A simplified method for extracting modal natural frequencies of power electronic converters

Abstract

The present application discloses a simplified method for extracting the modal natural frequency of a power electronic converter, the method comprising: step 1, first modeling the basic structural topology of the power electronic converter, recorded as the first model and obtaining the first vibration mode of each order of natural frequency; step 2, characteristic modal analysis of the first vibration mode, selecting the structure with the largest vibration amplitude in the first model, and determining the vibration source and vibration analysis object in the first model; step 3, second modeling according to the vibration source and the vibration analysis object, recorded as the second model and obtaining the natural frequency at a preset frequency; step 4, according to the characteristic modal analysis result of the first vibration mode, calculating the participation factor of each order of mode corresponding to the natural frequency of the second model at the preset frequency; step 5, selecting the natural frequency corresponding to the participation factor greater than the judgment threshold as the natural frequency of the power electronic converter. Through the technical solution in the present application, the modeling workload and the selection error of the natural frequency are reduced.

Description

Technical Field

[0001] The present application relates to the technical field of power electronic converters, and in particular to a simplified method for extracting the modal natural frequency of a power electronic converter. Background Art

[0002] In various power systems driven by power electronic converters, the switching action of power devices in the power electronic converters will not only generate high-frequency harmonic currents mainly at the PWM switching frequency, but also high-frequency vibrations mainly at the PWM switching frequency. At the same time, the power electronic converter will also be affected by external vibrations. When the frequency of these vibration excitations is consistent with the natural frequency of the power electronic converter, it will cause "resonance phenomenon".

[0003] Under the long-term effect of the "resonance phenomenon", stress and fatigue damage will be caused to the structure of the power electronic converter, which will cause the solder joints of the precision components inside the power electronic converter to break, causing the power electronic converter to fail. Therefore, in order to avoid the problem of power electronic converter failure caused by the "resonance phenomenon", under the premise that external vibration cannot be measured, it is necessary to design the switching frequency of the power electronic converter so that the high-frequency vibration mainly based on the PWM switching frequency avoids the resonant frequency of the power electronic converter. The basis for doing this is to be able to obtain the natural frequencies of each order of the power electronic converter.

[0004] In the prior art, most traditional modal simulation methods obtain natural frequencies similar to those of modal tests by special modeling of connection parts and parameter correction methods. This method has a good effect on extracting the natural frequencies of simple devices, but there are still the following problems:

[0005] 1. The simulation modeling time for devices with high structural complexity is too long;

[0006] 2. When the research object is connected to other devices, if the vibration correlation is not large, the natural frequency of the non-research object may be obtained, resulting in abnormal extraction of the natural frequency of the research object;

[0007] 3. Complex device simulation will obtain natural frequencies with denser spectrum distribution. The contribution of each order of natural frequencies to vibration is not the same. Subsequent processing is required to screen out the natural frequencies that have a greater impact on vibration, which invisibly adds a lot of additional work and increases the complexity of natural frequency extraction. Summary of the invention

[0008] The purpose of this application is to reduce the modeling workload in the natural frequency simulation process of a power electronic converter and to improve the accuracy of natural frequency selection.

[0009] The technical solution of the present application is: to provide a simplified method for extracting the modal natural frequency of a power electronic converter, the method comprising: step 1, first modeling the basic structural topology of the power electronic converter, recorded as the first model, and obtaining the first vibration mode of each order of natural frequency in the first model; step 2, performing characteristic modal analysis on the first vibration mode, selecting the structure with the largest vibration amplitude in the first model, and determining the vibration source and the vibration analysis object in the first model; step 3, performing second modeling based on the vibration source and the vibration analysis object, recorded as the second model, and obtaining the natural frequency of the second model at a preset frequency; step 4, calculating the participation factor of each order mode corresponding to the natural frequency of the second model at the preset frequency based on the characteristic modal analysis result of the first vibration mode; step 5, selecting the natural frequency corresponding to the participation factor greater than the judgment threshold according to the participation factor and the judgment threshold, and recording it as the natural frequency of the power electronic converter.

[0010] In any of the above technical solutions, further, the vibration source is a structure with the largest vibration amplitude in the first model, and the vibration analysis object is a structure in the first model that is closely fitted with the vibration source.

[0011] In any of the above technical solutions, further, step 3 specifically includes: step 31, calculating the relative amount of the maximum displacement of each part in the first vibration mode; step 32, determining the associated analysis object of the vibration source and / or the vibration analysis object according to the preset association conditions and the relative amount of the maximum displacement; step 33, performing second modeling based on the vibration source and / or the vibration analysis object, and the associated analysis object, recorded as the second model.

[0012] In any of the above technical solutions, further, the preset association condition is that the difference in the relative amount of the maximum displacement is within three orders of magnitude.

[0013] In any of the above technical solutions, further, the calculation expression of the participation factor is:

[0014]

[0015] Among them, d j is called the influence vector, the mode is i, the excitation direction is j, M is the mass matrix, u i is the eigenmode matrix.

[0016] In any of the above technical solutions, further, the judgment threshold is 1 / 3 of the maximum value of the participation factor.

[0017] The beneficial effects of this application are:

[0018] The technical solution in this application simplifies the modeling of complex structure circuit boards, reduces the modeling workload, and can effectively select the modal simulation natural frequency, reduce the error between the natural frequency and the experimental natural frequency, and improve the simulation accuracy. It is applicable to all types of board structure devices and has universal applicability.

[0019] (1) Compared with traditional simulation modeling methods, this application can reduce modeling time and improve modal simulation efficiency.

[0020] (2) For power electronic converters with multi-layer structures, the vibration transmission correlation between each layer of the structure is taken into account, and the basis for judging the vibration mode and displacement is introduced, and analyzed together with the experimental conditions, so that the modeling can focus on the research subject, which is conducive to the accuracy of modeling, modal analysis, and natural frequency selection.

[0021] (3) The appropriate natural frequency is further screened through the participation factor so that the natural frequency obtained by simulation can reflect the effective dynamic response to the excitation, thus ensuring the accuracy of the simulation.

[0022] (4) The method of the present application is concise and easy to promote. It can be extended to various systems with a single connection structure, making precise modal simulation of complex structural models possible, which has profound significance for subsequent vibration simulation. BRIEF DESCRIPTION OF THE DRAWINGS

[0023] The advantages of the above and / or additional aspects of the present application will become apparent and easily understood in the description of the embodiments in conjunction with the following drawings, in which:

[0024] Figure 1 This is a physical diagram of an isolated bidirectional DC / DC power electronic converter DAB control board according to an embodiment of the present application;

[0025] Figure 2 It is a complete model diagram of an isolated bidirectional DC / DC power electronic converter DAB control board according to an embodiment of the present application;

[0026] Figure 3 is a modal simulation natural frequency curve diagram of a DAB control panel according to an embodiment of the present application;

[0027] Figure 4 is a modal test simulation FFT curve diagram according to an embodiment of the present application;

[0028] Figure 5 is a schematic flow chart of a simplified method for extracting modal natural frequencies of a power electronic converter according to an embodiment of the present application;

[0029] Figure 6 is a basic topology model diagram of a DAB control board according to an embodiment of the present application;

[0030] Figure 7 is a schematic diagram of modal simulation vibration shapes and maximum displacement values ​​of a basic topological model of a DAB control panel according to an embodiment of the present application;

[0031] Figure 8 is a simplified model diagram of a DAB control panel according to one embodiment of the present application;

[0032] FIG9( a ) is a schematic diagram of modal simulation natural frequencies of a simplified model of a DAB control board according to an embodiment of the present application;

[0033] FIG9( b ) is a schematic diagram of a modal simulation vibration shape of a simplified model of a DAB control panel according to an embodiment of the present application;

[0034] Fig.10 is a simplified model diagram of a fine-modeled DAB control panel according to one embodiment of the present application;

[0035] Fig.11 is a modal simulation natural frequency curve of a simplified model of a DAB control panel according to an embodiment of the present application;

[0036] Fig.12 1 is a normalized Y-axis displacement participation factor curve of a simplified model of a finely modeled DAB control panel according to an embodiment of the present application. DETAILED DESCRIPTION

[0037] In order to more clearly understand the above-mentioned purposes, features and advantages of the present application, the present application is further described in detail below in conjunction with the accompanying drawings and specific implementation methods. It should be noted that the embodiments of the present application and the features in the embodiments can be combined with each other without conflict.

[0038] In the following description, many specific details are elaborated to facilitate a full understanding of the present application. However, the present application may also be implemented in other ways different from those described herein. Therefore, the protection scope of the present application is not limited to the specific embodiments disclosed below.

[0039] First, to illustrate the limitations of existing modal simulation methods, Solidworks software is used as the modeling software. Figure 1 and Figure 2 As shown in FIG. 1 , a model platform of an isolated bidirectional DC / DC converter (Dual Active Bridge, DAB) control board is established. The software used for natural frequency simulation is COMSOL Multiphysics, which is a finite element analysis software. The built DAB control board model is imported into COMSOL software. After modal simulation, the natural frequency curve below 2kHz is obtained as shown in FIG. Figure 3As shown, the obtained FFT curve and natural frequency are as follows Figure 4 shown.

[0040] By comparison, it can be seen that the natural frequencies obtained by simulation are too dense, with an average natural frequency difference of about 10 Hz per order, which cannot correspond to the experimental natural frequencies. Even with parameter correction, the distribution characteristics of the natural frequencies cannot be changed.

[0041] like Figure 5 As shown, this embodiment provides a simplified method for extracting the modal natural frequency of a power electronic converter, the method comprising:

[0042] Step 1: perform a first modeling on the basic structural topology of the power electronic converter, recorded as a first model, and obtain the first vibration mode of each order of natural frequency in the first model.

[0043] Specifically, from the physical Figure 1 It can be seen that the power electronic converter (DAB control board) is a layered structure, which is roughly divided into five layers, from top to bottom: Complex Programmable Logic Device (CPLD) board, sampling board, busbar, insulated gate bipolar transistor (IGBT) and heat sink. At the same time, there are two driver boards on the back connected to the heat sink. Among them, the connection between the upper three layers, the driver board and other layers is connected by fixed rods and fixed points, and the lower two layers are fixed by a close fit. Based on the principle of vibration transmission, it can be analyzed that the vibration transmission between the lower two layers is greater than that of other layers.

[0044] In Solidworks software, the basic structural topology modeling of each layer of the power electronic converter (DAB control board) is carried out, and the overall DAB special board simulation model is obtained after assembly, such as Figure 6 As shown in the figure, only the basic size and shape are modeled, and the heat sink is completely cut out because of the large gaps. The model of the power electronic converter is then imported into the COMSOL simulation software.

[0045] Step 2: Perform characteristic modal analysis on the first vibration mode, select the structure with the largest vibration amplitude in the first model, and determine the vibration source and vibration analysis object in the first model, wherein the vibration source is the structure with the largest vibration amplitude in the first model, and the vibration analysis object is the structure in the first model that is closely fitted with the vibration source.

[0046] In this embodiment, the material parameters of the power electronic converter (DAB control board) are set as shown in Table 1.

[0047] Table 1

[0048] Material Material parameters value steel density <![CDATA[7850kg / m 3 ]]> steel Young's modulus 240e9Pa steel Poisson's ratio 0.3 chip density <![CDATA[2420kg / m 3 ]]> chip Young's modulus 22e9Pa chip Poisson's ratio 0.3 Circuit Board density <![CDATA[1840kg / m 3 ]]> Circuit Board Young's modulus 14e9Pa Circuit Board Poisson's ratio 0.28

[0049] Set each layer of circuit board as circuit board material; IGBT as chip material; heat sink, busbar and connecting rod as steel material.

[0050] like Figure 7 As shown, it can be seen from the simulation results that the characteristic mode vibration type of the power electronic converter basically only vibrates on a certain layer of structure, and the vibrations between the layers of structure do not affect each other.

[0051] According to the orthogonality of the eigenmodes, the two eigenmode matrices u of the undamped structure are i and u j It can be shown to have the orthogonality shown in Formula 1.

[0052]

[0053] Where M is the mass matrix, K is the stiffness matrix, and the two eigenmode matrices u i and u j This is the result of the characteristic mode analysis of the first vibration mode.

[0054] Since characteristic modes can always be superimposed, two orthogonal modes can be obtained. That is, when the structure is excited by multiple frequencies and vibrates, each mode will only be linearly superimposed on its own vibration shape. Therefore, the vibration effects of different modes on different structures can be considered separately and then superimposed.

[0055] In this embodiment, the measuring points in the modal experiment are distributed on the radiator, and different DC voltages (such as 0V and 100V) are set in the related DAB control board vibration experiment. Through analysis, it is found that the vibration spectrum under different DC voltage conditions is basically unchanged, the peak of the vibration spectrum is mainly located at the switching frequency, and the vibration amplitude is the largest near the IGBT, indicating that the source of the vibration is the IGBT, and the IGBT is only tightly connected to the heat sink. Therefore, the heat sink is taken as the vibration analysis object, that is, the main research objects in this embodiment are the IGBT and the heat sink.

[0056] Step 3, performing a second modeling according to the vibration source and the vibration analysis object, recorded as a second model, and obtaining the natural frequency of the second model at a preset frequency;

[0057] Specifically, remove the parts of the above analysis that are not closely related to the vibration of the research object and remodel them, such as Figure 8After the modal simulation, 30 natural frequencies distributed around 1400 Hz (preset frequency) were obtained, as shown in Figure 9(b). There are only four natural frequencies below 3 kHz, and the vibration modes of each natural frequency can be reflected at each measuring point of the second model.

[0058] Then the above model is refined, such as Fig.10 As shown in the figure, the structure and size of the IGBT are determined according to the manufacturer's data sheet. The main body is modeled as an integrated whole due to its compact internal structure. The four-corner bolt connection holes are filled because they are actually fixed with bolts. Other parts are modeled realistically and given their own material properties.

[0059] For the heat sink, since its hollow structure has a great influence on the spatial distribution of mass stiffness, a real model is made. Repeat the previous simulation process and obtain 30 natural frequencies around 1400Hz (preset frequency) as follows Fig.11 As shown, it can be seen that its distribution is still relatively dense, and the average natural frequency of each order differs by about 25Hz.

[0060] Furthermore, step 3 specifically includes:

[0061] Step 31, calculating the relative maximum displacement of each part in the first vibration mode, including the relative maximum displacement of the CPLD board, the relative maximum displacement of the sampling board, the relative maximum displacement of the busbar, the relative maximum displacement of the IGBT, and the relative maximum displacement of the heat sink;

[0062] Step 32, determining the associated analysis object of the vibration source and / or the vibration analysis object according to the preset association condition and the relative amount of the maximum displacement; wherein the preset association condition is that the difference of the relative amount of the maximum displacement is within three orders of magnitude;

[0063] Step 33, performing second modeling based on the vibration source and / or the vibration analysis object, and the associated analysis object, recorded as the second model.

[0064] Specifically, the characteristic frequency search range is set, the minimum real part is 1 Hz, the maximum real part is 2 kHz, the maximum number of characteristic frequencies is 200, and the characteristic frequencies of each order are calculated.

[0065] For an undamped, multi-DOF system, the set of equations describing the characteristic motion of the system can be expressed in matrix form as Equation (2):

[0066]

[0067] Where [M] and [k] represent the mass matrix and stiffness matrix respectively. is the acceleration vector, {} is the displacement vector, and {()} is the external force vector.

[0068] After solving the finite element simultaneous eigenvalues, the eigenvalues ​​and eigenvectors of the characteristic motion equations can be obtained, where the eigenvalue is the natural frequency and the eigenvector is the mode vibration shape corresponding to the natural frequency of this order.

[0069] After that, the existing software is used to calculate the relative displacement of each part in different modal vibration modes (i.e., the normalized modal vector modulus), and the maximum relative displacement of the research subject (IGBT and heat sink) and other connection parts (CPLD board, sampling board, busbar) are taken to compare the relative displacement of the maximum displacement. If the difference between the two is more than three orders of magnitude, that is, it does not meet the preset correlation conditions, then it can be considered that the vibration correlation between the two is very small and can be analyzed independently of each other. If the difference between the two is within three orders of magnitude, then it is considered that the vibration correlation is large and the two need to be modeled together.

[0070] like Figure 7 In the case where the vibrating subjects are different from each other, the maximum relative displacements of the other structural layers differ from those of the research subjects IGBT and heat sink by more than three orders of magnitude. Therefore, the other structural layers that have little effect on the vibration of the IGBT and the heat sink can be ignored, and the modes of the IGBT and the heat sink that is tightly fitted to it can be studied separately.

[0071] Step 4: Calculate the participation factors of each mode corresponding to the natural frequency of the second model at a preset frequency according to the characteristic mode analysis result of the first vibration mode.

[0072] Specifically, the modal participation factor is a measure of the effectiveness of each excitation degree of freedom for each order of modal excitation. It is a relative quantity. If the value of the modal under the same excitation is larger, it means that the contribution of the modal to the dynamic response of the excitation is greater, that is, under the external excitation of the same amplitude, the vibration amplitude caused at the natural frequency of this order is relatively larger. Among them, the calculation expression of the participation factor is as follows:

[0073]

[0074] Among them, d j It is called the influence vector, which is the unit direction vector of displacement. All its components along the j-direction of freedom are 1, while all other components are 0. The mode is i, the excitation direction is j, M is the mass matrix, and u i is the characteristic mode matrix, that is, the eigenvalue matrix after solution.

[0075] Calculate the displacement participation factor of the main vibration direction (in this case, the Y axis) of each order of natural frequency. You can use unit mass normalization. The algorithm used is (mass matrix * displacement unit matrix Y component) / sqrt (modal normalization factor). The result is plotted as a curve graph as follows: Fig.12 shown.

[0076] Step 5: According to the participation factor and the judgment threshold, the natural frequency corresponding to the participation factor greater than the judgment threshold is selected and recorded as the natural frequency of the power electronic converter. The judgment threshold is 1 / 3 of the maximum value of the participation factor.

[0077] Specifically, the natural frequency with greater excitation effectiveness is screened according to the participation factor. Fig.12 The maximum value of the participation factor in is 3.19E-11, and its 1 / 3 value is taken as the judgment threshold, that is, the natural frequency with a participation factor greater than 1.06E-11 is taken as the natural frequency effective for the excitation response. There are four places in the figure where the participation factors meet the conditions, and the average value of the adjacent natural frequencies is taken.

[0078] In order to verify the accuracy of the simplified extraction method of the natural frequency involved in this embodiment, the obtained simulated natural frequency is compared with the experimental natural frequency, and the error is calculated, as shown in Table 2.

[0079] Table 2

[0080] Experimental natural frequency / Hz Simulation natural frequency / Hz error 737.543 749.95 1.620％ 997.020 1064.4 6.758％ 1531.607 1518.1 -0.882％ 1732.063 1703.8 -1.632％

[0081] It can be seen that the errors of the fourth-order modal natural frequency in the verification experiment are all within 10%, and except for the second-order error, the errors of the other orders are around 1%, indicating that the simulation results are basically consistent with the experimental results. This method can significantly improve the accuracy of natural frequency extraction.

