Resonant cavity parameter optimization design method for energy storage bidirectional DC converter LLC circuit

Abstract

The invention relates to a resonant cavity parameter optimization design method for an energy storage bidirectional DC converter LLC circuit, and the method comprises the steps: selecting N resonant inductance values, calculating the corresponding excitation inductance value and resonant capacitance value, and calculating the corresponding resonant inductance current, excitation inductance current and resonant capacitance voltage; respectively designing a resonant inductor, an excitation inductor and a resonant capacitor, and calculating a feasibility label, cost and loss; the method comprises the following steps: integrating inductance and capacitance design data into an input data set, taking a comprehensive index calculated by using cost and loss as a first output data set, taking a feasibility label as a second output data set, and calculating maximum mutual information coefficients of each input data feature and each output data set to form a final input data set and a final output data set; and inputting the input data into the SVM for training, randomly generating input data, inputting the input data into the SVM for outputting, and screening out an optimal design scheme after removing infeasible schemes. The method is beneficial to improving the performance and efficiency of the resonant cavity parameter optimization design.

Description

Technical Field

[0001] The invention relates to the technical field of power electronics, and in particular to a resonant cavity parameter optimization design method for an energy storage bidirectional direct current converter LLC circuit. Background Art

[0002] As an important topology of DC / DC converter with resonant switching technology, bidirectional LLC DC resonant converter has the advantages of high efficiency and low loss, zero voltage switching, zero current switching and high integration, and its application is very extensive. The parameters of the internal resonant circuit will directly affect the conversion efficiency and performance of the converter. However, the resonant circuit contains more magnetic devices and components, and has more design parameters. At present, the hardware parameter design of the resonant circuit mainly relies on engineering experience to calculate the parameters and calculate the circuit model. The engineering experience method has low design cost, but for systems with more parameters and more coupling relationships, it is difficult to fully consider factors such as cost, efficiency, and performance, and the design efficiency is low, and it is difficult to obtain the optimal parameter design. Hardware parameter design can also be calculated through an algorithm model, and its parameter design efficiency is high, but the design effect is limited by the accuracy of the model itself, and the accuracy of the model is largely determined by the training data. If the quality of the data set used is not high, the performance of the design scheme output by the trained model will not be very good. Therefore, the two design methods have their own advantages and disadvantages, and there are obvious disadvantages in using only one method for parameter design. Currently, there is a solution that combines engineering experience with algorithm models to design hardware parameters. The specific method is to use engineering experience data to provide a data set for algorithm model training. The subsequent parameter design is performed by the trained algorithm model. However, this method needs to ensure the quality of the provided data set. How to improve the quality of the data set generated by engineering experience for algorithm model training is an urgent problem to be solved.

[0003] Patent CN112069758A discloses a method for optimizing the soft start parameters of LLC converters based on a simplified numerical model, including sampling the input voltage of LLC converters, setting the maximum overshoot current I during startup according to the current stress of switching devices, and r_peak , set the minimum soft switching current I during startup according to the soft switching conditions r_zvs , and use it as the initial condition for parameter iteration. Secondly, according to the simplified circuit model under the four modes, the duration of each mode and the instantaneous resonant capacitor voltage and resonant current at the start-up time are obtained through iteration to determine the adaptive start-up frequency. Again, the system stability is judged according to the conditions for the LLC converter to oscillate. If the oscillation conditions are met, the start-up frequency at this time is obtained; if the oscillation conditions are not met, the I is gradually increased. r_peak , while reducing I r_zvs, re-iterate the calculation until the oscillation condition is met, and the starting frequency is obtained. Finally, the initial starting frequency and the initial duty cycle are output, and soft-start is performed with these parameters. Patent CN115495936A discloses a parameter design method for LLC converter based on adaptive polynomial approximation, which relates to the technical field of LLC resonant converter design; first, the parameter space of the resonant cavity that only works in the PO mode is determined by numerical analysis; then the final parameter space is determined by setting constraints, and the optimization target and the proxy model of the parameter space are established based on the adaptive polynomial approximation method, and the proxy model of the parameter space is also established by the adaptive polynomial approximation method; the NSGA-II algorithm is used for multi-objective optimization to design the optimal parameters. The method of the present invention is effective and reasonable, with high efficiency. Under the worst working condition, the efficiency can reach 94.2%, and the peak efficiency is 94.93%. The above patents all disclose the optimization design of LLC converter related parameters through specific algorithm models, but none of them involve how to improve the quality of the data set used for algorithm model training, so as to further improve the performance and efficiency of the optimization design of resonant cavity parameters through the algorithm model. Summary of the invention

[0004] The object of the present invention is to provide a method for optimizing the design of resonant cavity parameters for an energy storage bidirectional DC converter LLC circuit, which method is beneficial to improving the performance and efficiency of the optimized design of resonant cavity parameters.

