ${\varepsilon _r}$εr is the relative dielectric constant of the insulating material.

The inductance of the DC connection cable connecting the inverter will affect the high-frequency EMI spectrum and should be modelled. The inductance, named L_cable, can be calculated analytically by the following equation: 7 Lcableμ0l2π(ln⁡(2lr)−1) \begin{equation} {L_{cable}}\frac{{{\mu _0}l}}{{2\pi \left( \ln \left( \frac{{2l}}{r}\right) - 1\right)}} \label{eq7} %(7) \end{equation}

Where ${\mu _0}$μ0 is the permeability of free space, l and r are the length and cross-section radius of DC cable, respectively.

The total DM EMI source should be a combination of IdmV(s) and, IdmI(s) and then write: 8 Idms(s)=∑j=a,b,cvj(s)TdmVj(s)+∑j=a,b,cij(s)Tdmlj(s) \begin{equation} {I_{dms}}(s) = \sum\limits_{j = a,b,c} {{v_j}} (s){T_{dmVj}}(s) + \sum\limits_{j = a,b,c} {{i_j}} (s){T_{dmlj}}(s) \label{eq8} %(8) \end{equation}

${T_{dmVj}}(s)$TdmVj(s) is the transfer function from ${v_j}(s)$vj(s) to ${I_{dmV}}(s)$Idmv(s), ${T_{dmlj}}(s)$Tdmlj(s) is the transfer function from ${i_j}(s)$ij(s) to ${I_{dml}}(s)$Idml(s).

The total CM EMI source VCM should be a combination of VcmV and VcmI, which can also be written as: 9 Vdms(s)=∑j=a,b,cvj(s)TcmVj(s)+∑j=a,b,cij(s)Tcmlj(s) \begin{equation} {V_{dms}}(s) = \sum\limits_{j = a,b,c} {{v_j}(s){T_{cmVj}}(s)} + \sum\limits_{j = a,b,c} {{i_j}(s){T_{cmlj}}(s)} \label{eq9} %(9) \end{equation}

${T_{cmVj}}(s)$TcmVj(s) is the transfer function from ${v_j}(s)$vj(s) to ${I_{cmV}}(s)$IcmV(s), ${T_{cmlj}}(s)$Tcmlj(s) is the transfer function from ${i_j}(s)$ij(s) to ${I_{cml}}(s)$Icml(s).

It can be seen that in Eqs (8) and (9), to analytically calculate IDM(s) and VDM(s), the frequency domain expressions of ${v_j}(s)$vj(s) and ${i_j}(s)$ij(s) are the first.

The EMI test of the motor system mainly includes two aspects: one is to test the conducted radiation; second, PCB board near field radiation is tested. The conducted radiation includes the voltage method and the test of instantaneous pulses in the main loop route, usually using CISPR15 and GB18655-2011.

The schematic diagram of the voltage method test is shown in Figure 6. According to the international standard manual, the positive and negative power cables of the test equipment should be connected to two manual networks respectively. The length of the power cables should exceed 250 mm.

Fig. 6

Figure 7 shows the schematic diagram of the testing structure of the instantaneous pulse of the main loop route. This method has a more accurate testing effect than the voltage method. In this circuit, the connecting wire should be arranged in an insulation material of 70 mm or so. The oscilloscope and voltage wave are added to the circuit, and the waveform diagram of repeated changes can be obtained by controlling the circuit through switches. To keep the accuracy of the results, it waveform data needs to be collected10 times and the best quality is selected. A schematic diagram of the circuit structure is shown in Figure 7.

Fig. 7

Fig. 8

Taking the motor drive system as the research object, the EMI results are tested. In this paper, the model XC2287 three-phase AC permanent magnet synchronous motor is taken as the research object, and the main parameters are as follows: The voltage of the DC power supply is 26 V, the frequency of the PWM wave is 900 kHz, the rated power is 3600 W, the rated current is 7500 mA, the maximum instantaneous current is 1370 mA, the line-phase electromotive force is 5.4 V, the speed is fixed at 3200 RPM, and the maximum is 4000 RPM. The internal resistance of the coil is 0.264 ω, and the inductance of the coil is 0.934 mH. The conducted emission test adopts instantaneous pulse test mode, and the schematic diagram of PCB near field radiation is shown in Figure 7. An electromagnetic wave is converted into an analogue signal after receiving it by the coil in space, and a digital signal waveform is acquired by the digital oscilloscope. The frequency domain signal emitted by electromagnetic radiation is converted into a time domain signal, which is obtained by Fourier transform. According to the signal, B (field intensity) and F (frequency) curves are drawn.

In view of the complex and changeable EMI in the motor drive system of new energy vehicles, this paper analysed the generation mechanism of EMI, found the EMI source, and modelled and analszed the motor interference components. The conclusions are as follows: