Two-level PWM converter switching waveform determination method for electromagnetic interference prediction

Abstract

The invention discloses a two-level PWM converter switching waveform determination method for electromagnetic interference prediction, and belongs to the field of power electronic electromagnetic compatibility. According to the method, based on steady-state quantity and transient quantity solution of a state equation set, a switching waveform containing a nonlinear transient state can be efficiently obtained, and then through conversion and detection processing of an equivalent interference source, the switching waveform containing the nonlinear transient state can be obtained; a broadband high-precision switch waveform frequency spectrum for electromagnetic interference prediction can be obtained; benefited from the specified initial value of the state variable, the nonlinear device model, the simplified calculation of the steady-state value of the state variable and the parallel calculation of the transient value, the method has the advantages of good convergence, high precision and high calculation efficiency at the same time, and realizes the calculation of the switching waveform spectrum for the electromagnetic interference prediction from the physical parameters of the switching device; the spectral range covers 0-1GHz, and the calculation time is only within hundred seconds; therefore, the electromagnetic interference modeling speed can be increased, the prediction precision can be improved, and finally the electromagnetic interference suppression design of the two-level PWM converter can be accurately guided.

Description

Technical Field

[0001] The invention belongs to the field of electromagnetic compatibility of power electronics, and more specifically, relates to a method for determining a switching waveform of a two-level PWM converter for electromagnetic interference prediction. Background Art

[0002] In recent years, two-level PWM converters have been widely used in modern power systems, rail transit and consumer electronics. However, the inherent switching mode of two-level PWM converters will cause serious electromagnetic interference problems, endangering the normal operation of surrounding electronic equipment or affecting the reliability and life of electronic systems. Before deployment, two-level PWM converters must pass strict EMC certification. Good electromagnetic interference suppression design is the only way to succeed in EMC certification. Compared with the suppression design based on prototype tests, the suppression design based on models consumes less manpower and material resources, is more efficient, has more feasible solutions, and can ultimately help improve product performance. Therefore, electromagnetic interference modeling of two-level PWM converters is currently widely concerned and applied.

[0003] The electromagnetic interference modeling of the two-level PWM converter is based on the physical parameters of active and passive devices, and uses field / path simulation or code programming to predict the electromagnetic interference spectrum measured at the LISN or antenna when the two-level PWM converter prototype to be developed is running. Electromagnetic interference modeling is mainly divided into the modeling of interference sources and interference paths. Compared with the mature interference path modeling, the modeling of interference sources is still understudied. The modeling of interference sources is to obtain the spectrum amplitude of the switching waveform. The switching waveform modeling used for electromagnetic interference prediction, that is, interference source modeling, is to establish a function model between the physical parameters of the switching device and related auxiliary circuits and the spectrum amplitude of the switching waveform when the switching device is working.

[0004] Switching waveform modeling is the top priority of electromagnetic interference modeling for two-level PWM converters. It not only affects the design of the converter port EMI filter, but is also closely related to the design of the drive and buffer circuits, which significantly affects the efficiency of the converter. Switching waveform modeling for electromagnetic interference prediction is mainly divided into three methods: time domain, frequency domain, and time-frequency domain modeling: the time domain modeling method based on nonlinear switching waveform has slightly higher accuracy, but the simulation time is extremely long and the simulation convergence is poor; the frequency domain modeling method based on ideal linear switching waveform has a short simulation time but low accuracy; there is also a type of time-frequency domain modeling method that combines the advantages of both, but the method is complex to operate, has poor reproducibility, fails to consider the nonlinear effects of switching devices, and the accuracy is still low. In addition, the applicable frequency bands of these methods are very low, and the maximum electromagnetic interference prediction of about 10MHz can be achieved. However, the required frequency range for electromagnetic interference modeling of two-level PWM converters is 0Hz-1GHz, and no model has yet been able to achieve such a wideband electromagnetic interference prediction. In summary, the current switching waveform models used for electromagnetic interference prediction still face the problems of low accuracy and low efficiency. Summary of the invention

[0005] In view of the above defects or improvement needs of the prior art, the present invention provides a method for determining a two-level PWM converter switching waveform for electromagnetic interference prediction, thereby solving the problems of low accuracy and low efficiency of the existing two-level PWM converter switching waveform modeling methods.

[0006] To achieve the above object, according to a first aspect of the present invention, a method for determining a switching waveform of a two-level PWM converter for electromagnetic interference prediction is provided, comprising:

[0007] S1, compare the modulated wave of the converter with the carrier wave to obtain the steady-state value V of the gate voltage of the first and second switching tubes in the converter gs1_sta 、V gs2_sta , and according to V gs1_sta 、V gs2_sta and load current I L Determine the steady-state value V of the voltage across the first and second switch tubes ds1_sta 、V ds2_sta and the steady-state value of the output current I s1_sta ,I s2_sta ;

[0008] Among them, at a certain moment, if the modulation wave amplitude is greater than the carrier amplitude, then at this moment, V gs1_sta =V H , V gs2_sta =V L , I s1_sta =-I L , I s2_sta =0, when I L ≤0, V ds1_sta=V moson 、V ds2_sta =V DC -V moson , when I L >0, V ds1_sta =-V Don 、V ds2_sta =V DC +V Don ; If the modulation wave amplitude is less than or equal to the carrier amplitude, then at this moment, V gs1_sta =V L , V gs2_sta =V H , I s1_sta =0,I s2_sta =I L , when I L When V ds1_sta =V DC +V Don 、V ds2_sta =-V Don , when I L When negative, V ds1_sta =V DC -V moson 、V ds2_sta =V moson ; V H 、V L They are high and low level driving voltage, V DC 、V Don 、V moson They are input voltage, diode conduction voltage, and MOS tube conduction voltage respectively.