[0005] In order to achieve the above object, the technical solution adopted by the present invention is: a resonant cavity parameter optimization design method for an energy storage bidirectional DC converter LLC circuit, comprising:

[0006] Select N inductance values ​​as resonant inductance values, and calculate corresponding excitation inductance values ​​and resonant capacitance values ​​based on the relationship between the resonant inductance and the excitation inductance and the resonant capacitance, so as to obtain N groups of data consisting of the resonant inductance values ​​and the corresponding excitation inductance values ​​and resonant capacitance values; use the N groups of data to calculate the corresponding resonant inductance current, excitation inductance current, and resonant capacitance voltage data in turn;

[0007] Based on the obtained resonant inductance value, excitation inductance value, simulated resonant inductance current and excitation inductance current data, the resonant inductance and excitation inductance are designed in combination with the core data in the core database, and the feasibility label, cost and loss are calculated; based on the obtained resonant capacitance value and the resonant capacitance voltage obtained by simulation, the resonant capacitance is designed in combination with the capacitance data of different manufacturers, and the feasibility label, cost and loss are calculated;

[0008] Integrate the inductor and capacitor related design data into the initial input data set, use the comprehensive index data calculated using the cost and loss data as the first output data set, use the feasibility label as the second output data set, calculate the maximum mutual information coefficient between each feature of the input data set and the first output data set and the second output data set, and use the maximum mutual information coefficient to screen out the input data features with high correlation with the first output data set and the second output data set, respectively, to form two new input data sets;

[0009] The two new input data sets are integrated to obtain the final input data set, and the first output data set and the second output data set are integrated to obtain the final output data set. The final input data and the final output data are input into the SVM for training, and then a large amount of input data with the same format as the final input data is randomly generated and input into the SVM for output. After eliminating infeasible solutions, the optimal design solution is selected according to the comprehensive indicators.

[0010] Further, N inductance values ​​are selected as resonant inductance values, and the corresponding excitation inductance values ​​and resonant capacitance values ​​are calculated based on the relationship between the resonant inductance and the excitation inductance and the resonant capacitance, so as to obtain N groups of data consisting of the resonant inductance values ​​and the corresponding excitation inductance values ​​and resonant capacitance values; the corresponding resonant inductance current, excitation inductance current and resonant capacitance voltage data are calculated in turn using the N groups of data, and the specific method is as follows:

[0011] Select N inductance values ​​from the inductance database as the resonant inductance value, and the corresponding resonant capacitance value and excitation inductance value are calculated as follows:

[0012]

[0013] Lm=k×Lr (2)

[0014] Wherein, Cr is the resonant capacitance value corresponding to the resonant inductance value, fr is the resonant frequency, Lr is the resonant inductance value, Lm is the excitation inductance value corresponding to the resonant inductance value, and k is the proportional coefficient between the excitation inductance and the resonant inductance;

[0015] After calculating the N resonant capacitance values ​​and N excitation inductance values ​​corresponding to the N resonant inductance values, they are combined into N groups of data consisting of the resonant inductance value and the corresponding excitation inductance value and resonant capacitance value, and the N groups of data are input into the simulation model in turn for simulation, and the maximum current I1max and the effective value I1 flowing through the resonant inductance and the resonant capacitance are calculated, and the switching frequency f corresponding to the rated voltage output from the low-voltage DC side to the high-voltage DC side is s , the maximum voltage Ucmax at the resonant capacitor and the maximum current I2max and effective value I2 flowing through the excitation inductor.

[0016] Furthermore, based on the obtained resonant inductance value, excitation inductance value, simulated resonant inductance current and excitation inductance current data, the resonant inductance and excitation inductance are designed in combination with the core data in the core database, and the feasibility label, cost and loss are calculated. The specific method is as follows:

[0017] Select U magnetic core data from the magnetic core database, and perform resonant inductance or excitation inductance design respectively to obtain different design schemes of resonant inductance or excitation inductance:

[0018] First, the winding conductor is designed. The calculation formula for the diameter of the enameled wire conductor of the inductor winding is as follows:

[0019]

[0020] Among them, d w is the diameter of the enameled wire conductor, I3 is the effective value of the inductor current, J max is the maximum current density effective value, h ins is the thickness of the insulation layer of the winding wire;

[0021] After calculating the diameter of the enameled wire conductor, the maximum number of turns of the core is calculated based on the core data in the core database. The calculation formula is as follows:

[0022]

[0023] Among them, N max is the maximum number of turns of the magnetic core, d core is the inner diameter of the core, h ins is the thickness of the insulation layer of the winding wire, d w is the diameter of the enameled wire, S wire is the winding wire spacing;

[0024] After calculating the maximum number of turns of the magnetic core, the number of turns of the inductor winding, the cross-sectional area of ​​the winding wire and the filling factor of the inductor window are calculated. The calculation formula is as follows:

[0025]

[0026] Among them, N Lr is the number of turns of the inductor winding, L3 is the inductance value, I3max is the maximum current flowing through the inductor, ΔB is the magnetic flux swing, A e is the effective magnetic flux cross-sectional area;

[0027]

[0028] Among them, A1 is the cross-sectional area of ​​the winding wire, I3 is the effective value of the inductor current, J max is the maximum effective value of current density;

[0029]

[0030] Among them, K L is the inductor window filling factor, A w is the core window area;

[0031] After calculating the design parameters of the inductor, the feasibility label is calculated based on the design parameters. The calculation formula is as follows:

[0032]

[0033] Among them, Label is the feasibility label, N Lr is the number of turns of the inductor winding, N max K is the maximum number of turns of the core, L is the inductance window filling factor;

[0034] After obtaining the feasibility label, the inductor cost and loss of different design solutions are calculated;

[0035] The average cost of the same type of core produced by different manufacturers is calculated as the inductor cost. The calculation formula is as follows:

[0036]

[0037] Among them, F L is the inductor cost, F1, F2, ..., F n is the cost corresponding to the production of inductors by different manufacturers, and n is the total number of manufacturers of different manufacturers;

[0038] The average value of the inductance loss calculated by calculators from different manufacturers is used as the inductance loss. The calculation formula is as follows:

[0039]

[0040] Among them, P L is the inductor loss, P1, P2, ..., P n is the inductor loss calculated using calculators from different manufacturers, and n is the total number of manufacturers from different manufacturers.

[0041] Furthermore, based on the obtained resonant capacitance value and the resonant capacitance voltage obtained by simulation, the resonant capacitance is designed in combination with the capacitance data of different manufacturers, and the feasibility label and cost and loss are calculated. The specific method is as follows:

[0042] Select T capacitor data from the capacitor database and design the resonant capacitors respectively to obtain different design schemes of the resonant capacitors:

[0043] Calculate the number of different capacitors in series, the calculation formula is as follows:

[0044]

[0045] Among them, n2 is the number of capacitors required to be connected in series, C i is the capacitance value corresponding to different capacitors, C r is the capacitance value of the resonant capacitor, and round() is the rounding function;

[0046] After calculating the number of different capacitors in series, the feasibility label is calculated, and the calculation formula is as follows:

[0047]

[0048] Where n is the number of capacitors that need to be connected in series, N cmax is the maximum number of stacks, U ci is the voltage value that a single capacitor can withstand, U cmax The maximum voltage that the total capacitor can withstand, Label is the feasibility label;

[0049] After obtaining the feasibility label, the capacitor cost is calculated, and the calculation formula is as follows:

[0050] F c =F ci ×n i (13)

[0051] Among them, F c is the total price required for different capacitors, F ci The price of a single capacitor with different capacitance values, n i The number of capacitors required for different capacitance values;

[0052] After calculating the cost of different capacitors, the equivalent series resistance and loss are calculated. The calculation formula is as follows:

[0053]

[0054] Among them, ESR is the equivalent series resistance of the capacitor, tanδ is the loss angle of the capacitor, and f s is the switching frequency corresponding to the output of rated voltage from the low-voltage DC side to the high-voltage DC side, and C is the capacitance value of the capacitor;

[0055] P C =I C 2 ×ESR (15)

[0056] Among them, I C is the effective value of the current flowing through the capacitor, P C is the loss of the capacitor.

[0057] Furthermore, the design data related to the inductor and capacitor are integrated into the initial input data set, the comprehensive index data calculated using the cost and loss data is used as the first output data set, and the feasibility label is used as the second output data set. The maximum mutual information coefficients of each feature of the input data set and the first output data set and the second output data set are calculated respectively. The input data features with high correlation with the first output data set and the second output data set are screened out using the maximum mutual information coefficient to form two new input data sets respectively. The specific method is as follows:

[0058] Select the design data related to inductance and capacitance, including resonant inductance value, resonant capacitance value, excitation inductance value, magnetic core inner diameter, maximum current density effective value, magnetic core effective flux cross-sectional area, magnetic core window area, and actual capacitance value, and integrate them into the initial input data set;

[0059] After obtaining the initial input data set, the total cost and total loss data are used to calculate the comprehensive index, and it is used as the first output data set. The feasibility labels corresponding to each input data are used as the second output data set. The calculation formulas for total cost, total loss and comprehensive index are as follows:

[0060] F i总 =F iLr +F iLm +F iCr (16)

[0061] P i总 =P iLr +P iLm +P iCr (17)

[0062]

[0063] Among them, t i is the comprehensive index of cost and loss of the i-th design solution, F i总 is the total design cost of the ith design solution, F max is the maximum design cost among all design options, P i总 is the total loss of the ith design solution, P max is the maximum loss among all design solutions, F iLr The total cost of the resonant inductor designed for the ith design, F iLm The total cost of the magnetizing inductance designed for the ith design, F iCr The total cost of the resonant capacitor designed for the ith design, P iLr is the total loss of the resonant inductor, P iLm is the total loss of the excitation inductance, P iCr is the total loss of the resonant capacitor;

[0064] After obtaining the first output data set and the second output data set, calculating the maximum mutual information coefficient between each feature of the input data set and the first output data set and the second output data set;

[0065] After calculating the maximum mutual information coefficient between each feature and the output data set, the features in the input data set are sorted according to the maximum mutual information coefficient, and some features with large maximum mutual information coefficients are selected as the features of the new input data set according to the required data dimension.

[0066] Furthermore, the calculation process of the maximum mutual information coefficient between the features in the input dataset and a single output dataset is as follows:

[0067] (1) Selecting data of one feature in the input data set and the output data set in turn to form a two-dimensional data set, and plotting them in the form of coordinate points in a two-dimensional coordinate plane;

[0068] (2) The dividing lines of the divided areas are calculated according to the precision traversal, where the calculation formulas for the dividing lines of the x-axis and y-axis are as follows:

[0069] x 分割线 = k × d x (19)

[0070] y 分割线 = k × d y (20)

[0071] Among them, x 分割线 is the x-coordinate corresponding to the current x-axis dividing line, y 分割线 is the y coordinate corresponding to the current y-axis dividing line, k is the number of current traversals, d x The step size used to traverse the x-axis, d y The step size used to traverse the y axis;

[0072] (3) After the segmentation line is calculated, it is segmented into four regions: "upper left", "lower left", "upper right", and "lower right". At this time, the coordinate points in different regions are counted respectively, and the probability of the coordinate points in each region is calculated. The calculation formula is as follows:

[0073]

[0074] Among them, p is the probability of the coordinate point appearing in the area, n i is the number of coordinate points in the area, n 总 is the total number of coordinate points;

[0075] (4) After calculating the probability of the coordinate point in each area, the mutual information value between the feature and the output data set under the current traversal is calculated according to the mutual information calculation formula. The calculation formula is as follows:

[0076]

[0077] Among them, I(x, y) is the mutual information value of the two variables x and y, p(X=x i , Y=y j ) is in (x i ,y j ) region, p(x=x i ) is X in x i The probability of the region, p(Y = y j ) is Y in y j The probability of the region; when x i When the subscript i=1, it represents the left side, and when i=2, it represents the right side; when y j When the subscript j=1, it represents the upper side, and when j=2, it represents the lower side;

[0078] (5) After changing the segmentation line by traversing and calculating the mutual information value of each traversal feature and the output data set, when the segmentation line meets the traversal condition, stop traversing and calculate the maximum mutual information coefficient. The calculation formula is as follows:

[0079]

[0080] Among them, mic(x, y) is the maximum mutual information coefficient between the feature and the output data set, max() is the maximum value function, I(x, y) is the mutual information coefficient between the feature and the output data set, a is the number of regions divided in the x direction, and b is the number of regions divided in the y direction.