[0082] The above describes the technical scheme of the present application in detail in combination with the accompanying drawings. The present application proposes a simplified method for extracting the modal natural frequency of a power electronic converter, including: step 1, first modeling the basic structural topology of the power electronic converter, recorded as the first model, and obtaining the first vibration mode of each order of natural frequency in the first model; step 2, characteristic modal analysis of the first vibration mode, selecting the structure with the largest vibration amplitude in the first model, and determining the vibration source and vibration analysis object in the first model; step 3, second modeling according to the vibration source and the vibration analysis object, recorded as the second model, and obtaining the natural frequency of the second model at a preset frequency; step 4, according to the characteristic modal analysis result of the first vibration mode, calculating the participation factor of each order of the natural frequency of the second model at the preset frequency; step 5, according to the participation factor and the judgment threshold, selecting the natural frequency corresponding to the participation factor greater than the judgment threshold, and recording it as the natural frequency of the power electronic converter. Through the technical scheme in the present application, the modeling workload is reduced, and the natural frequency of the modal simulation can be effectively selected, the selection error of the natural frequency is reduced, and the simulation accuracy is improved.

[0083] The steps in this application can be adjusted in order, combined, and deleted according to actual needs.

[0084] The units in the device of the present application can be combined, divided and deleted according to actual needs.

[0085] Although the present application is disclosed in detail with reference to the accompanying drawings, it should be understood that these descriptions are merely exemplary and are not intended to limit the application of the present application. The scope of protection of the present application is defined by the appended claims and may include various modifications, alterations and equivalents made to the invention without departing from the scope and spirit of the present application.

Claims

1. A simplified method for extracting the modal natural frequency of a power electronic converter, characterized in that: The method comprises: Step 1, performing a first modeling on the basic structural topology of the power electronic converter, recorded as a first model, and obtaining the first vibration mode of each order natural frequency in the first model; Step 2, performing characteristic mode analysis on the first vibration mode, selecting the structure with the largest vibration amplitude in the first model, and determining the vibration source and vibration analysis object in the first model; Step 3, performing a second modeling according to the vibration source and the vibration analysis object, recorded as a second model, and obtaining the natural frequency of the second model at a preset frequency; Step 4, calculating the participation factors of each mode corresponding to the natural frequency of the second model at a preset frequency according to the characteristic mode analysis result of the first vibration mode; Step 5, according to the participation factor and the judgment threshold, selecting the natural frequency corresponding to the participation factor greater than the judgment threshold, and recording it as the natural frequency of the power electronic converter; The step 3 specifically includes: Step 31, calculating the relative maximum displacement of each part in the first vibration mode; Step 32, determining the associated analysis object of the vibration source and / or the vibration analysis object according to the preset association condition and the relative amount of the maximum displacement; Step 33, performing a second modeling according to the vibration source and / or the vibration analysis object, and the associated analysis object, recorded as the second model; The calculation expression of the participation factor is: Among them, d j is called the influence vector, the mode is i, the excitation direction is j, M is the mass matrix, u i is the eigenmode matrix.

2. The simplified method for extracting the modal natural frequency of a power electronic converter according to claim 1, characterized in that: The vibration source is a structure with the largest vibration amplitude in the first model, and the vibration analysis object is a structure in the first model that is closely fitted with the vibration source.

3. The simplified method for extracting the modal natural frequency of a power electronic converter according to claim 1, characterized in that: The preset association condition is that the difference in the relative amount of the maximum displacement is within three orders of magnitude.

4. The simplified method for extracting the modal natural frequency of a power electronic converter according to any one of claims 1 to 3, characterized in that: The judgment threshold is 1 / 3 of the maximum value of the participation factor.

Images