Optimization design method for forced air cooling high-frequency transformer fin radiator suitable for source network load storage integrated system

Abstract

The invention discloses a forced air cooling high-frequency transformer fin radiator optimization design method suitable for a source network load storage integrated system. A polynomial response surface model capable of reflecting the heat spot temperature of a radiator is established; establishing a three-dimensional model of the fin radiator under different combination designs, and determining the highest temperature under different combination designs; taking the highest temperature of the fin radiator as an optimization target, taking the screened decisive influence factor as an independent variable, carrying out test design by adopting a response surface method, obtaining the influence degree of a response surface model and each variable on a result through residual analysis of the response surface method, and obtaining a final response surface function; and obtaining an optimal value of each variable by adopting a genetic algorithm according to the response surface model. According to the method, the influence of different variables on the highest temperature can be analyzed, the optimal solution can be directly obtained, and the calculation amount and the test frequency are reduced. The highest temperature of the high-frequency transformer can be reduced, and the highest temperature of the high-frequency transformer can be reduced to meet the engineering requirement.

Description

Technical Field

[0001] The invention relates to the technical field of high-frequency transformer design, and in particular to an optimization design method for a forced air-cooled high-frequency transformer fin heat sink suitable for a source-grid-load-storage integrated system. Background Art

[0002] Source-grid-load-storage integration is an operation mode and technology that can maximize the utilization of energy resources. Through various interactive forms such as source-source complementarity, source-grid coordination, grid-load interaction, grid-storage interaction, and source-load interaction, it can improve the dynamic balance capacity of the power system in a more economical, efficient and safe manner. It is an important development path for building a new power system. The power electronic transformer (PET) in the source-grid-load-storage integration system is a new type of power conversion device that integrates power electronic converters and high-frequency transformers. It is the core equipment for AC and DC power conversion and control. With the improvement of the power transmission capacity of high-voltage direct current transmission systems, the volume of high-frequency transformers is getting smaller and smaller, and the capacity is getting larger and larger. The loss of high-frequency transformers will increase accordingly with the increase of frequency, and most of it will be converted into heat energy. The reduction in volume will cause local temperature rise, which may damage the insulation, so the heat dissipation surface area of ​​high-frequency transformers continues to decrease. The reliability and life of high-frequency transformers may be affected if the design is unreasonable.

[0003] At present, power transformers generally use air cooling, water cooling and other cooling methods to dissipate heat. In terms of heat dissipation, there are mainly liquid cooling technology, increasing the heat dissipation area, optimizing the heat dissipation structure, etc. The principle of air cooling is to use the heat exchange with the flowing air in the main transformer room to dissipate the heat converted from a large amount of transformer losses through the plate radiator installed on the transformer. The plate radiator can effectively control the operating temperature of the transformer to ensure the safe operation of the transformer. Therefore, it is necessary to carry out reasonable heat dissipation optimization design to improve the heat exchange capacity of the transformer plate radiator and ensure the safe and stable operation of the transformer.

[0004] The heat sink for transformers mainly adopts the traditional sheet heat sink as the structural form. Nowadays, the widely used fin heat sink cannot effectively dissipate heat under high temperature and heavy load due to the increasing load on the power end, and cannot meet the heat dissipation needs of the transformer well. Many scholars have carried out a lot of research from different angles, mainly focusing on the optimization design of heat sinks with structural parameters to improve the heat dissipation efficiency of heat sinks for transformers, including changing the angle, increasing the number of heat sinks, adjusting the perforation diameter of the heat sink, optimizing the shape of the fins, and improving the external heat dissipation conditions of the heat sink surface material. Many scholars have carried out extensive research and in-depth discussions on strengthening the heat dissipation efficiency of heat sinks for transformers from different angles, so as to effectively enhance the heat dissipation capacity.

[0005] (1) Reference [1]: Liang Yiming, Ao Ming, Zhu Daming, et al. Optimization analysis of heat transfer characteristics of transformer plate heat sink [J]. Transformer, 2015, 52(04): 32-34. Through simulation analysis, this paper studies how the overall structural changes of different groups of heat sinks affect the performance of the heat sink, as well as factors such as the number of plate heat sinks. However, this document is designed for the plate heat sink structure of power frequency transformers and is not suitable for high-frequency transformers.

[0006] (2) Reference [2]: Shao Zhiwei, Huang Yaji, Zhang Qiang, et al. Numerical study on heat dissipation characteristics of plate-type heat sink [J]. East China Electric Power, 2014, 42(02): 431-434. The CFD method was used to study the effect of the number of heat sinks on the heat dissipation efficiency. The results showed that when the number of heat sinks increased from 10 to 20, the heat dissipation efficiency increased from 42.85% to 62.17%. However, this document analyzed the heat dissipation characteristics of the plate-type heat sink structure of an oil-immersed power frequency transformer. It quantitatively studied the effects of different transformer oil flow rates, the number of heat sinks, the heat sink spacing and the heat sink height on the heat dissipation capacity. No corresponding design method was given, and the conclusions obtained are not applicable to high-frequency transformers.

[0007] (3) Reference [3]: MAHDI, Valeartsв, et al. Two dimensional analysis of the edge effect field and losses in high frequency transformer foils [J]. IEEE Transactions on Magnetics, 2005, 41 (8): 2377-2383. The study investigated the effects of five different fin structures on the fluid flow and internal temperature of the radiator, and found that the trapezoidal fins with a wide top and narrow bottom exhibited the best heat exchange performance, which was better than the standard rectangular fins. This paper focused on analyzing the influence of various parameters of different fin structures on the heat exchange performance, but did not propose a design method for the radiator.

[0008] (4) Reference [4]: ​​Yang Jingshan, Numerical simulation of heat transfer enhancement of electronic device cooling radiator [D]. Huazhong University of Science and Technology, 2020. The influence of transformer fluid temperature and velocity was compared using Fluent software, and the different heights of the heat dissipation center of the radiator were compared. Increasing the heat dissipation center height of the fin can effectively improve the heat dissipation efficiency of the transformer oil, thereby reducing the temperature rise of the transformer. This document focuses on analyzing the influence of radiator structural parameters on heat dissipation efficiency, and does not propose a design method for the radiator. Summary of the invention

[0009] For the high-frequency transformer in the source-grid-load-storage integrated system, the present invention provides a forced air-cooled high-frequency transformer fin heat sink optimization design method suitable for the source-grid-load-storage integrated system, and obtains the conclusion on the optimization design of the heat dissipation structure of the high-frequency transformer based on the sensitivity analysis method and the response surface method. This method can not only analyze the influence of different variables on the maximum temperature, but also directly obtain the optimal solution, reducing the amount of calculation and the number of experiments. The present invention can reduce the maximum temperature of the high-frequency transformer, reduce the maximum temperature of the high-frequency transformer to meet the engineering requirements, and provide support for the design of high-reliability, high-power density high-frequency transformers in the source-grid-load-storage integrated system.

[0010] The technical solution adopted by the present invention is:

[0011] The optimization design method of the forced air-cooled high-frequency transformer fin heat sink applicable to the source-grid-load-storage integrated system includes the following steps:

[0012] Step 1: Establish a polynomial response surface model that can reflect the temperature of the heat sink;

[0013] Step 2: Establish a three-dimensional model of the fin heat sink under different combination designs and determine the maximum temperature under different combination designs;

[0014] Step 3: Taking the maximum temperature of the finned heat sink as the optimization target and the selected decisive influencing factors as independent variables, the response surface method is used for experimental design. Through the residual analysis of the response surface method, the response surface model and the influence of each variable on the result are obtained, and the final response surface function is obtained;

[0015] Step 4: Based on the response surface model, a genetic algorithm is used to obtain the optimal value of each variable.

[0016] In the step 1, the Box-Behnken surface response method is used to optimize the design for the heat dissipation performance of the heat sink under different combinations of structural parameters. In the Box-Behnken design, the experimental design table includes a set of center point and several star point experiments, as well as the combination of various factors at different levels. The second-order Box-Behnken response surface model is actually a multivariate quadratic equation, which is used to fit the functional relationship between the design parameters and the effect value. The order of the second-order response surface model is 2. Based on the selected second-order Box-Behnken response surface model, the specified fin thickness A, fin height B, and fin spacing C design parameters are experimentally sampled and sample values ​​are obtained. Finally, the coefficients to be determined are determined by the least squares method, and the explicit function of the actual model is finally obtained.

[0017] The second-order Box-Behnken response surface model contains the following cross terms:

[0018]

[0019] Where: represents the response variable; β0, β j , β jj , β ij are the unknown coefficients, which are the coefficients of the constant term, the first-order term, the second-order non-intersecting term, and the second-order intersecting term; x i 、x j is the design variable; n is the number of design variables.

[0020] In step 2, the central combination design factor level table is an experimental design method used to explore and verify the influence of the three factors of fin thickness A, fin height B, and fin spacing C on the maximum temperature. In this design, the experiment is divided into multiple combinations, each of which controls a factor and simulates its influence on the result. The central combination design factor level table is determined, and the electromagnetic-temperature-fluid multi-physics field finite element method is used to calculate the maximum temperature under different combination designs when the fins are kept at a constant temperature.

[0021] In step 2, according to the high-frequency transformer temperature results under different factor combination design conditions, the parameters affecting the maximum temperature of the transformer are determined, including fin thickness A, fin height B, fin spacing C, fin number N and other parameters.

[0022] Determine whether the variables to be optimized in the fin heat sink have an impact on the maximum temperature and the extent of the impact, thereby determining the main factors affecting the maximum temperature. Specifically: adopt the idea of ​​the control variable method, select any one of the parameters such as fin thickness A, fin height B, fin spacing C, fin number N, etc. one by one, select a suitable step size and change this parameter according to the step size, and keep other variables unchanged, establish a three-dimensional finite element model of the high-frequency transformer and the corresponding finite element model of the heat sink in the finite element simulation software Comsol, and calculate the maximum temperature under different heat sink parameters.

[0023] In step 3, assuming that the response surface function to be obtained requires k trials, the second-order Box-Behnken response surface model in step 1 can be transformed into a matrix form as follows:

[0024]

[0025] Where: y1 represents the maximum temperature simulation value under the first combination design; y2 represents the maximum temperature simulation value under the second combination design; y k represents the maximum temperature simulation value under the kth combination design;

[0026] x 11 represents the value of the first design factor under the first combination design; x 12 represents the value of the second design factor under the first combination design; x 1n-1represents the value of the n-1th design factor under the first combination design; x 1n It represents the value of the nth design factor under the first combination design;

[0027] x 21 represents the value of the first design factor under the second combination design; x 22 represents the value of the second design factor under the second combination design; x 2n-1 represents the value of the n-1th design factor under the second combination design; x 2n Indicates the value of the nth design factor under the second combination design;

[0028] x k1 represents the value of the first design factor under the kth combination design; x k2 represents the value of the second design factor under the kth combination design; x kn-1 represents the value of the n-1th design factor under the kth combination design; x kn represents the value of the nth design factor under the kth combination design;

[0029] β0 is a constant term; β1...β n is the first-order coefficient corresponding to the 1st...nth factors; β 11 ......β nn is the second-order non-intersecting coefficient corresponding to the 1st...nth factors; β 12 ......β nn-1 is the coefficient of the second-order cross terms between the 1st and 2nd factors, ..., the nth and n-1th factors; T represents the matrix transpose;

[0030] μ1 represents the fitting error under the first combination design; μ2 represents the fitting error under the second combination design; μ k represents the fitting error under the kth combination design.

[0031] According to the extreme value problem of unbiased estimation, the above formula is used to minimize the sum of squares of errors, thus obtaining the polynomial coefficients:

[0032] β=(X T X) -1 X T y

[0033] Where: β represents the vector of undetermined coefficients of the polynomial; X T represents the transpose of the design variable matrix; X represents the design variable matrix; y represents the Comsol simulation results of the corresponding variables.

[0034] Substitute the polynomial coefficient vector into step 1 to obtain the second-order Box-Behnken response surface model. In order to verify the applicability of the model, the model is subjected to residual analysis according to Design-Expert. In the residual normal probability plot, if the residuals are normally distributed and the residual points are evenly dispersed near a straight line, it means that the model fits well.

[0035] The overall efficiency predicted by the response surface model is given by the coefficient of determination R 2 Explanation, coefficient of determination R 2 It is the ratio of the regression sum of squares to the total sum of squares. The closer its value is to 1, the higher the accuracy. Its calculation formula is:

[0036]

[0037] Where: y is the sample value; is the sample mean; M represents the number of sample values ​​obtained through experimental design sampling; y i represents the finite element simulation results; The response variable represented by is calculated.

[0038] In step 3, the calculation process of the response surface method is:

[0039] Step (1): Determine the sampling space by variables such as fin thickness A, fin height B, and fin spacing C, and determine the level of the experimental factor; fin thickness A, fin height B, and fin spacing C all have three levels: low level (-1), center level (0) and high level (1), representing the lower limit, middle value, and upper limit of the factor.

[0040] Step (2): using the central composite design (CCD) method, the electromagnetic-temperature-fluid multi-physics field finite element analysis calculation is performed on the sampling points according to the experimental factor levels;

[0041] Step (3): taking the maximum temperature of the fin heat sink as the optimization target, taking the selected decisive influencing factors as independent variables, fitting the response surface function, and performing variance analysis;

[0042] Step (4): Based on the residual normal probability plot analysis, insignificant items are eliminated and significant influencing factors are determined;

[0043] Step (5): determine whether the accuracy is met;

[0044] If the conditions are met, the response surface model can be obtained; if not, the number of terms or order of the response surface model is increased and step (1) is repeated. In step 4, the optimal value of each variable of the radiator is obtained by using a genetic algorithm according to the response surface model.

[0045] Taking the minimization of the maximum temperature t of the high-frequency transformer as the objective function, considering the value ranges of the fin thickness A, fin height B, and fin spacing C of the radiator, an optimization design model is established:

[0046] The objective function is:

[0047]

[0048] The constraints are:

[0049] A∈[A min ,A max ]; B∈[B min ,B max ]; C∈[C min ,C max ]

[0050] The individual fitness evaluation function is:

[0051]

[0052] The genetic algorithm takes the initial solution as the starting point, and generates the next generation of population through genetic operations such as selection, crossover, and mutation. The individual fitness evaluation function is used to examine the quality of each individual in the population, and high-quality individuals are selected to enter the next generation. After multiple iterations, the optimal solution is finally obtained.

[0053] The present invention provides an optimization design method for a forced air-cooled high-frequency transformer fin heat sink suitable for a source-grid-load-storage integrated system, and the technical effects are as follows:

[0054] 1) The present invention proposes to suppress the maximum temperature of the high-frequency transformer by adding a fin heat sink, and calculate the temperature distribution of the high-frequency transformer by using the electromagnetic-temperature-fluid multi-physics field finite element analysis method; combine the sensitivity analysis method to study the influence of various size parameters of the fin heat sink on the maximum temperature of the high-frequency transformer, and determine the decisive influencing factors of the maximum temperature; use the response surface method to construct a response surface function related to the maximum temperature, and optimize the transformer fin heat sink.

[0055] 2) The design method of the present invention can be applied to the optimization design of the heat dissipation structure of a transformer with a double-winding or multi-winding structure. The maximum temperature can be calculated according to the response surface function, which reduces the amount of calculation and is beneficial to engineering applications.

[0056] 3) Response surface method (RSM) is a mathematical and statistical method for optimization. It has been widely used for its independence and convenience in establishing approximate models. It can quickly and conveniently handle various optimization problems under the premise of ensuring accuracy and efficiency. Therefore, the present invention applies the response surface method to the structural optimization problem of the fin heat sink of the high-frequency transformer, and can fit an explicit response surface function, which is convenient for the structural design of the fin heat sink of the high-frequency transformer, and quantitatively analyzes the influence of various design parameters. BRIEF DESCRIPTION OF THE DRAWINGS

[0057] The present invention will be further described below in conjunction with the accompanying drawings and embodiments:

[0058] Figure 1 A flow chart established for the design method of the present invention.

[0059] FIG2(a) is a schematic diagram of the front view of the high-frequency transformer structure;

[0060] FIG2( b ) is a schematic side view of the high frequency transformer structure;

[0061] FIG2( c ) is a schematic diagram of a top view of the high-frequency transformer structure.

[0062] Figure 3(a) is a two-dimensional model of a fin heat sink;

[0063] Figure 3(b) is a three-dimensional finite element model of the fin heat sink.

[0064] Figure 4(a) is a diagram showing the core loss distribution of a high-frequency transformer without a finned heat sink installed;

[0065] Figure 4(b) is a diagram showing the loss distribution of the high-frequency transformer winding without a finned heat sink installed;

[0066] Figure 4(c) shows the temperature distribution of the high-frequency transformer without a finned heat sink.

[0067] Figure 5 The residuals follow a normal distribution.

[0068] FIG6( a ) is a schematic diagram of heat dissipation performance under different combinations of fin height and fin thickness parameters;

[0069] FIG6( b ) is a schematic diagram of heat dissipation performance under different combinations of fin spacing and fin thickness parameters;

[0070] Figure 6(c) shows the heat dissipation performance under different combinations of fin spacing and fin height parameters.

[0071] Figure 7 This is the genetic algorithm operation flow chart.

[0072] Figure 8 This is the temperature distribution diagram of the high-frequency transformer under forced air cooling conditions after fin optimization.

[0073] Fig. 9 This is the air flow distribution diagram of the high-frequency transformer under forced air cooling conditions after fin optimization. DETAILED DESCRIPTION

[0074] Example:

[0075] For a single-phase double-winding core-type high-frequency transformer, the outer winding uses a 20-turn primary winding with a rated voltage of 500 V. The inner winding is a secondary winding with 20 turns and a rated voltage of 500 V. Figure 2 shows the design structure of the core-type high-frequency transformer, and Table 1 lists the size parameters of each component of the high-frequency transformer.

[0076] Table 1 High frequency transformer model design parameters

[0077]

[0078] The structure and size of the fin heat sink are determined according to the transformer model to ensure that the maximum temperature meets the insulation requirements under rated operating conditions. The two-dimensional schematic diagram of the fin heat sink is shown in Figure 3(a), and the established three-dimensional simulation model of the fin heat sink is shown in Figure 3(b). It is necessary to clarify the influence of each geometric parameter of the fin heat sink on the maximum temperature and select the determining factors of the maximum temperature. The sensitivity analysis method will be used to obtain the influence of each factor on the maximum temperature. The thickness A of the fin, the height B of the fin, and the spacing C of the fin all have a great influence on the maximum temperature.

[0079] According to the design process of the response surface model, the experimental data under different factors are shown in Table 2. In Table 2, the fin thickness A, fin height B, and fin spacing C all have three levels: low level (-1), center level (0), and high level (1), representing the lower limit, middle value, and upper limit of the factor.

[0080] Table 2 Central composite design factor level table

[0081]

[0082]

[0083] The finite element method is used to calculate the core loss, winding loss, and maximum temperature when the finned heat sink is not installed. The distribution cloud diagram is as follows Figure 4(a) to Figure 4(c) As shown. Figure 4(a) to Figure 4(c)It can be seen that the core loss and winding loss are not evenly distributed, but the loss density is large at the inner corner of the core, and the loss density between the primary and secondary windings is large. Since the secondary winding is wrapped by the core and winding, the heat dissipation path is not smooth, so the highest temperature is located in the secondary winding. The core loss is 51.21W, the winding loss is 18.39W, and the maximum temperature is 105℃. The core loss and winding loss will be used as heat sources to analyze the effect of installing a fin heat sink. Under the constant temperature condition of the fin, the temperature field calculation results of each measuring point are shown in Table 3. It can be seen from Table 3 that when the radiator maintains a constant temperature, the maximum temperature of the radiator in the heat dissipation stage is 32.73℃, and the minimum temperature is 29.94℃.

[0084] Table 3 Temperature simulation results of different experimental points when the fin is kept at constant temperature

[0085]

[0086] Using the second-order Box-Behnken response surface model, the temperature of the fin under the working conditions shown in Table 3 was regressed and analyzed, and Table 4 was obtained. Here R 2 It indicates the conformity of the fitting function with the experimental data, and R 2 The closer it is to 1, the more significant the model is. As can be seen from Table 4, the standard deviation of the secondary model is very small, R 2 It is also close to 1, indicating that the second-order Box-Behnken response surface model is stable. In order to further study the correctness of the second-order model, a variance analysis was performed on the fin temperature in Table 3, and the results are shown in Table 5.

[0087] Table 4 Regression results of the model

[0088]

[0089] Table 5 Variance analysis of fin heat sink temperature

[0090]

[0091] It can be clearly observed in Table 5 that when the second-order model is fitted for the heat dissipation temperature, the significance level P value is calculated to be 0.0001, which is far below the significance level threshold of 0.05, thus proving the validity and relevance of the established model. The applicability of the model has been statistically confirmed. Through the analysis and mathematical fitting of the data, a second-order mathematical expression closely related to the heat dissipation temperature t is obtained. This equation can more accurately describe the relationship between the actual measured value and the predicted value:

[0092] t=31.21-0.10059A-0.7978B-0.4855C+0.0699AB-0.0984AC+0.0460BC

[0093] -0.3239A 2 +0.1721B 2 +0.2718C 2

[0094] In order to further verify the applicability of the model, residual analysis of the model was performed according to Design-Expert. Figure 5 This is the residual normal probability plot. The residuals are normally distributed, and the residual points are evenly dispersed near a straight line. Although there is a slight offset at a certain point, it has little effect on the result, indicating that the model fits well.

[0095] To ensure the accuracy and scientificity of the research, the Box-Behnken response surface method can be used to explore in detail the effects of different parameters such as fin thickness, spacing, and material properties on the radiator efficiency, thereby providing theoretical support for the optimal design of the radiator.

[0096] according to Figure 6(a) to Figure 6(c) It can be concluded that as the heat sink temperature decreases, the heat dissipation performance of the heat sink increases with the increase of the heat sink height and thickness; at the same time, reducing the distance between the heat sinks and increasing the thickness of the heat sink will also cause the heat sink temperature to decrease, thereby improving the heat dissipation. The heat dissipation effect will be improved when the heat sink temperature decreases, the heat sink spacing decreases, and the height increases. That is, the heat dissipation performance will be different under different fin structure parameter combinations, and each group of parameter combinations has roughly the same degree of influence on the heat dissipation performance.

[0097] Through the above detailed analysis, it is determined that three key parameters have the most significant impact on the heat dissipation performance of the fin. Figure 7 The genetic algorithm shown in the figure determines the optimal geometric parameters of the fins. The first is the thickness A of the fin, and its value is set to 1.997mm. This parameter is directly related to the size of the contact area between the fin and the air; the second is the fin height B, and the selected value is 11.994mm, which determines the height of the fin and the resistance to air flow; the last is the fin spacing C, which is set to 3.952mm. This distance affects the efficiency of heat transfer. Combining these key factors, a fin model with the best heat dissipation performance can be obtained, as shown in Table 6. In order to verify the effectiveness of the prediction model, the temperature of the high-frequency transformer in natural convection and without a heat sink is simulated and verified by comparing the temperature of the forced air cooling and finned heat sink in the constant temperature stage. The results are shown in Figure 8 The maximum temperature of the fin during the heat dissipation period is 58.64°C, which is 105°C lower than the maximum temperature of the high-frequency transformer during the natural air cooling period, indicating that the heat dissipation performance of the fin of this parameter model is optimal.

[0098] In order to simulate the heat dissipation process of the high-frequency transformer, the wind speed at the forced air cooling inlet is set to 0.5m / s. Fig. 9As shown, the cross-sectional velocity distribution inside the high-frequency transformer can be observed. Through careful analysis and measurement, it is found that the core and winding areas near the fluid inlet position show the best heat dissipation effect. This result is determined by the flow path of the wind speed, which mainly flows from the right port of the high-frequency transformer to the left port, and the wake formed behind the heat sink further confirms the key role of the convection cooling mechanism in the heat transfer process.

[0099] Table 6 Prediction results of fin model structural parameters

[0100]

Claims

1. The optimization design method of forced air-cooled high-frequency transformer fin heat sink suitable for source-grid-load-storage integrated system is characterized by The following steps are involved: Step 1: Establish a polynomial response surface model that can reflect the temperature of the heat sink; Step 2: Establish a three-dimensional model of the fin heat sink under different combination designs and determine the maximum temperature under different combination designs; Step 3: Taking the maximum temperature of the finned heat sink as the optimization target and the selected decisive influencing factors as independent variables, the response surface method is used for experimental design. Through the residual analysis of the response surface method, the response surface model and the influence of each variable on the result are obtained, and the final response surface function is obtained; Step 4: According to the response surface model, the optimal value of each variable is obtained.

2. The optimization design method for a forced air-cooled high-frequency transformer fin heat sink suitable for a source-grid-load-storage integrated system according to claim 1 is characterized in that: In the step 1, the Box-Behnken surface response method is used to optimize the design for the heat dissipation performance of the heat sink under different structural parameter combinations; the second-order Box-Behnken response surface model multivariate quadratic equation is used to fit the functional relationship between the design parameters and the effect values; based on the second-order Box-Behnken response surface model, experimental design sampling is performed on the specified fin thickness A, fin height B, and fin spacing C design parameters to obtain sample values.

3. The optimization design method for a forced air-cooled high-frequency transformer fin heat sink suitable for a source-grid-load-storage integrated system according to claim 2 is characterized in that: The second-order Box-Behnken response surface model contains the following cross terms: Where: represents the response variable; β0, β j , β jj , β ij are the unknown coefficients, which are the coefficients of the constant term, the first-order term, the second-order non-intersecting term, and the second-order intersecting term; x i 、x j is the design variable; n is the number of design variables.

4. The optimization design method for a forced air-cooled high-frequency transformer fin heat sink suitable for a source-grid-load-storage integrated system according to claim 1 is characterized in that: In step 2, the central combination design factor level table is used to explore and verify the influence of the three factors of fin thickness A, fin height B, and fin spacing C on the maximum temperature; in this design, the experiment is divided into multiple combinations, each combination controls one factor, and simulates its influence on the result; the central combination design factor level table is determined, and the electromagnetic-temperature-fluid multi-physics field finite element method is used to calculate the maximum temperature under different combination designs when the fins are kept at a constant temperature.

5. The optimization design method for a forced air-cooled high-frequency transformer fin heat sink suitable for a source-grid-load-storage integrated system according to claim 4 is characterized in that: Determine whether the variables to be optimized in the fin heat sink have an impact on the maximum temperature and the extent of the impact, thereby determining the main factors affecting the maximum temperature; specifically: Select any one of the parameters such as fin thickness A, fin height B, fin spacing C, number of fins N, etc. one by one, select an appropriate step size and change the parameter according to the step size, and keep other variables unchanged. Establish a three-dimensional finite element model of the high-frequency transformer and the corresponding finite element model of the heat sink in the finite element simulation software Comsol, and calculate the maximum temperature under different heat sink parameters.

6. The optimization design method for a forced air-cooled high-frequency transformer fin heat sink suitable for a source-grid-load-storage integrated system according to claim 1 is characterized in that: In step 3, assuming that the response surface function to be obtained requires k experiments, the second-order Box-Behnken response surface model in step 1 can be converted into a matrix form as follows: (β0β1…β n β 11 …β nn β 12 …β nn-1 ) T +(μ1μ2…μ k ) T Where: y1 represents the maximum temperature simulation value under the first combination design; y2 represents the maximum temperature simulation value under the second combination design; y k represents the maximum temperature simulation value under the kth combination design; x 11 represents the value of the first design factor under the first combination design; x 12 represents the value of the second design factor under the first combination design; x 1n-1 represents the value of the n-1th design factor under the first combination design; x 1n It represents the value of the nth design factor under the first combination design; x 21 represents the value of the first design factor under the second combination design; x 22 represents the value of the second design factor under the second combination design; x 2n-1 represents the value of the n-1th design factor under the second combination design; x 2n Indicates the value of the nth design factor under the second combination design; x k1 represents the value of the first design factor under the kth combination design; x k2 represents the value of the second design factor under the kth combination design; x kn-1 represents the value of the n-1th design factor under the kth combination design; x kn represents the value of the nth design factor under the kth combination design; β0 is a constant term; β1...β n is the first-order coefficient corresponding to the 1st...nth factors; β 11 ......β nn is the second-order non-intersecting coefficient corresponding to the 1st...nth factors; β 12 ......β nn-1 is the coefficient of the second-order cross terms between the 1st and 2nd factors, ..., the nth and n-1th factors; T represents the matrix transpose; μ1 represents the fitting error under the first combination design; μ2 represents the fitting error under the second combination design; μ k represents the fitting error under the kth combination design; According to the extreme value problem of unbiased estimation, the above formula is used to minimize the sum of squares of errors, thus obtaining the polynomial coefficients: β=(X T X) -1 X T y Where: β represents the vector of undetermined coefficients of the polynomial; X T represents the transpose of the design variable matrix; X represents the design variable matrix; y represents the Comsol simulation results of the corresponding variables.