[0081] Furthermore, the new input data set obtained has two rows, which are respectively a data set after the features are filtered by the input data set and the first output data set, and a data set after the features are filtered by the input data set and the second output data set. These two rows of data are combined according to the upper and lower rows of data sequence numbers to obtain the final input data set; the first output data set and the second output data set are combined according to the upper and lower rows of data sequence numbers to obtain the final output data set.

[0082] Furthermore, the final input data and the final output data are input into SVM for model training; after the model training is completed, M input data are randomly generated according to the data format of the final input data, wherein each input data represents a design scheme; these randomly generated input data are input into SVM to make it output the corresponding comprehensive index and the feasibility label of each design scheme; the input data, i.e., the design scheme, whose output feasibility label is 0 is eliminated, the comprehensive indexes corresponding to the remaining input data are sorted, and the input data, i.e., the design scheme, with the largest comprehensive index is selected as the optimal design scheme after optimization.

[0083] Compared with the prior art, the present invention has the following beneficial effects: the method adds the maximum mutual information coefficient to perform feature screening on the input data obtained by engineering experience based on the combination of engineering experience and algorithm model design, thereby providing a higher quality data set for intelligent algorithm training, so that the trained algorithm model can perform parameter optimization design more efficiently. The parameters optimized by this method can reduce the resonant cavity power consumption of the LLC circuit of the energy storage bidirectional DC converter, improve the energy conversion efficiency of the energy storage bidirectional DC converter, and reduce the hardware cost. BRIEF DESCRIPTION OF THE DRAWINGS

[0084] In order to more clearly illustrate the technical solutions of the exemplary embodiments of the present invention, the drawings required for use in the embodiments will be briefly introduced below. It should be understood that the following drawings only show certain embodiments of the present invention and therefore should not be regarded as limiting the scope. For ordinary technicians in this field, other related drawings can be obtained based on these drawings without paying creative work.

[0085] Figure 1 A flowchart of the implementation of the resonant cavity parameter optimization design method for the energy storage bidirectional DC converter LLC circuit provided by the embodiment of the present invention;

[0086] Figure 2 Schematic diagram of a circuit of a bidirectional LLC resonant converter according to an embodiment of the present invention;

[0087] Figure 3 A flowchart for obtaining test data in an embodiment of the present invention;

[0088] Figure 4 This is a flow chart of inductor design in an embodiment of the present invention;

[0089] Figure 5 A flow chart of capacitor design in an embodiment of the present invention;

[0090] Figure 6 A flowchart of screening initial data set features in an embodiment of the present invention;

[0091] Figure 7 is a schematic diagram of an initial input data set, a first output data set, and a second output data set in an embodiment of the present invention;

[0092] Figure 8 A schematic diagram of the area division process for calculating the maximum mutual information coefficient in an embodiment of the present invention;

[0093] Fig. 9 This is a flow chart of using SVM to select the optimal design solution in an embodiment of the present invention;

[0094] Fig.10Schematic diagram of the final input data set and the final output data set in an embodiment of the present invention. DETAILED DESCRIPTION

[0095] The core of the present invention is to provide a resonant cavity parameter optimization design method for an energy storage bidirectional DC converter LLC circuit. The initial input data set and the first output data set and the second output data set of the resonant cavity design are generated by using an engineering experience method. The maximum mutual information coefficient is used to screen out the features related to the two output data sets in the initial input data set and form two new input data sets. The two new input data sets in the previous step are integrated into the final input data set, and the two output data sets are integrated into the final output data set. The final input data set and the final output data set are put into SVM for training, and the trained SVM is used to optimally screen a large number of randomly generated design schemes. The design method introduces the maximum mutual information coefficient on the basis of combining engineering experience and intelligent algorithm design to perform feature screening on the input data obtained using engineering experience, providing a higher quality data set for intelligent algorithm training, and the trained algorithm model can perform parameter optimization design more efficiently. The parameters optimized by this method can reduce the power consumption of the resonant cavity of the energy storage bidirectional DC converter LLC circuit, improve the energy conversion efficiency of the energy storage bidirectional DC converter, and reduce hardware costs.

[0096] In order to make the purpose, technical solution and advantages of the embodiments of the present invention clearer, the technical solution in the embodiments of the present invention will be clearly and completely described below in conjunction with the drawings in the embodiments of the present invention. Obviously, the described embodiments are part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0097] The existing resonant converter resonant cavity design methods mainly include engineering experience design method and algorithm design method. Engineering experience design mainly uses the professional knowledge of engineers to calculate the relevant parameters of the resonant cavity in the resonant converter, and selects appropriate magnetic components and capacitors based on the calculated parameters. The engineering experience design method is not efficient when facing optimization designs with many parameters and high complexity. The algorithm design method mainly calculates relevant data by designing algorithms and using computers as carriers, which can process optimization designs with many parameters and high complexity more efficiently. However, the algorithm design method depends on the accuracy of the model. If the model accuracy is not high, the performance of the designed resonant cavity is not high. At present, there is an idea of ​​combining the two design methods for design, that is, using engineering experience to provide a training data set for the model, and then using the trained model for design. However, the quality of the training data set generated by engineering experience is not necessarily high, and the accuracy of the trained model cannot be guaranteed.

[0098] Therefore, the present invention provides a resonant cavity parameter optimization design method for an energy storage bidirectional DC converter LLC circuit. By preliminarily setting N resonant inductors, resonant capacitors and excitation inductor values ​​and performing simulation runs, relevant data for subsequent design of resonant inductors, resonant capacitors and excitation inductors are obtained; by using the previously obtained inductor test data and the core data manual to design resonant inductors and excitation inductors, and performing feasibility labels and cost loss calculations, a part of the data is provided for the initial input data set, the first output data set and the second output data set; by using the previously obtained capacitor test data and the capacitor data manual to design resonant capacitors, and performing feasibility labels and cost loss calculations, another part of the data is provided for the initial input data set, the first output data set and the second output data set; after the resonant inductor, resonant capacitor and excitation inductor are designed, the comprehensive index of cost and loss is calculated, the data commonly used in the design is selected as the initial data set, the comprehensive index is selected as the first output data set, and the feasibility label is selected as the second output data set, and the features with a large correlation between the initial input data set and the output data set are screened according to the maximum mutual information coefficient between the initial data set and the first output data set and the second output data set, forming two new input data sets; finally, the two new input data sets obtained in the previous step are integrated into the final input data set, and the two output data sets are integrated into the final data set. The final input data set and the final output data set are used to train the SVM model. After the SVM model training is completed, a large number of input data with the same format as the final input data, i.e., design schemes, are randomly generated and put into the SVM model to output comprehensive indicators and feasibility labels. After eliminating unfeasible design schemes, the optimal results of feasible design schemes are screened out according to the comprehensive indicators of cost loss. Based on the combination of engineering experience and intelligent algorithm design, this method adds the maximum mutual information coefficient to perform feature screening on the input data obtained using engineering experience, thereby providing a higher quality data set for intelligent algorithm training, so that the trained algorithm model can perform parameter optimization design more efficiently. The parameters optimized by this method can reduce the resonant cavity power consumption of the LLC circuit of the energy storage bidirectional DC converter, improve the power conversion efficiency of the energy storage bidirectional DC converter, and reduce the hardware cost.

[0099] Figure 1 is a flow chart for implementing the method of this embodiment; Figure 2 It is a schematic diagram of a bidirectional LLC resonant converter circuit; Figure 3 To obtain the test data flow chart; Figure 4 Design a flow chart for the inductor; Figure 5 Design a flow chart for capacitors; Figure 6 Flowchart for screening initial dataset features; Figure 7 is a schematic diagram of an initial input data set, a first output data set, and a second output data set; Figure 8Schematic diagram of the area division process for calculating the maximum mutual information coefficient; Fig. 9 Flowchart for using SVM to select the optimal design solution; Fig.10 Schematic diagram of the final input data set and the final output data set.

[0100] like Figure 1 As shown, this embodiment provides a resonant cavity parameter optimization design method for an energy storage bidirectional DC converter LLC circuit, including:

[0101] S1. Select N inductance values ​​as resonant inductance values, and calculate the corresponding excitation inductance values ​​and resonant capacitance values ​​based on the relationship between the resonant inductance and the excitation inductance and the resonant capacitance, and obtain N groups of data consisting of the resonant inductance values ​​and the corresponding excitation inductance values ​​and resonant capacitance values; input the N groups of data into the simulation model in turn, and calculate the corresponding resonant inductance current, excitation inductance current, and resonant capacitance voltage data. The specific implementation method is as follows.

