Method for calculating magnetic core loss of solid-state transformer adaptive to characteristics of power system

Abstract

The invention discloses a solid-state transformer magnetic core loss calculation method adaptive to electric power system characteristics. The method comprises the following steps: firstly, collecting actually measured ferrite core loss data, and fitting hysteresis and eddy current loss parameters of a classical two-item model; setting 100kHz as a low-frequency and medium-high-frequency demarcation point through an eddy current and magnetic hysteresis loss balance point and a relaxation effect, and dividing three frequency bands by taking a skin effect characterization parameter greater than 2 as a medium-frequency and high-frequency demarcation point; adding a specific compensation item according to the loss characteristics of each frequency band, and constructing a segmented two-item model with compensation items; and fitting the parameters by adopting an L-M algorithm to calculate the total loss. According to the method, through frequency and flux density interval double division and compensation item introduction, the influence of nonlinear factors is effectively considered, and the magnetic core loss calculation precision is remarkably improved.

Description

Technical Field

[0001] The present invention relates to the technical field of power equipment loss calculation, and in particular to a solid-state transformer core loss calculation method adapted to power system characteristics. Background Art

[0002] In the context of developing new power systems, solid-state transformers (SSTs) are considered a popular solution for improving the distribution network's ability to accommodate the large-scale integration of renewable energy and enhance the control and maintenance capabilities of distribution networks. Due to their capabilities in grid voltage / reactive power regulation, frequency / active power regulation, and protection, solid-state transformers (SSTs) are considered a promising solution for enhancing the network's ability to accommodate renewable energy. During SST development, accurately calculating core loss and establishing a quantitative analysis model for the multi-dimensional influencing factors are key technical challenges in overcoming the bottlenecks of high-frequency and high-power density designs. Multi-dimensional factors, such as the broadband harmonic magnetic fields generated by high-frequency switching in SST topologies, the nonlinear dynamic hysteresis characteristics of the core material under high-frequency alternating magnetic fields, local flux distortion caused by multi-physics coupling, and the complex spatiotemporal harmonic distribution introduced by modular multilevel converters, lead to significant errors in traditional core loss calculation models based on power-frequency sinusoidal excitation in SST applications. Experimental data shows that when the switching frequency exceeds 5 kHz, the core loss prediction error of conventional models can exceed 40%, severely restricting the optimal design of high-frequency magnetic components. Therefore, constructing a refined calculation model of core loss suitable for high-frequency working conditions of solid-state transformers and accurately analyzing the synergistic mechanism of harmonic spectrum, hysteresis nonlinearity and multi-skin effect have become the core research directions for promoting the development of solid-state transformer technology towards high efficiency and miniaturization.

[0003] Currently, there are two main methods for calculating core loss. One is the Steinmetz equation (SE), used in engineering to estimate core loss, and its modified and generalized formulas. Because the Steinmetz equation and its modified formulas are empirical formulas, their applicable frequency and amplitude ranges are limited, making them inapplicable to all core materials. The other is a loss separation model based on the principle of core loss generation. A typical example is Jordan's 1924 separation of ferromagnetic loss into two terms: hysteresis and eddy current. This model is also known as the classical two-term model. The loss term coefficient is typically obtained by least squares fitting based on the loss and flux density test curves provided by the material manufacturer. Building on this, G. Bertotti supplemented the anomalous loss term based on the loss mechanism to derive the widely used classical three-term loss separation model, which includes hysteresis loss, eddy current loss, and anomalous loss. Although the two separation models differ in the number of loss terms, the loss coefficients are obtained in the same way, and the fitted curve obtained by summing all losses is similar to the measured curve. However, both the binomial model and the trinomial model are constant coefficient models, so the calculation is relatively simple and easy to implement. However, after considering the nonlinear characteristics of ferromagnetic materials and the influence of internal harmonic magnetic fields, the error is also relatively large.

[0004] In order to solve this problem, studying the variable coefficient loss model in which the loss coefficient changes with magnetic flux density and frequency has become an effective method. However, the traditional variable coefficient model expresses the total loss coefficient as K(B,f)=k1B through numerical fitting method. α f β Using a polynomial form, a global fitting strategy can lead to problems such as the inability to resolve local distortions in the hysteresis loop, the dynamic drift of permeability caused by temperature gradients, and the impact of pathological polynomial fitting. In particular, in the SST modular multi-level topology, the harmonic spectra experienced by the cores of each submodule vary significantly, significantly reducing the accuracy of traditional variable coefficient models.