[0009] S2, taking the voltage across both ends of each switch tube, the gate voltage and the output current as its state variables, and determining the transient initial value and driving voltage of the state variables of each switch tube;

[0010] Among them, the transient initial values ​​of the voltage across the ends of each switch tube and the output current are the steady-state values ​​of the voltage across the ends and the output current at the transient start time n respectively; if I L is positive, then the transient initial values ​​of the gate voltage and drive voltage of the first switch tube are both V L , the transient initial values ​​of the gate voltage and the driving voltage of the second switch tube are the steady-state values ​​of the gate voltage at time n and time n+1 respectively; if I L If V is negative, the initial transient values ​​of the gate voltage and drive voltage of the first switch tube are the steady-state values ​​of the gate voltage at time n and time n+1 respectively, and the initial transient values ​​of the gate voltage and drive voltage of the second switch tube are both V L ;

[0011] S3, establishing a differential equation group of the state variables of each switch tube with respect to the transient moment according to the switch tube model, the diode model and the voltage equation of the junction capacitance, and solving the differential equation group in combination with the transient initial value and the driving voltage of the state variables of each switch tube to obtain the transient value of each state variable;

[0012] S4, the steady-state value and transient value of each state variable of each switch tube are spliced ​​to obtain the complete time domain waveform of each state variable, so as to calculate the common-mode and differential-mode interference source switching waveforms of the converter, and the detection processing is performed to obtain the two-level PWM converter switching waveform for electromagnetic interference prediction.

[0013] According to a second aspect of the present invention, there is provided an electronic device, comprising: a computer-readable storage medium and a processor;

[0014] The computer-readable storage medium is used to store executable instructions;

[0015] The processor is used to read the executable instructions stored in the computer-readable storage medium and execute the method as described in the first aspect.

[0016] According to a third aspect of the present invention, a computer-readable storage medium is provided, wherein the computer-readable storage medium stores computer instructions, and the computer instructions are used to cause a processor to execute the method as described in the first aspect.

[0017] According to a fourth aspect of the present invention, there is provided a computer program product, comprising a computer program or instructions, wherein when the computer program or instructions are executed by a processor, the method according to the first aspect is implemented.

[0018] In general, the above technical solutions conceived by the present invention can achieve the following beneficial effects compared with the prior art:

[0019] 1. The method provided by the present invention calculates the initial value of the state variable of the switch tube in a specified manner rather than iteratively step by step, thereby ensuring that the I / V characteristics of the nonlinear device (switching devices such as IGBT or MOS and diodes) are in the correct conduction or shutdown interval, thereby ensuring the computational convergence of the switch waveform model, and will not produce the computational errors of the existing time domain modeling method that does not converge; the method provided by the present invention can realize the switch waveform calculation with both computing power selectivity and support for parallel processing, wherein the state equation group is not solved in the steady state of non-switch switching, and the state equation group is only solved in the switch switching transient state, focusing the limited computing power on the switch transient state that requires high precision, thereby improving the computational efficiency; and the transient calculations of different switch switching times are independent of each other, and parallel calculations can be used to further improve the computational efficiency; in addition, the method provided by the present invention, when establishing the differential equation group of the state variable, takes into account the pressure change characteristics of the device junction capacitance, which can further improve the computational accuracy. After simulation verification, the computational efficiency (minute level) and convergence of this method far exceed the existing foreign commercial software (hour level) such as Pspice.

[0020] 2. Furthermore, the method provided by the present invention takes into account that the existing transconductance model of the switching device only considers the device drain current I d With gate voltage V gs The change of drain-source voltage V ds The influence can only be on the drain-source voltage V ds The larger amplification area is used, which is not applicable in the device conduction stage or the initial stage of shutdown. The transconductance model is improved, and the improved model also takes into account V ds and V gs For drain current I d It is applicable to all stages of the operation of switching devices with higher accuracy and greater versatility.

[0021] 3. The method provided by the present invention adopts the Runge-Kutta algorithm to solve the state equation group to obtain the transient value of the state variable, which has higher accuracy and is not easy to diverge.

[0022] 4. The method provided by the present invention takes into account the skin effect of ringing when solving the state equation group. When the switching waveform enters the oscillation stage, the skin resistance is made effective, which can further improve the calculation accuracy. BRIEF DESCRIPTION OF THE DRAWINGS

[0023] Figure 1 A flow chart of a method for determining a switching waveform of a two-level PWM converter for electromagnetic interference prediction provided by an embodiment of the present invention;

[0024] Figure 2 (a) and (b) are schematic diagrams of function curves when the switch device model adopts a transconductance model and an improved model provided by an embodiment of the present invention, respectively;

[0025] Figure 3 (a) and (b) are Figure 2 The schematic diagram of the input, output and reverse transfer capacitance voltage variation curves provided in the data sheet of the voltage variation junction capacitance model of the IGBT or MOS device, and the schematic diagram of the converted junction capacitance voltage variation curve;

[0026] Figure 4 (a) and (b) are schematic diagrams of a diode model using equivalent junction capacitance to characterize reverse recovery and a schematic diagram of a diode forward characteristic curve, respectively;

[0027] Figure 5 (a) and (b) are Figure 4 Schematic diagram of the switching device and the combined model of the anti-parallel diode junction capacitance;

[0028] Figure 6 It is a schematic diagram of the switching transient circuit model of a two-level PWM converter;

[0029] Figure 7 (a) and (b) are typical topology diagrams of single-phase two-level PWM converter and three-phase two-level PWM converter respectively.

[0030] Figure 8 The state variable V provided by the embodiment of the present invention ds Waveform diagram;

[0031] Fig. 9 The state variable I provided in the embodiment of the present invention s Waveform diagram;

[0032] Fig.10 A schematic diagram of a spectrum of a conducted common-mode interference source provided by an embodiment of the present invention;

[0033] Fig.11 A schematic diagram of a spectrum of a conducted differential mode interference source provided by an embodiment of the present invention;

[0034] Fig.12 A schematic diagram of the source frequency of radiated common-mode interference provided by an embodiment of the present invention;

[0035] Fig.13 A schematic diagram of a spectrum of a radiated differential mode interference source is provided for an embodiment of the present invention. DETAILED DESCRIPTION

[0036] In order to make the purpose, technical solutions and advantages of the present invention more clearly understood, the present invention is further described in detail below in conjunction with the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are only used to explain the present invention and are not intended to limit the present invention. In addition, the technical features involved in the various embodiments of the present invention described below can be combined with each other as long as they do not conflict with each other.

[0037] Due to the inherent switching operation mode of the two-level PWM converter, it will generate strong electromagnetic interference during operation. Electromagnetic interference that is not well suppressed will seriously pollute the surrounding electromagnetic environment and power supply system through the coupling mode of conduction and radiation. The electromagnetic interference suppression design of the two-level PWM converter is extremely dependent on an accurate switching waveform model. At present, the frequency domain model based on the ideal switching waveform lacks consideration of the nonlinear characteristics and ringing effect of the device, and has low accuracy. The existing time domain model based on the nonlinear switching waveform has low computational convergence and low efficiency, especially for the radiated electromagnetic interference switching waveform that requires an extremely high prediction frequency, and even the correct result cannot be calculated. Therefore, the existing switching waveform determination method for electromagnetic interference prediction faces the pain points of low accuracy and low computational efficiency.

[0038] An embodiment of the present invention provides a method for determining a switching waveform of a two-level PWM converter for electromagnetic interference prediction, comprising:

[0039] S1, compare the modulated wave of the converter with the carrier wave to obtain the steady-state value V of the gate voltage of the first and second switching tubes in the converter gs1_sta 、V gs2_sta , and according to V gs1_sta 、V gs2_sta and load current I L Determine the steady-state value V of the voltage across the first and second switch tubes ds1_sta 、V ds2_sta and the steady-state value of the output current I s1_sta ,I s2_sta ;

[0040] At a certain moment, if the modulation wave amplitude is greater than the carrier amplitude, then at that moment, V gs1_sta =V H , V gs2_sta =V L , I s1_sta =-I L , I s2_sta =0, when I L ≤0, V ds1_sta =V moson 、V ds2_sta =V DC -V moson , when I L >0, V ds1_sta =-V Don 、V ds2_sta =V DC +V Don ; If the modulation wave amplitude is less than or equal to the carrier amplitude, then at this moment, V gs1_sta =V L , V gs2_sta=V H , I s1_sta =0,I s2_sta =I L , when I L When V ds1_sta =V DC +V Don 、V ds2_sta =-V Don , when I L When negative, V ds1_sta =V DC -V moson 、V ds2_sta =V moson ; V H 、V L They are high and low level driving voltage, V DC 、V Don 、V moson They are input voltage, diode voltage, and MOS tube conduction voltage respectively.

[0041] Specifically, in step S1, a steady-state value of a state variable is generated by coding according to a modulation strategy.

[0042] In the present invention, the state variable refers to the voltage V across each switch tube. ds , gate voltage V gs , the switch output current I s ; Among them, V gs For calculating V ds and I s The intermediate variable of the two-level PWM converter; Since the two-level PWM converter has two switching tubes (MOS tube or IGBT), the switching waveform of the two-level PWM converter refers to the V ds and I s waveform.

[0043] Since the conduction voltage V of the diode and MOS tube is calculated Don and V moson The diode model and the switch device model are needed when the differential equation group is established in step S3, and the diode model, the switch device model and the junction capacitance voltage change model of the switch tube are needed when the differential equation group is established in step S3. Therefore, the above models are described below. It is worth noting that the above models can all adopt the existing models (as shown in equations (1), (3) and (5)), wherein, preferably, the switch device model can adopt the improved model of the present invention (as shown in equation (2)).

[0044] Extract physical parameters from the data sheet of the switching device. Extract the switch tube model based on the output and transfer characteristic curves. d =g(V ds ,V gs), extract the diode model I according to the diode forward characteristic D =f(V D ). According to the curve of junction capacitance and voltage across the switch tube, the voltage-dependent characteristic curve of junction capacitance C = p(V ds ). And according to the diode model and the switch tube model, the steady-state voltage and current of each switch tube when there is no transient switching process and it is only in the state of conduction or shutdown are calculated.

[0045] 1. Output / transfer characteristic model of IGBT or MOS device (i.e. switching device model)

[0046] The transconductance model widely used in the existing switching waveform modeling method is shown in formula (1). The present invention takes into account that the transconductance model (1) only considers the device drain current I d With gate voltage V gs The change of drain-source voltage V ds The influence can only be on the drain-source voltage V ds The larger amplification area is used, and it is not applicable in the device conduction stage or the initial stage of shutdown. Therefore, preferably, the switch device model used in the present invention is an improved model, which takes into account V ds and V gs For drain current I d The model is applicable to all stages of the operation of switching devices, has higher accuracy and versatility, and improves the model accuracy of the device in the saturation region.

[0047] The improved model is shown in formula (2). In these two expressions, V gsth , V dsth , K is a parameter that needs to be fitted according to the device data sheet. Figure 2 (a) and (b) compare the two model function curves and point out the key parameter V of the model. gsth , V dsth The remaining parameter K needs to be calculated by reading a point in the data sheet.

[0048] I mos =g(V gs -V gsth ) (1)

[0049]

[0050] Among them, I mos is the channel current of the switch tube, V gs 、V ds They are the gate voltage and the voltage across the switch tube, V gsth 、V dsth , K are fitting parameters, V this the gate threshold voltage, ∨ and ∧ are the calculation symbols of or and respectively.

[0051] 2. Junction capacitance voltage change model of IGBT or MOS device

[0052] The junction capacitance of the device is closely related to the rising and falling slopes of the switching transient. The voltage-dependent effect of the junction capacitance leads to the variable slope of the switching transient. The voltage-dependent effect is characterized by equation (3). gd0 and m gd Need to be extracted from the data sheet. Figure 3 (a) in the figure plots the measured capacitance versus drain-source voltage curve provided in the data sheet. Figure 3 (b) in the figure plots the Figure 3 The junction capacitance of the device after the conversion of (a) data. The intersection of the curve and the y-axis is C gd0 and C ds0 In addition, a data point needs to be read to complete the modeling of the voltage-variation effect. Since the present invention is applicable to both IGBT and MOS, and the modeling method is the same, all the following discussions of the present invention will only introduce MOS tubes.

[0053]

[0054] 3. Device model of anti-parallel diode

[0055] The diode in the two-level PWM converter is an anti-parallel diode. The forward conduction model of the diode is selected from equation (4) and Figure 4 The broken line model shown in (a) in Figure 1 is shown in Figure 1. It should also be noted that the diode also has junction capacitance, which is the direct cause of the reverse recovery characteristics. The modeling of the junction capacitance can be based on the recovery charge Q provided in the data sheet. R and the reverse voltage V when measuring R Determine, as shown in formula (5). Due to the junction capacitance C of the diode D and MOS tube junction capacitance C ds Since they are in parallel, it is not necessary to calculate the diode junction capacitance C separately. D The current is combined to calculate C D +C ds The current can be Figure 5 As shown in (a) and (b) in .

[0056] This capacitance is hereinafter referred to as C' ds

[0057]

[0058] C d ' s =C D +C ds (6)

[0059] In step S1, for the voltage V across the switch tube ds , the encoding generates the known modulation wave and carrier sequence, and the two sequences are compared to obtain V ds The ideal square wave of the square wave sequence is generated in parallel, which is much more efficient than the traditional time domain step simulation. ds The amplitude of the square wave depends on whether the MOS tube is in forward conduction or the diode is in freewheeling. gs The method of obtaining V ds The acquisition method is similar to that of the known modulation wave and carrier sequence, which are compared and driven according to the high level V H and drive low level V L , we get V gs For the switch output current I s The output current waveform amplitude of the switch tube is determined according to the current on or off state of the switch tube and the positive direction of the device current. This waveform is also generated in parallel with high calculation efficiency.

[0060] The ideal switching waveform is generated by coding rather than circuit simulation, which makes it easy to modify the waveform parameters. The pseudo code of the coding algorithm is as follows.

[0061] Algorithm: Calculation of steady-state values ​​of state variables.

[0062] Input: Input voltage V DC , output voltage V o , load current effective value I Load , power factor PF, switching frequency F s , modulation wave frequency F u .

[0063] Output: The steady-state matrix V of the state variables of the two switching tubes ds_sta =[V ds1_sta ,V ds2_sta ],V gs_sta =[V gs1_sta ,V gs2_sta ], I s_sta =[I s1_sta ,I s2_sta ],V ds_sta ,V gs_sta ,I s_sta Each element is a row with 1 and column with f s / F u Array of .

[0064] 1. Initialize the waveform duration t max 1 / F u , sampling frequency f sis 2GHz, and the calculation time vector t=[1 / f s :1 / f s :t max ], the length is N. According to the input voltage V DC Output voltage V o Calculate the modulation ratio m = 1.414 × V o / V DC .

[0065] 2. According to the power factor PF and modulation wave frequency F u and load current effective value I Load Generate load sinusoidal current time domain waveform I L Vector, the length of the vector is N.

[0066] 3. Generate a switching frequency of F s Triangular carrier Carrier, with amplitude [0,1]. Carrier is a vector of N.

[0067] 4. The single-phase two-level PWM converter generates a phase with an amplitude of m and a frequency of F according to the sinusoidal modulation (SPWM) strategy. u The sinusoidal modulation wave D, that is, D = sin(2×F u t). If it is a three-phase two-level PWM converter, the space vector modulation (SVPWM) measurement generates a phase amplitude of m and a frequency of F. u The SVPWM sinusoidal modulation wave D ra , the formula is:

[0068]

[0069] 5. Compare the amplitudes of the modulation wave D and the carrier at each moment one by one, and combine the high level of the driving voltage V H and low level V L Get V gs =[V gs1 ,V gs2 ]. That is, when D is greater than Carrier, V gs1 =V H , V gs2 =V L , when D is less than or equal to Carrier, V gs1_sta =V L , V gs2_sta =V H , as shown in Table 1.

[0070] 6. According to the load current I L , diode model V D =f -1 (I D) and switch device models (e.g., MOS device model V ds =g -1 (I mos ,V gs )) Calculate the conduction voltage V of the diode and MOS tube Don and V moson . Both are vectors of length N.

[0071] 7. According to V gs High or low level and load current I L Is it positive or negative? Determine the on-state device at this time and calculate the voltage V across each switch tube. ds and output current I s .Right now:

[0072] Table 1. Rules for determining steady-state values ​​of state variables

[0073] D <![CDATA[I L ]]> <![CDATA[V gs1_sta ]]> <![CDATA[V gs2_sta ]]> <![CDATA[V ds1_sta ]]> <![CDATA[V ds2_sta ]]> <![CDATA[I s1_sta ]]> <![CDATA[I s2_sta ]]> >Carrier <＝0 <![CDATA[V H ]]> <![CDATA[V L ]]> <![CDATA[V moson ]]> <![CDATA[V DC -V moson ]]> <![CDATA[-I L ]]> 0 >Carrier >0 <![CDATA[V H ]]> <![CDATA[V L ]]> <![CDATA[-V Don ]]> <![CDATA[V DC +V Don ]]> <![CDATA[-I L ]]> 0 <=Carrier <＝0 <![CDATA[V L ]]> <![CDATA[V H ]]> <![CDATA[V DC +V Don ]]> <![CDATA[-V Don ]]> 0 <![CDATA[I L <!-- 6 -->]]> <=Carrier >0 <![CDATA[V L ]]> <![CDATA[V H ]]> <![CDATA[V DC -V moson ]]> <![CDATA[V moson ]]> 0 <![CDATA[I L ]]>

[0074] At this point, the steady-state values ​​of the three state variables V ds_sta ,I s_sta and V gs_sta All are obtained.

[0075] S2, taking the voltage across both ends of each switch tube, the gate voltage and the output current as its state variables, and determining the transient initial value and driving voltage of the state variables of each switch tube;

[0076] Among them, the transient initial values ​​of the voltage across the ends of each switch tube and the output current are the steady-state values ​​of the voltage across the ends and the output current at the transient start time n respectively; if I L is positive, then the transient initial values ​​of the gate voltage and drive voltage of the first switch tube are both V L , the transient initial values ​​of the gate voltage and the driving voltage of the second switch tube are the steady-state values ​​of the gate voltage at time n and time n+1 respectively; if I L If V is negative, the initial transient values ​​of the gate voltage and drive voltage of the first switch tube are the steady-state values ​​of the gate voltage at time n and time n+1 respectively, and the initial transient values ​​of the gate voltage and drive voltage of the second switch tube are both V L .

[0077] Specifically, the initial value of the state variable and the driving voltage of the switch transient state are determined according to the steady-state value of the state variable. gs The waveform of the switching transient is determined to determine the start time of the switching transient. Then, the forward transistor of the two IGBT / MOS switching transistors is determined according to the direction of the output current. For the forward transistor and the non-forward transistor, the specific initial value of the state variable and the driving voltage V are respectively specified according to the switching waveform at the current moment. drive =[V drive1 ,V drive2 ].

[0078] Calculate V gs_sta The moment when the amplitude starts to jump (that is, the sequence number of the array) is stored in the transient start time matrix T _pul =[t _pul1 ,t _pul2 ,…,t _pulm ], where m is the number of transient states.

[0079] State variable I s and V ds The initial value is equal to the steady-state value of the state variable at the beginning of the switching transient, and the steady-state value has been calculated in the previous step.

[0080] State variable V gs Initial value and driving voltage V drive The determination is based on the opening or closing of the forward tube. The forward tube refers to the MOS tube whose device current is in the same direction (from drain to source) as the current. At the beginning of each switching transient, the forward tube must be switched, and the non-forward tube is already in a steady state. The gate voltage V gs The corresponding steady-state value at the beginning of the transient state should be selected, and the driving voltage V drive The corresponding steady-state value at the end of the transient should be selected. gs and the driving voltage V drive The shutdown level V L This is because in order to avoid the bridge arm from being directly connected, when the forward tube is turned on or off, the non-forward tube must be in the off steady state.

[0081] Based on this, Table 2 lists the initial values ​​of state variables and the rules for determining the driving voltage. n is the time when the transient starts. The initial values ​​of state variables at the beginning of different transients all adopt this rule.

[0082] Table 2. Rules for determining the initial values ​​of state variables and driving voltage of two-level PWM converter

[0083] <![CDATA[V ds10 ]]> <![CDATA[V gs10 ]]> <![CDATA[I s10 ]]> <![CDATA[V drive1 ]]> <![CDATA[V ds20 ]]> <![CDATA[V gs20 ]]> <![CDATA[I s20 ]]> <![CDATA[V drive2 ]]> <![CDATA[I L >0]]> <![CDATA[V ds1 (n)]]> <![CDATA[V L ]]> <![CDATA[I s1 (n)]]> <![CDATA[V L ]]> <![CDATA[V ds2 (n)]]> <![CDATA[V gs2_sta (n)]]> <![CDATA[I s2 (n)]]> <![CDATA[V gs2_sta (n+1)]]> <![CDATA[I L ≤0]]> <![CDATA[V ds1 (n)]]> <![CDATA[V gs1_sta (n)]]> <![CDATA[I s1 (n)]]> <![CDATA[V gs1_sta (n+1)]]> <![CDATA[V ds2 (n)]]> <![CDATA[V L ]]> <![CDATA[I s2 (n)]]> <![CDATA[V L ]]>

[0084] Determine the driving voltage V drive Finally, according to V drive Select the drive resistor R g , when V drive When it is high level, the driving resistor R g Select R gon , when V drive When it is low level, the driving resistor R g Select R goff .

[0085] S3, a differential equation group of the state variables of each switch tube about the transient moment is established according to the switch tube model, the diode model and the voltage equation of the junction capacitance, and the transient initial value and the driving voltage of the state variables of each switch tube are combined to solve them to obtain the transient value of each state variable.

[0086] Specifically, the state equations are written and encoded according to the two-level PWM converter topology. The encoded gate voltage V gs , output voltage V ds and the device output current I s The state equations of these three state variables give rise to a function model for calculating the differentials of the state variables at different times.

[0087] The switching transient model of the two-level PWM converter is as follows: Figure 6 As shown, nonlinear switching devices, voltage-dependent junction capacitance, and skin effect are all included. The following circuit equations are written according to Kirchhoff's current law (KCL) and voltage law (KVL) and converted into state equations.

[0088] for Figure 6 The red node in is the gate or g pole of the MOS tube. The equations can be written according to KCL as follows. There are two MOS tubes here, so i = 1, 2, and there are two equations.

[0089]

[0090] for Figure 6 The blue node in is the drain or d-pole of the MOS tube. The equations can be written according to KCL as follows. There are two MOS tubes here, so i = 1, 2, and there are two equations.

[0091]

[0092] for Figure 6 The loop consisting of two MOS tubes, parasitic inductance and skin resistance, and DC power supply can be written as follows according to KVL. There are two MOS tubes here, so i = 1, 2, and there are two equations. It should also be noted that the current differentials of the two MOS tubes are the same, because the difference between the two currents is I L is a constant value. Here I ring It refers to the oscillating current flowing through the skin resistance, and its determination is introduced in the next section.

[0093]

[0094] In addition, the three capacitor voltages of each switch tube also satisfy the KVL theorem, that is,

[0095] V gsi +V gdi -Vdsi =0, i=1,2 (11)

[0096] After transforming equations (7)-(10) together, we can obtain the following set of state equations.

[0097]

[0098] Current I d By MOS current I mos and the diode current I D Synthesis, that is:

[0099] I di =I mosi -I Di , i=1,2 (13)

[0100] According to equation (2), the MOS current I mosi The state variable V dsi and V gsi Calculation. According to formula (4), the diode current I D The state variable V dsi Calculation, this is because the forward voltage of the diode is opposite to the voltage of the MOS tube V D =-V dsi According to equations (3) and (6), the voltage variable capacitor can be expressed by the state variable V dsi Calculate. So far, the right side of the equal sign in the equation group (11) can be obtained by the state variable V ds 、V gs and I s Solve, that is, transform the equation into a state equation group of state variables and state variable differentials. For the convenience of the following explanation, the three state variable equations can be written together in the following form. x represents the three state variables packaged, and y is the functional relationship between the state variable x and its differential, that is, equation (11).

[0101]

[0102] It can be understood that, in formula (12), the modulation strategy of the converter is SVPWM, and the diode model, the switch device model and the junction capacitance pressure change model of the switch tube adopt formulas (2), (3) and (5) respectively. The state variable differential equation group of the switch tube established is only used as an example, and the present invention does not make a unique limitation on this. For other modulation strategies and diode models, switch device models, and junction capacitance pressure change models of the switch tube, those skilled in the art can make selections according to actual conditions and derive the state variable differential equation group according to common knowledge.

[0103] The derivation of the state equation group adopts an existing algorithm, such as the improved Euler method, etc. Considering that the step size required for the improved Euler method and other solution methods to calculate the non-divergence is too small when dealing with this equation, therefore, preferably, the present invention selects the Runge-Kutta algorithm with higher accuracy and less divergence to solve and obtain the transient value of the state variable.

[0104] The pseudo code of the solution process is as follows. For the transient start time vector T pul Each element t in pul By repeating this process, the transient values ​​of all state variables can be calculated.

[0105] Algorithm: Calculation of transient values ​​of state variables.

[0106] Input: state equation model y, state variable transient initial value V ds0 V gs0 I s0 , transient time t pul .

[0107] Output: transient value of state variables.

[0108] 1. Set the maximum calculation time to 20us and the calculation step h to 0.25ns. To ensure accuracy, the calculation step of transient state variables is smaller than that of steady-state state variables. The initial value of skin resistance is set to zero. According to the load current I L and transient time t pul Calculate the transient load current i L .

[0109] 2. According to the two state variables I s1 I s2 The initial value of determines the ringing current term I ring The basis for this choice is that the state variable with a larger absolute amplitude will only have a ringing current during the ringing period at the end of the transient process, and no AC power current. The ringing current at this time is I ring .

[0110] Table 3. Ringing current term I ring Selection rules

[0111] Criteria <![CDATA[I ring ]]> <![CDATA[|I s1 |>|I s2 |]]> <![CDATA[I s1 ]]> <![CDATA[|I s1 |<=|I s2 |]]> <![CDATA[I s2 ]]>

[0112] 3. Use the Runge-Kutta method to calculate the state variables at the next time n+1.

[0113]

[0114] 4. Determine whether the diode is in action at each step h of each transient process. If it starts to act, the switching waveform enters the oscillation stage, and the skin resistance R qIt should take effect. It can be understood that this step is a preferred step.

[0115] When solving the state equation group, the transient end criterion used is: Taking N = 50, M = J = 10000 as an example, calculate the variance of all state variables from n-49 to n, if the state variable V ds The variance is less than the input voltage V DC One ten-thousandth of, and V gs The variance is less than the high and low level difference V H -V L One ten-thousandth of s The variance of the load current i L , then it can be considered that the state variable has reached a steady state, that is, the transient process is over, and the transient value V of the state variable obtained at this time is stored. ds_pul and I s_pul , and the transient occurrence time t pul If not, go back to step 3 and use the calculated x(n+1) to continue calculating x(n+2) according to the Runge-Kutta algorithm and repeat the above process.

[0116] In summary, in S3, the state equations are solved by combining the initial state value and the driving voltage. During the solution process, it is determined whether the transient state has ended and the skin resistance is updated. For each transient state of the switching waveform obtained in S1, the following process is calculated in parallel. Set the sampling frequency f to 4 GHz. s , and get the time series. At each moment, based on the state variable value of the previous moment, first calculate the voltage-variable junction capacitance of the switch tube according to the capacitance-voltage-variable model, calculate the working current of the current switch device according to the device I / V curve model, and then call the function model in S3 to calculate the state variable differential, determine the state variable value at the current moment according to the state variable differential, and iterate the calculation. When the calculation reaches the time length (about N / f s =12.5ns), the variance of the last N values ​​of the calculated state variables is used to determine whether the state variable waveform has ended the transient state. In addition, the diode current calculated from the state variables is used to determine whether the waveform has entered the ringing stage. At this time, the skin resistance is corrected to make the skin effect take effect.

[0117] S4, the steady-state value and transient value of each state variable of each switch tube are spliced ​​to obtain the complete time domain waveform of each state variable, so as to calculate the common-mode and differential-mode interference source switching waveforms of the converter, and the detection processing is performed to obtain the two-level PWM converter switching waveform for electromagnetic interference prediction.

[0118] Specifically, step S4 includes:

[0119] S41, synthesize the complete state variable time domain waveform using the steady-state value and transient value of the state variable.

[0120] The steady-state value of the state variable obtained by S3 is replaced by the transient state variable value of the switch obtained in S4 during the switching transient time, and a device state variable time domain waveform is synthesized that takes into account the device nonlinear I / V characteristics, device junction capacitance pressure variation effect, and skin effect.