7. The optimization design method for a forced air-cooled high-frequency transformer fin heat sink suitable for a source-grid-load-storage integrated system according to claim 6 is characterized in that: Substitute the polynomial coefficient vector into step 1 to obtain the second-order Box-Behnken response surface model; in order to verify the applicability of the model, the model is subjected to residual analysis according to Design-Expert; in the residual normal probability plot, if the residuals are normally distributed and the residual points are evenly dispersed near a straight line, it means that the model fits well.

8. The optimization design method for a forced air-cooled high-frequency transformer fin heat sink suitable for a source-grid-load-storage integrated system according to claim 6, characterized in that: The overall efficiency predicted by the response surface model is given by the coefficient of determination R 2 Explanation, coefficient of determination R 2 It is the ratio of the regression sum of squares to the total sum of squares. The closer its value is to 1, the higher the accuracy. Its calculation formula is: Where: y is the sample value; is the sample mean; M represents the number of sample values ​​obtained through experimental design sampling; y i represents the finite element simulation results; The response variable represented by is calculated.

9. The optimization design method for a forced air-cooled high-frequency transformer fin heat sink suitable for a source-grid-load-storage integrated system according to claim 1, characterized in that: In step 3, the calculation process of the response surface method is: Step (1): Determine the sampling space by variables such as fin thickness A, fin height B, and fin spacing C, and determine the level of the experimental factor; the fin thickness A, fin height B, and fin spacing C all have three levels, namely, low level: -1; center level: 0; and high level: 1, which respectively represent the lower limit value, middle value, and upper limit value of the factor; Step (2): using the central composite design (CCD) method, the electromagnetic-temperature-fluid multi-physics field finite element analysis calculation is performed on the sampling points according to the experimental factor levels; Step (3): taking the maximum temperature of the fin heat sink as the optimization target, taking the selected decisive influencing factors as independent variables, fitting the response surface function, and performing variance analysis; Step (4): Based on the residual normal probability plot analysis, insignificant items are eliminated and significant influencing factors are determined; Step (5): Determine whether the accuracy is met; if it is met, the response surface model can be obtained; if it is not met, increase the number of terms or order of the response surface model and repeat step (1).

10. The optimization design method for a forced air-cooled high-frequency transformer fin heat sink suitable for a source-grid-load-storage integrated system according to claim 1, characterized in that: In step 4, according to the response surface model, a genetic algorithm is used to obtain the optimal value of each variable of the radiator; Taking the minimization of the maximum temperature t of the high-frequency transformer as the objective function, considering the value ranges of the fin thickness A, fin height B, and fin spacing C of the radiator, an optimization design model is established: The objective function is: The constraints are: A∈[A min ,A max ];B∈[B min ,B max ];C∈[C min ,C max ] The individual fitness evaluation function is: The genetic algorithm takes the initial solution as the starting point, and generates the next generation of population through genetic operations such as selection, crossover, and mutation. The individual fitness evaluation function is used to examine the quality of each individual in the population, and high-quality individuals are selected to enter the next generation. After multiple iterations, the optimal solution is finally obtained.

Images