[0005] In view of this, in order to solve the problems existing in the variable coefficient loss model, it is necessary to start from the mechanism of loss generation and study the loss variation law of the main loss items under different magnetic density and frequency conditions, and then propose a calculation model with wider applicability and higher calculation accuracy. Since the parameters in the two-term model are derived from the assumptions under low frequency, the physical explanation is clear, and the law of error variation with frequency and magnetic density is the same as the law of saturation effect, skin effect and other influences, which is convenient for further considering the influence of these effects on the basis of the original model. In order to obtain a loss model that can be more universal in engineering practice, the present invention proposes a ferrite core loss calculation model based on the two-term model, which is segmented and compensated for frequency and magnetic density, so as to achieve a refined analysis of the core loss and provide important technical support for promoting the development of SST technology towards high efficiency and miniaturization. Summary of the Invention

[0006] In view of the shortcomings of the prior art, the present invention provides a solid-state transformer core loss calculation method adapted to the characteristics of the power system, which aims to solve the problems in the background technology.

[0007] To achieve the above-mentioned purpose, the present invention provides the following technical solution: a method for calculating the core loss of a solid-state transformer adapted to the characteristics of a power system, comprising the following steps: collecting measured ferrite core loss data at a preset frequency and magnetic flux density, and obtaining the parameters of a classical two-term model through linear fitting; the parameters of the classical two-term model include hysteresis loss and eddy current loss; calculating the balance point of eddy current loss and hysteresis loss, and using the frequency corrected based on the relaxation effect as the dividing point between low frequency and medium-high frequency; calculating the skin effect characterization parameters of the ferrite core ,by The frequency when ≥2 is used as the dividing point between medium frequency and high frequency. Based on the dividing points between low frequency and medium and high frequency, and the dividing points between medium frequency and high frequency, the frequency is divided into low frequency band, medium frequency band and high frequency band. According to the loss characteristics of different frequency bands, specific loss compensation terms are added for correction to obtain the final segmented compensation term two-term loss calculation model. The LM algorithm is used to fit the parameters of the segmented compensation term two-term loss calculation model to calculate the total core loss. Specifically, specific loss compensation terms are added to correct the loss characteristics of different frequency bands to obtain the final segmented-band compensation term two-term loss calculation model. The specific process is as follows: based on the classic two-term constant coefficient model, a compensation coefficient that considers the effect of frequency on the loss coefficient and an eddy current loss flux density compensation term that considers the nonlinear factors of ferromagnetic materials are introduced in the low-frequency band; a hysteresis loss flux density compensation term that considers local hysteresis loops is introduced in the mid-frequency band; and an eddy current loss frequency correction term that considers the skin effect is introduced in the high-frequency band to obtain the final segmented-band compensation term two-term loss calculation model. Among them, the LM algorithm is used to fit the parameters of the two-term loss calculation model with segmented compensation terms, and the calculation of the total core loss includes: solving the parameters of the eddy current loss flux density compensation term of the two-term loss calculation model with segmented compensation terms; solving the parameters of the hysteresis loss flux density compensation term of the two-term loss calculation model with segmented compensation terms; solving the parameters of the eddy current loss frequency correction term of the two-term loss calculation model with segmented compensation terms.

[0008] Furthermore, the classic two-term constant coefficient model is shown in formula (1); (1); Where, Represents the total loss of the ferrite core; Represents hysteresis loss ; Indicates eddy current loss ; represents the constant coefficient related to hysteresis loss; represents the magnetic flux density; Indicates an index related to material properties; Indicates frequency; represents the constant coefficient related to eddy current loss; Expressed as: (2); Where, Indicates the thickness of the ferrite core; Indicates the mass density of the ferrite core; Represents resistivity.

[0009] Furthermore, the specific process of calculating the balance point of eddy current loss and hysteresis loss and using the frequency corrected based on the relaxation effect as the dividing point between low frequency and medium and high frequency is as follows: Taking the frequency dividing point where eddy current loss is equal to hysteresis loss, eddy current loss is proportional to the 1.5 power of frequency, and we get: (3); Considering the relaxation effect at high frequencies, which causes the magnetic moment reversal delay and reduces the effective magnetic permeability, the boundary point between low frequency and medium and high frequency is obtained to be 100kHz.

[0010] Furthermore, based on the classical two-term constant coefficient model, a compensation coefficient that considers the effect of frequency on the loss coefficient and an eddy current loss flux density compensation term that considers the nonlinear factors of ferromagnetic materials are introduced in the low-frequency band. A hysteresis loss flux density compensation term that considers the local hysteresis loop is introduced in the mid-frequency band, and an eddy current loss frequency correction term that considers the skin effect is introduced in the high-frequency band. The final segmented two-term loss calculation model with compensation terms is shown in Equation (4). (4); Where, Represents the total loss of the ferrite core; It represents the compensation factor that takes into account the effect of frequency on the loss factor. , It represents the maximum error of the segmented two-term loss calculation model with compensation terms within the compensation range at a frequency of 50 Hz. represents the eddy current loss flux density compensation term; represents the hysteresis loss flux density compensation term; represents the eddy current loss frequency correction term; Indicates removal Eddy current loss coefficient; Represents the coefficient related to the eddy current loss flux density compensation term; Represents the coefficient related to the hysteresis loss flux density compensation term; Represents the index related to the eddy current loss flux density compensation term; Represents the index related to the hysteresis loss flux density compensation term; Represents the index related to the eddy current loss frequency correction term; Represents an index related to the eddy current loss frequency correction term.