[0121] For m transient states, repeat the following process: set the steady-state value of the state variable V ds_sta and I s_sta Starting from the transient state t_ pul The following waveform is replaced by the transient value V ds_pul and I s_pul , the complete state variables can be synthesized. The pseudo code is as follows:

[0122] Algorithm: Synthesize the complete state variable V ds , I s .

[0123] Input: Steady-state value of state variable V ds_sta and I s_sta , transient start time t _pul , transient state variable V ds_pul and I s_pul .

[0124] Output: Contains transient and steady-state state variables V ds , I s .

[0125] 1. Calculate the duration p of the transient state variable.

[0126] 2. If the transient duration p is equal to the transient start time t _pul The sum does not exceed the longest time t of the steady-state state variable end , then the steady-state value of the state variable is changed from the transient start time t _pul to p+t _pul All steady-state variables are replaced by transient values.

[0127] 3. If the transient duration p is equal to the transient start time t _pul The sum has exceeded the longest time t of the steady-state state variable end , then only the steady-state variables are changed from the transient start time t _pul Until the last time t end The steady-state state variables are replaced by transient values.

[0128] S42, according to V ds and I s The state variables calculate the switching waveforms of common-mode and differential-mode interferers.

[0129] In view of the symmetry of the different phase bridge arms in the inverter, based on the Vds and I s The state variables generate the state variables of other bridge arms. According to the functional relationship between the common-differential mode interference source and the bridge arm state variables, the switching waveform of the common-differential mode interference source is calculated.

[0130] Two-level PWM converters are divided into three-phase and single-phase, which need to be discussed separately.

[0131] For a single-phase PWM converter, such as Figure 7 (a) in the figure generally adopts unipolar frequency multiplication modulation technology. At this time, the switch voltage of the other bridge arm can be calculated as follows: V ds3 Equal to V ds2 Phase shift half a cycle, V ds4 Equal to V ds1 Phase shifted by half a cycle. s3 Equal to I s2 Phase shift half a cycle, I s4 Equal to I s1 The phase shift is half a cycle. This is because, in unipolar frequency modulation, the carriers of the two bridge arms are 180° apart.

[0132] The switching waveform of the common-mode interference source of the single-phase inverter is equal to the average value of the voltage waveform of the two lower tubes of the bridge arm and the sum of the current of the two lower tubes of the bridge arm, that is:

[0133]

[0134] For a three-phase PWM converter, such as Figure 7 In (b), space vector modulation technology is generally used. At this time, the switch voltage of the other bridge arm can be calculated as follows: V ds3 Equal to V ds1 Phase shift 1 / 3 cycle, V ds5 Equal to V ds1 Phase shift 2 / 3 cycle. V ds4 Equal to V ds2 Phase shift 1 / 3 cycle, V ds6 Equal to V ds2 Phase shift 2 / 3 cycle. s3 Equal to I s1 Phase shift 1 / 3 cycle, I s5 Equal to I s1 Phase shift 2 / 3 cycle. s4 Equal to I s2 Phase shift 1 / 3 cycle, I s6 Equal to I s2 The phase is shifted by 2 / 3 of a cycle. This is because, in space vector modulation, the carriers of the two bridge arms are 120° apart.

[0135] The switching waveform of the common-mode interference source of the three-phase inverter is equal to the average value of the voltage waveform of the three bridge arms and the sum of the current of the three bridge arms, that is:

[0136]

[0137] S43, detecting the switching waveforms of the common-mode and differential-mode electromagnetic interference sources to obtain quasi-peak and average switching waveform spectra for electromagnetic interference prediction.

[0138] For example, the algorithm of simulating EMI receiver is used to convert the input time domain electromagnetic interference into a spectrum result similar to EMI receiver test according to the civilian or military electromagnetic interference standard. The spectrum result can be used for electromagnetic interference emission prediction. That is, the input of the function is the interference time domain waveform, and the output is the quasi-peak value and average value result after the simulation EMI standard specification detection.

[0139] The physical layer parameters of the known single-phase two-level PWM converter are shown in Table 4. The switching waveform spectrum calculation time of the common-mode and differential-mode interference sources based on the model proposed in the present invention is only 80 seconds.

[0140] Table 4 Physical layer parameters of single-phase two-level PWM converter

[0141]

[0142]

[0143] After model calculation, the drain-source voltage V ds can be predicted, the time domain waveform is as follows Figure 8 As shown in the figure, it can be seen from the switching transient waveform that the slope change characteristics when the voltage rises or falls can be simulated, which shows the accuracy of the model.

[0144] The output current of the switching device I s It can also be predicted that the time domain waveform is Fig. 9 As shown, the reverse recovery current peak when the switch is turned on can be accurately simulated.

[0145] For the switching waveform spectrum of the common differential mode interference source in the conduction frequency band, the results are as follows Fig.10 and Fig.11 Both the quasi-peak value and the average value can be modeled and predicted, and the ringing effect around 18MHz can be predicted.

[0146] The proposed model can also predict common and differential mode interference sources in the radiation band up to 1 GHz, such as Fig.12 and Fig.13 shown.

[0147] In summary, the present invention aims at the pain points of low accuracy and low efficiency of the existing switching waveform model used for electromagnetic interference prediction of two-level PWM converters, and proposes a wide-band, high-efficiency, high-precision switching waveform determination method based on a state equation group. The switching transient is solved based on the state equation group, and each switching transient is solved jointly according to the specified initial value of the state variable, the equivalent state equation group, the nonlinear device transfer function, and the voltage-change characteristic of the junction capacitance; the device switching waveform is synthesized with the steady-state value and the transient value of the state variable, and then the device switching waveform and the common-differential mode interference source switching waveform are converted to obtain the switching waveform for electromagnetic interference prediction, and finally the switching waveform spectrum is obtained after the detection processing of the analog EMI receiver.