[0011] Furthermore, the eddy current loss and magnetic flux density compensation term parameters of the segmented two-term loss calculation model with compensation terms are solved: when the frequency f≤100kHz and the magnetic flux density B>1.4T, the eddy current loss and magnetic flux density compensation term parameters k2 and β2 are solved: (5); (6); Where, It represents the ratio of the measured value of the core loss divided by the frequency when the magnetic flux density takes different values ​​under the current frequency sinusoidal excitation; Indicates the measured value of core loss when the magnetic flux density takes different values ​​under the current frequency sinusoidal excitation; It represents the ratio of the calculated value of the classical two-term constant coefficient model within the current frequency and magnetic flux density range divided by the frequency; Indicates the calculated value of the classical two-term constant coefficient model under the current frequency and magnetic density range; Combining equations (5) and (6), we get: (7); Where, represents the eddy current loss flux density compensation term to be solved; Taking the logarithm of both sides of formula (7) yields: (8); According to formula (8), the linear fitting method is used to obtain the and For low frequency band and , is obtained by performing nonlinear surface fitting based on the LM algorithm through the loss data at different magnetic densities at multiple frequency points in the low frequency band according to formula (7).

[0012] Furthermore, the solution of the hysteresis loss flux density compensation term parameters of the segmented compensation term two-term loss calculation model is: when the frequency f>100kHz, the hysteresis loss flux density compensation term parameters are and Solve: (9); (10); Where, Represents the calculated value of the two-term loss model with eddy current loss flux density compensation term at the current frequency; express Ratio to the current frequency; Combining equations (9) and (10), we get: (11); Where, Represents the parameters of the hysteresis loss flux density compensation term to be solved; Taking the logarithm of both sides of equation (11), we get: (12); According to formula (12), the linear fitting method is used to obtain the and .

[0013] Furthermore, the solution of the eddy current loss frequency correction term parameters of the segmented two-term loss calculation model with compensation term is: when the frequency f≥500kHz, and Solve: (13); (14); Where, Indicates the calculated value of the two-term loss model with the hysteresis loss flux density compensation term and the eddy current loss flux density compensation term added at the current frequency; express Ratio to frequency; Combining equations (13) and (14), we get: (15); Taking the logarithm of both sides of equation (15), we get: (16); According to formula (16), the linear fitting method is used to obtain the and .

[0014] An electronic device includes a processor, a memory, and a bus, wherein the processor and the memory are connected via the bus, wherein the memory is used to store a set of program codes, and the processor is used to call the program codes stored in the memory to execute a solid-state transformer core loss calculation method adapted to the characteristics of the power system.

[0015] A non-volatile computer storage medium stores computer executable instructions, which can execute a solid-state transformer core loss calculation method adapted to power system characteristics.

[0016] Compared with the existing technology, the present invention has the following beneficial effects:

[0017] (1) The segmented two-term loss calculation model with compensation terms constructed in the present invention addresses the problem that the traditional model does not take into account magnetic saturation, harmonic phase superposition and skin effect under high-frequency working conditions of solid-state transformers, resulting in large errors. By dual division of frequency and magnetic flux intervals, eddy current and hysteresis loss flux density compensation terms are introduced on the basis of the two-term model, effectively considering the influence of nonlinear factors such as magnetic saturation, local hysteresis loops, eddy current skin effect and so on on the loss characteristics of ferrites, and significantly improving the calculation accuracy of core loss.

[0018] (2) The model of the present invention adds a piecewise constant function that changes with frequency and magnetic flux density, and is fitted and solved by the LM algorithm. It reduces the amount of calculation while avoiding the pathological characteristics of polynomial fitting, and does not change the loss parameters of the binomial model. It can intuitively reflect the influence of harmonic magnetic fields, nonlinearity of ferromagnetic materials and skin effect on hysteresis and eddy current losses. It has clear physical meaning and is convenient for loss mechanism analysis in engineering applications.

[0019] (3) The present invention divides the low-frequency, medium-frequency and high-frequency intervals based on the loss mechanism, and introduces specific compensation items according to the loss characteristics of each frequency band (such as the frequency compensation coefficient of the low-frequency band, the hysteresis loss flux density compensation item of the medium and high-frequency bands, and the skin effect correction item of the high-frequency band), so that the model can maintain high accuracy in a wide frequency range; the parameter solution combines linear fitting with the LM algorithm, and the corresponding magnetic density data is selected according to the frequency band to avoid overfitting, providing accurate loss calculation support for the optimization design of high-frequency magnetic components of solid-state transformers. BRIEF DESCRIPTION OF THE DRAWINGS

[0020] Figure 1 It is a flow chart of the steps of the present invention. DETAILED DESCRIPTION

[0021] like Figure 1 As shown, the present invention provides a technical solution: a method for calculating the core loss of a solid-state transformer adapted to the characteristics of a power system, comprising:

[0022] Step S1: Collecting the measured ferrite core loss data at a preset frequency (closest to 0) and magnetic flux density, and obtaining the parameters of the classical two-term model through linear fitting; the parameters of the classical two-term model include hysteresis loss and eddy current loss.

[0023] Step S2: Calculate the balance point of eddy current loss and hysteresis loss, and use the frequency corrected based on the relaxation effect as the dividing point between low frequency and medium and high frequency.

[0024] Step S3: Calculate the skin effect characterization parameters of the ferrite core ,by The frequency when >2 is used as the dividing point between the intermediate frequency and the high frequency.

[0025] Step S4: Based on the low frequency and medium-high frequency dividing points and the medium frequency and high frequency dividing points, the frequency is divided into a low frequency band, a medium frequency band and a high frequency band.

[0026] Step S5: According to the loss characteristics of different frequency bands, specific loss compensation items are added for correction to obtain the final segmented two-term loss calculation model with compensation items.

[0027] Step S6: Using the LM algorithm to fit the parameters of the segmented-band compensation term two-term loss calculation model to calculate the total core loss.

[0028] Among them, the classic two-term constant coefficient model is shown in formula (1); (1); Where, Represents the total loss of the ferrite core; Represents hysteresis loss ; Indicates eddy current loss ; represents the constant coefficient related to hysteresis loss; represents the magnetic flux density; Indicates an index related to material properties; Indicates frequency; represents a constant coefficient related to eddy current loss.

[0029] 、 It is obtained by global fitting of the measured ferrite core loss data, so Equation (1) only meets the accuracy requirements within a certain frequency and magnetic flux density range.

[0030] When the frequency is low, the demagnetization effect of eddy current can be ignored. It can be approximately obtained by formula (2); (2); Where, Indicates the thickness of the ferrite core; Indicates the mass density of the ferrite core; Represents resistivity.

[0031] Since the proportion of hysteresis loss and eddy current loss and the loss generation mechanism are different at different frequencies, and the core loss of the solid-state transformer is closely related to the internal and external working conditions, the corresponding frequency bands are divided according to the operating conditions of the solid-state transformer for analysis.

[0032] The operating frequency of a solid-state transformer is generally between 10kHz and 1MHz. When the transformer operating frequency is less than 100kHz, the operating principle of the high-frequency transformer is similar to that of a conventional transformer. Based on the law of electromagnetic induction, high-frequency alternating current is passed through the primary winding, generating an alternating magnetic flux in the magnetic core and inducing voltage in the secondary winding. At this time, hysteresis loss is generally greater than eddy current loss. When the transformer operating frequency exceeds 100kHz, eddy current loss increases significantly because hysteresis loss is proportional to frequency and eddy current loss is proportional to the 1.2 to 1.6 power of frequency.

[0033] The specific process of calculating the balance point of eddy current loss and hysteresis loss and correcting it to 100kHz as the dividing point between low frequency and medium and high frequency considering the relaxation effect is as follows: In order to better divide the typical frequency band affected by eddy current loss, the frequency dividing point is the point where eddy current loss is equal to hysteresis loss. Taking MnZn ferrite commonly used in solid-state transformers as an example, assuming = , eddy current loss is proportional to the 1.5th power of frequency, and we get: (3); Considering the relaxation effect at high frequency, the magnetic moment reversal delay (relaxation time 𝜏) reduces the effective magnetic permeability and weakens the magnetic flux change rate. , In order to suppress the generation of eddy currents, the corner frequency between the low frequency and medium and high frequency regions is corrected to 100kHz.

[0034] Among them, the calculation of the skin effect characterization parameters of the ferrite core ,by The specific process of using the frequency > 2 as the dividing point between the intermediate frequency and the high frequency is as follows: When the frequency continues to increase, the alternating magnetic flux generates an induced electromotive force when passing through the core, which in turn generates an eddy current. This eddy current generates a magnetic field inside the core, which in turn weakens the original magnetic field, and the magnetic flux is limited to the surface layer (penetration depth). ), the effective core volume is reduced, showing the magnetic skin effect. Multiply by the core thickness Skin effect characterization parameters To characterize the uneven distribution of eddy current density and magnetic flux density along the thickness direction of the core, when =0, indicating uniform distribution. The larger the value, the more uneven the magnetic flux density distribution. It is generally believed that when When it is less than 1, the skin effect can be ignored. When the value is greater than 2, the skin effect begins to become noticeable. Taking MnZn ferrite as an example, the skin effect has a significant impact on the frequency of 500kHz.

[0035] Among them, according to the loss characteristics of different frequency bands, specific loss compensation terms are added for correction to obtain the final segmented two-term loss calculation model with compensation terms. The specific process is as follows: The frequency is divided into low-frequency, medium-frequency and high-frequency analysis areas with 100kHz and 500kHz as the dividing points. Specific loss compensation items are added to make corrections based on the loss characteristics of different frequency bands.

[0036] In the low frequency band below 100kHz, the calculation error of the classic two-term constant coefficient model is relatively low, but it is still lower than the measured data in the area of ​​0~0.6T and above 1.4T, and the magnetic flux density during the operation of the solid-state transformer is mainly in the working range of 0.6T to 2T. By comparing the calculation errors in the low frequency band, it is not difficult to find that as the frequency increases, the calculation error of the classic two-term constant coefficient model in this range also gradually increases. This is mainly because the classic two-term constant coefficient model and The calculation is based on the assumption that the frequency is 50Hz, but in reality 、 and Since factors such as material nonlinearity will change with frequency, in the low-frequency band, when the original model does not introduce a correction term that takes into account the change in eddy current loss caused by material nonlinearity, it is necessary to introduce an error compensation coefficient that takes into account frequency changes in the classic two-term constant coefficient model. In the range above 1.4T, it was observed that the model calculation results were lower than the measured values, and the error increased with the increase of the magnetic flux amplitude. This is mainly because the eddy current loss term in the classic two-term constant coefficient model does not take into account the nonlinear factors of ferromagnetic materials. As the magnetic flux amplitude increases, the saturation degree deepens, and the additional value of eddy current loss caused by saturation increases. Therefore, in the high magnetic flux range above 1.4T, it is necessary to introduce an additional magnetic flux density term to take into account the additional value of eddy current loss as the magnetic flux amplitude increases.

[0037] In the medium and high frequency bands above 100kHz, regardless of the core flux density, its nonlinear characteristics are very severe. At this time, the increase in eddy current loss caused by the nonlinear characteristics of the ferromagnetic material will exist in the operating range of 0.6T to 2T. Therefore, the compensation range of the additional flux density term for eddy current loss will be expanded from above 1.4T in the low frequency band to the entire operating range above 0.6T. At the same time, at higher frequencies, the hysteresis loop distortion caused by the harmonic magnetic field of the ferromagnetic material is larger. As the frequency increases, the impact of the local hysteresis loop on the loss will be significantly strengthened. Therefore, an additional flux density term is introduced in the full magnetic density range of the medium and high frequency bands to consider the impact of the harmonic magnetic field on hysteresis loss.

[0038] At high frequencies above 500kHz, the results calculated using the classic two-term constant coefficient model are slightly larger than the measured values. In addition to the aforementioned reasons, the skin effect cannot be ignored. At this point, the demagnetizing effect of eddy currents gradually increases, and the skin effect becomes noticeable. Therefore, in this high frequency range, the exponential term needs to be corrected for frequency and material thickness.

[0039] In response to the above phenomenon, and in order to avoid the tedious calculation caused by fitting the measured ferrite core loss data at various frequencies and the influence of the pathological characteristics of polynomial fitting, based on the classical two-term constant coefficient model, a compensation coefficient that considers the influence of frequency on the loss coefficient and an eddy current loss flux density compensation term that considers the nonlinear factors of ferromagnetic materials are introduced in the low-frequency band. A hysteresis loss flux density compensation term that considers the local hysteresis loop is introduced in the mid-frequency band, and an eddy current loss frequency correction term that considers the skin effect is introduced in the high-frequency band to take into account the above influences. The final segmented two-term loss calculation model with compensation terms is shown in Equation (4).

[0040] (4); Where, Represents the total loss of the ferrite core; It represents the compensation factor that takes into account the effect of frequency on the loss factor. , It represents the maximum error of the two-term loss calculation model with compensation terms in the compensation range at a frequency of 50 Hz. represents the eddy current loss flux density compensation term; represents the hysteresis loss flux density compensation term; represents the eddy current loss frequency correction term; Indicates removal Eddy current loss coefficient; Represents the coefficient related to the eddy current loss flux density compensation term, which is used to describe the influence of magnetic flux density on eddy current loss; It represents the coefficient related to the hysteresis loss flux density compensation term, which is used to describe the influence of magnetic flux density on hysteresis loss; Represents an index related to the eddy current loss flux density compensation term, used to describe how the nonlinear effect of magnetic flux density affects eddy current loss; Represents an index related to the hysteresis loss flux density compensation term, which is used to describe how the nonlinear effect of flux density affects the hysteresis loss; An index related to the eddy current loss frequency correction term, used to describe the influence of characteristic length (possibly related to the skin effect) on eddy current loss; An index related to the frequency correction term of eddy current loss, used to describe the degree of influence of frequency on eddy current loss.

[0041] Among them, the segmented two-term loss calculation model with compensation terms has a total of 10 loss parameters, including 3 constant parameters ( 、 and ) and 7 variable parameters ( 、 、 、 、 、 、 );in, It changes with frequency, and because the proportion of eddy current loss increases at high frequency and high magnetic density, and the eddy current and hysteresis loss flux density compensation term is calculated through data at different frequencies in the frequency band, the influence of frequency is taken into account, so it is 1 at high frequency and high magnetic density; 、 、 、 、 、 In order to reflect the loss characteristics of different frequency bands, it is expressed as a piecewise constant function that varies with frequency. It can be seen that the piecewise two-term loss calculation model with compensation terms does not change the loss coefficient of the classic two-term constant coefficient model. Instead, it introduces compensation terms in different magnetic flux densities or frequency bands to ensure the applicability of the model by studying the physical nature of the effects of harmonic magnetic fields, ferromagnetic material nonlinearity, and skin effect on hysteresis and eddy current losses. Its characteristics are:

[0042] 1. Since the loss parameters of the classic two-term constant coefficient model are not changed, the influence of harmonic magnetic field, ferromagnetic material nonlinearity and skin effect on hysteresis and eddy current loss can be intuitively obtained through the segmented two-term loss calculation model with compensation terms.

[0043] 2. The method for solving model parameters is simpler and also avoids the pathological characteristics of polynomial fitting.

[0044] Among them, the LM algorithm is used to fit the parameters of the segmented compensation term two-term loss calculation model. The specific process of calculating the total core loss is: Solution of the compensation term parameters of eddy current loss and magnetic flux density in the segmented two-term loss calculation model with compensation terms: When the frequency f≤100kHz and the magnetic flux density B>1.4T, the eddy current loss magnetic flux density compensation term parameters k2 and β2 are solved based on the approximate assumption that the model calculation value is the measured value: (5); (6); Where, It represents the ratio of the measured value of the core loss when the magnetic flux density takes different values ​​under the current frequency sinusoidal excitation divided by the current frequency; Indicates the measured value of core loss when the magnetic flux density takes different values ​​under the current frequency sinusoidal excitation; It represents the ratio of the calculated value of the classical two-term constant coefficient model within the current frequency and magnetic flux density range divided by the current frequency; It represents the calculated value of the classical two-term constant coefficient model under the current frequency and magnetic density range. By substituting Available.

[0045] Combining equations (5) and (6), we get: (7); Where, Represents the eddy current loss flux density compensation term to be solved.

[0046] Taking the logarithm of both sides of formula (7) yields: (8); According to formula (8), the linear fitting method can be used to obtain the and For low frequency band and , is obtained by performing nonlinear surface fitting based on the LM algorithm through the loss data at different magnetic densities at multiple frequency points in the low frequency band according to formula (7).

[0047] When the frequency When the frequency is greater than 100kHz, the increase in hysteresis loss caused by local hysteresis loops cannot be ignored. To accurately calculate the parameters of the eddy current loss flux density compensation term while avoiding overfitting, the hysteresis loss flux density compensation term should be close to 1, that is, the measured loss data with a magnetic flux amplitude of about 1 T is selected. Therefore, when solving the eddy current loss flux density compensation term parameters in the medium and high frequency bands, the measured loss data with a magnetic flux amplitude of 0.6T to 1.5T is used for fitting. The eddy current loss flux density compensation term parameters in other magnetic flux density ranges are the same as the parameters for the magnetic flux density range of 0.6T≤B≤1.5T.

[0048] Solution of the hysteresis loss and magnetic flux density compensation term parameters of the segmented two-term loss calculation model with compensation terms: When the frequency f>100kHz, based on the approximate assumption that the model calculation value is the measured value, the hysteresis loss flux density compensation parameter and Solve: (9); (10); Where, represents the calculated value of the two-term loss model (i.e., Equation (5)) with eddy current loss flux density compensation at the current frequency; express The ratio to the current frequency.

[0049] Combining equations (9) and (10), we get: (11); Where, Represents the parameters of the hysteresis loss flux density compensation term to be solved.

[0050] Taking the logarithm of both sides of equation (11), we get: (12); According to formula (12), the linear fitting method can be used to obtain the and .

[0051] Since the amplitude of the harmonic magnetic field flux density is smaller than that of the fundamental magnetic field, in order to avoid the influence of the eddy current loss change on the calculation result of the hysteresis loss flux density compensation term, the eddy current loss flux density compensation term should be close to 1, that is, the flux density amplitude is close to 0; therefore, the measured loss data from 0 to 0.5T are selected for fitting.

[0052] Solution of the eddy current loss frequency correction term parameters of the segmented two-term loss calculation model with compensation terms: When the frequency f≥500kHz, based on the approximate assumption that the model calculated value is the measured value, and Solve: (13); (14); Where, Indicates the calculated value of the two-term loss model with the hysteresis loss flux density compensation term and the eddy current loss flux density compensation term added at the current frequency; express The ratio to the current frequency.

[0053] Combining equations (13) and (14), we get: (15); Taking the logarithm of both sides of equation (15), we get: (16); According to formula (16), the linear fitting method can be used to obtain the and .

[0054] Since the skin effect is very serious at any magnetic flux density amplitude at high frequencies, all measured loss data at each frequency are selected for fitting.

[0055] An electronic device includes a processor, a memory, and a bus, wherein the processor and the memory are connected via the bus, wherein the memory is used to store a set of program codes, and the processor is used to call the program codes stored in the memory to execute a solid-state transformer core loss calculation method adapted to the characteristics of the power system.

[0056] A non-volatile computer storage medium stores computer executable instructions, which can execute a solid-state transformer core loss calculation method adapted to power system characteristics.

[0057] While embodiments of the present invention have been shown and described, it will be appreciated by those skilled in the art that various changes, modifications, substitutions, and variations may be made to these embodiments without departing from the principles and spirit of the invention, and that the scope of the invention is defined by the appended claims and their equivalents.

Claims

1. A method for calculating core loss of solid-state transformers adapted to the characteristics of power systems, characterized in that: The method comprises the following steps: collecting measured ferrite core loss data at a preset frequency and magnetic flux density, and obtaining parameters of a classical two-term model through linear fitting; the parameters of the classical two-term model include hysteresis loss and eddy current loss; calculating the balance point of eddy current loss and hysteresis loss, and using the frequency corrected based on the relaxation effect as the dividing point between low frequency and medium and high frequency; Calculate the skin effect characterization parameters of ferrite cores ,by The frequency when ≥2 is used as the dividing point between medium frequency and high frequency. Based on the dividing points between low frequency and medium and high frequency, and the dividing points between medium frequency and high frequency, the frequency is divided into low frequency band, medium frequency band and high frequency band. According to the loss characteristics of different frequency bands, specific loss compensation terms are added for correction to obtain the final segmented compensation term two-term loss calculation model. The LM algorithm is used to fit the parameters of the segmented compensation term two-term loss calculation model to calculate the total core loss. Specifically, specific loss compensation terms are added to correct the loss characteristics of different frequency bands to obtain the final segmented-band compensation term two-term loss calculation model. The specific process is as follows: based on the classic two-term constant coefficient model, a compensation coefficient that considers the effect of frequency on the loss coefficient and an eddy current loss flux density compensation term that considers the nonlinear factors of ferromagnetic materials are introduced in the low-frequency band; a hysteresis loss flux density compensation term that considers local hysteresis loops is introduced in the mid-frequency band; and an eddy current loss frequency correction term that considers the skin effect is introduced in the high-frequency band to obtain the final segmented-band compensation term two-term loss calculation model. Among them, the LM algorithm is used to fit the parameters of the two-term loss calculation model with segmented compensation terms, and the calculation of the total core loss includes: solving the parameters of the eddy current loss flux density compensation term of the two-term loss calculation model with segmented compensation terms; solving the parameters of the hysteresis loss flux density compensation term of the two-term loss calculation model with segmented compensation terms; solving the parameters of the eddy current loss frequency correction term of the two-term loss calculation model with segmented compensation terms.

2. The method for calculating core loss of a solid-state transformer adapted to the characteristics of a power system according to claim 1, characterized in that: The classic two-term constant coefficient model is shown in formula (1); （1）； Where, Represents the total loss of the ferrite core; Represents hysteresis loss ; Indicates eddy current loss ; represents the constant coefficient related to hysteresis loss; represents the magnetic flux density; Indicates an index related to material properties; Indicates frequency; represents the constant coefficient related to eddy current loss; Expressed as: （2）； Where, Indicates the thickness of the ferrite core; Indicates the mass density of the ferrite core; Represents resistivity.

3. The method for calculating core loss of a solid-state transformer adapted to the characteristics of a power system according to claim 2, characterized in that: The specific process of calculating the balance point of eddy current loss and hysteresis loss and using the frequency corrected based on the relaxation effect as the dividing point between low frequency and medium and high frequency is as follows: Taking the frequency dividing point where eddy current loss is equal to hysteresis loss, eddy current loss is proportional to the 1.5 power of frequency, and we get: （3）； Considering the relaxation effect at high frequencies, which causes the magnetic moment reversal delay and reduces the effective magnetic permeability, the boundary point between low frequency and medium and high frequency is obtained to be 100kHz.