[0148] An embodiment of the present invention provides an electronic device, including: a computer-readable storage medium and a processor;

[0149] The computer-readable storage medium is used to store executable instructions;

[0150] The processor is used to read the executable instructions stored in the computer-readable storage medium and execute the method described in any of the above embodiments.

[0151] An embodiment of the present invention provides a computer-readable storage medium, wherein the computer-readable storage medium stores computer instructions, and the computer instructions are used to enable a processor to execute the method described in any embodiment.

[0152] An embodiment of the present invention provides a computer program product, including a computer program or instructions, and when the computer program or instructions are executed by a processor, the method described in any embodiment is implemented.

[0153] It will be easily understood by those skilled in the art that the above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. Any modifications, equivalent substitutions and improvements made within the spirit and principles of the present invention should be included in the protection scope of the present invention.

Claims

1. A method for determining a switching waveform of a two-level PWM converter for electromagnetic interference prediction, characterized in that: include: S1, compare the modulated wave of the converter with the carrier wave to obtain the steady-state value V of the gate voltage of the first and second switching tubes in the converter gs1_sta 、V gs2_sta , and according to V gs1_sta 、V gs2_sta and load current I L Determine the steady-state value V of the voltage across the first and second switch tubes ds1_sta 、V ds2_sta and the steady-state value of the output current I s1_sta ,I s2_sta ; Among them, at a certain moment, if the modulation wave amplitude is greater than the carrier amplitude, then at this moment, V gs1_sta =V H , V gs2_sta =V L , I s1_sta =-I L , I s2_sta =0, when I L ≤0, V ds1_sta =V moson 、V ds2_sta =V DC -V moson , when I L >0, V ds1_sta =-V Don 、V ds2_sta =V DC +V Don ; If the modulation wave amplitude is less than or equal to the carrier amplitude, then at this moment, V gs1_sta =V L , V gs2_sta =V H , I s1_sta =0,I s2_sta =I L , when I L When V ds1_sta =V DC +V Don 、V ds2_sta =-V Don , when I L When negative, V ds1_sta =V DC -V moson 、V ds2_sta =V moson ; V H 、V L They are high and low level driving voltage, V DC 、V Don 、V moson They are input voltage, diode conduction voltage, and MOS tube conduction voltage respectively. S2, taking the voltage across both ends of each switch tube, the gate voltage and the output current as its state variables, and determining the transient initial value and driving voltage of the state variables of each switch tube; Among them, the transient initial values ​​of the voltage across the ends of each switch tube and the output current are the steady-state values ​​of the voltage across the ends and the output current at the transient start time n respectively; if I L is positive, then the transient initial values ​​of the gate voltage and drive voltage of the first switch tube are both V L , the transient initial values ​​of the gate voltage and the driving voltage of the second switch tube are the steady-state values ​​of the gate voltage at time n and time n+1 respectively; if I L If V is negative, the initial transient values ​​of the gate voltage and drive voltage of the first switch tube are the steady-state values ​​of the gate voltage at time n and time n+1 respectively, and the initial transient values ​​of the gate voltage and drive voltage of the second switch tube are both V L ; S3, establishing a differential equation group of the state variables of each switch tube with respect to the transient moment according to the switch tube model, the diode model and the voltage equation of the junction capacitance, and solving the differential equation group in combination with the transient initial value and the driving voltage of the state variables of each switch tube to obtain the transient value of each state variable; S4, the steady-state value and transient value of each state variable of each switch tube are spliced ​​to obtain the complete time domain waveform of each state variable, so as to calculate the common-mode and differential-mode interference source switching waveforms of the converter, and the detection processing is performed to obtain the two-level PWM converter switching waveform for electromagnetic interference prediction.

2. The method according to claim 1, characterized in that The switch tube model is: Among them, I mos is the channel current of the switch tube, V gs 、V ds They are the gate voltage and the voltage across the switch tube, V gsth 、V dsth , K are fitting parameters, V th is the gate threshold voltage, ∨ and ∧ are the calculation symbols of or and respectively.

3. The method according to claim 2, characterized in that The differential equations are: Where i=1,2, L is the parasitic inductance, C gdi , C gsi , C d ' si They are the drain gate capacitance, gate capacitance, and parasitic capacitance C of the i-th switch tube. D The equivalent capacitance between the drain and source of a MOS tube or between the emitter and collector of an IGBT, V drivei 、V gsi ,I si ,I di 、V dsi are the driving voltage, gate voltage, output current, drain current, and voltage across the i-th switch tube, respectively. DC is the input voltage, I ring is the oscillating current of the skin resistance, R q is the skin resistance, R g is the driving resistor.

4. The method according to claim 3, characterized in that In step S3, the Runge-Kutta algorithm is used to solve the differential equations.

5. The method according to claim 1 or 4, characterized in that In step S3, when solving the differential equations, the variance of the state variables at N adjacent transient moments is calculated. If the voltage V ds The variance is less than the input voltage V DC One-Mth of the gate voltage V gs The variance is less than one-J of the high and low levels of the driving voltage, and the output current I s If the variance of is less than the load current IL, the state variable has reached a steady state; where N, M, and J are all positive numbers.

6. The method according to claim 1 or 4, characterized in that: In step S3, when solving the differential equations, it is determined whether the diode is activated at each step length h of each transient process. If so, the skin resistance R q Not zero, otherwise R q zero.

7. An electronic device, characterized in that: include: A computer readable storage medium and a processor; The computer-readable storage medium is used to store executable instructions; The processor is used to read the executable instructions stored in the computer-readable storage medium and execute the method according to any one of claims 1 to 6.

8. A computer-readable storage medium, characterized in that: The computer-readable storage medium stores computer instructions, and the computer instructions are used to enable a processor to execute the method according to any one of claims 1 to 6.

9. A computer program product comprising a computer program or instructions, characterized in that When the computer program or instruction is executed by a processor, the method according to any one of claims 1 to 6 is implemented.

Images