4. The method for calculating core loss of a solid-state transformer adapted to the characteristics of a power system according to claim 3, characterized in that: Based on the classical two-term constant coefficient model, a compensation coefficient that considers the effect of frequency on the loss coefficient and an eddy current loss flux density compensation term that considers the nonlinear factors of ferromagnetic materials are introduced in the low-frequency band. A hysteresis loss flux density compensation term that considers the local hysteresis loop is introduced in the mid-frequency band, and an eddy current loss frequency correction term that considers the skin effect is introduced in the high-frequency band. The final segmented two-term loss calculation model with compensation terms is shown in Equation (4). （4）； Where, Represents the total loss of the ferrite core; It represents the compensation factor that takes into account the effect of frequency on the loss factor. , It represents the maximum error of the segmented two-term loss calculation model with compensation terms within the compensation range at a frequency of 50 Hz. represents the eddy current loss flux density compensation term; represents the hysteresis loss flux density compensation term; represents the eddy current loss frequency correction term; Indicates removal Eddy current loss coefficient; Represents the coefficient related to the eddy current loss flux density compensation term; Represents the coefficient related to the hysteresis loss flux density compensation term; Represents the index related to the eddy current loss flux density compensation term; Represents the index related to the hysteresis loss flux density compensation term; Represents the index related to the eddy current loss frequency correction term; Represents an index related to the eddy current loss frequency correction term.

5. The method for calculating core loss of a solid-state transformer adapted to the characteristics of a power system according to claim 4, characterized in that: Solution of the eddy current loss and magnetic flux density compensation term parameters of the segmented two-term loss calculation model with compensation terms: When the frequency f≤100kHz and the magnetic flux density B>1.4T, the eddy current loss and magnetic flux density compensation term parameters k2 and β2 are solved: （5）； （6）； Where, It represents the ratio of the measured value of the core loss divided by the frequency when the magnetic flux density takes different values ​​under the current frequency sinusoidal excitation; Indicates the measured value of core loss when the magnetic flux density takes different values ​​under the current frequency sinusoidal excitation; It represents the ratio of the calculated value of the classical two-term constant coefficient model within the current frequency and magnetic flux density range divided by the frequency; Indicates the calculated value of the classical two-term constant coefficient model under the current frequency and magnetic density range; Combining equations (5) and (6), we get: （7）； Where, represents the eddy current loss flux density compensation term to be solved; Taking the logarithm of both sides of formula (7) yields: （8）； According to formula (8), the linear fitting method is used to obtain the and For low frequency band and , is obtained by performing nonlinear surface fitting based on the LM algorithm through the loss data at different magnetic densities at multiple frequency points in the low frequency band according to formula (7).

6. The method for calculating core loss of a solid-state transformer adapted to the characteristics of a power system according to claim 5, characterized in that: Solution of the hysteresis loss flux density compensation term parameters of the segmented belt compensation term two-term loss calculation model: When the frequency f>100kHz, the hysteresis loss flux density compensation term parameters and Solve: （9）； （10）； Where, Represents the calculated value of the two-term loss model with eddy current loss flux density compensation term at the current frequency; express Ratio to the current frequency; Combining equations (9) and (10), we get: （11）； Where, Represents the parameters of the hysteresis loss flux density compensation term to be solved; Taking the logarithm of both sides of equation (11), we get: （12）； According to formula (12), the linear fitting method is used to obtain the and .

7. The method for calculating core loss of a solid-state transformer adapted to the characteristics of a power system according to claim 6, characterized in that: Solution of the eddy current loss frequency correction term parameters of the segmented two-term loss calculation model with compensation terms: When the frequency f≥500kHz, and Solve: （13）； （14）； Where, Indicates the calculated value of the two-term loss model with the hysteresis loss flux density compensation term and the eddy current loss flux density compensation term added at the current frequency; express Ratio to frequency; Combining equations (13) and (14), we get: （15）； Taking the logarithm of both sides of equation (15), we get: （16）； According to formula (16), the linear fitting method is used to obtain the and .

8. An electronic device, characterized in that: The invention comprises a processor, a memory and a bus, wherein the processor and the memory are connected via the bus, wherein the memory is used to store a set of program codes, and the processor is used to call the program codes stored in the memory to execute the solid-state transformer core loss calculation method adapted to the power system characteristics according to any one of claims 1 to 7.

9. A non-volatile computer storage medium storing computer-executable instructions, characterized in that: The computer executable instructions execute the solid-state transformer core loss calculation method adapted to power system characteristics as described in any one of claims 1-7.

Images