[0102] Select N inductance values ​​from the inductance database as the resonant inductance value, and the corresponding resonant capacitance value and excitation inductance value are calculated as follows:

[0103]

[0104] Lm=k×Lr (2)

[0105] Among them, Cr is the resonant capacitance value corresponding to the resonant inductance value, fr is the resonant frequency (set according to demand), Lr is the resonant inductance value, Lm is the excitation inductance value corresponding to the resonant inductance value, and k is the proportional coefficient between the excitation inductance and the resonant inductance (generally 3-7, here 5 is selected as the proportional coefficient).

[0106] After calculating the N resonant capacitance values ​​and N excitation inductance values ​​corresponding to the N resonant inductance values, they are combined into N groups of data consisting of the resonant inductance value and the corresponding excitation inductance value and resonant capacitance value, and the N groups of data are input into the simulation model in turn for simulation, and the maximum current I1max and the effective value I1 flowing through the resonant inductance and the resonant capacitance are calculated, and the switching frequency f corresponding to the rated voltage output from the low-voltage DC side to the high-voltage DC side is s , the maximum voltage Ucmax at the resonant capacitor and the maximum current I2max and effective value I2 flowing through the excitation inductor.

[0107] This step uses the selected N resonant inductor data to generate resonant capacitor and excitation inductor values, and puts them into the simulation model to obtain data such as resonant inductor current, excitation inductor current, resonant capacitor voltage, etc. The purpose is to use the relevant data obtained in this step to design the resonant inductor, resonant capacitor and excitation inductor in the next two steps.

[0108] S2. Based on the obtained resonant inductance value, excitation inductance value, simulated resonant inductance current and excitation inductance current data, combined with the core data in the core database, the resonant inductance and excitation inductance are designed, and the feasibility label, cost and loss are calculated; based on the obtained resonant capacitance value and the simulated resonant capacitance voltage, combined with the capacitance data of different manufacturers, the resonant capacitance is designed, and the feasibility label, cost and loss are calculated. The specific implementation method is as follows.

[0109] U magnetic core data are selected from the magnetic core database, and resonant inductance or excitation inductance are designed respectively to obtain different design schemes of the resonant inductance or excitation inductance. It should be noted that both the resonant inductance and the excitation inductance can be designed using the following method.

[0110] First, use the data obtained above to design the winding wire. The calculation formula for the diameter of the enameled wire conductor of the inductor winding is as follows:

[0111]

[0112] Among them, d w is the diameter of the enameled wire conductor, I3 is the effective value of the inductor current, J max is the maximum current density effective value (set according to requirements), h ins is the thickness of the insulation layer of the winding wire.

[0113] After calculating the diameter of the enameled wire conductor, the maximum number of turns of the core is calculated based on the core data in the core database. The calculation formula is as follows:

[0114]

[0115] Among them, N max is the maximum number of turns of the magnetic core, d core is the inner diameter of the core, h ins is the thickness of the insulation layer of the winding wire, d w is the diameter of the enameled wire, S wire The winding wire spacing (the wire spacing is set according to requirements).

[0116] After calculating the maximum number of turns of the magnetic core, the number of turns of the inductor winding, the cross-sectional area of ​​the winding wire and the filling factor of the inductor window are calculated. The calculation formula is as follows:

[0117]

[0118] Among them, N Lr is the number of turns of the inductor winding, L3 is the inductance value, I3max is the maximum current flowing through the inductor, ΔB is the magnetic flux swing, A e is the effective magnetic flux cross-sectional area.

[0119]

[0120] Among them, A1 is the cross-sectional area of ​​the winding wire, I3 is the effective value of the inductor current, J max is the maximum effective value of current density;

[0121]

[0122] Among them, K L is the inductor window filling factor, A w is the core window area.

[0123] After calculating the design parameters of the inductor, the feasibility label is calculated based on the design parameters. The calculation formula is as follows:

[0124]

[0125] Among them, Label is the feasibility label, N Lr is the number of turns of the inductor winding, N max K is the maximum number of turns of the core, L is the inductor window filling factor.

[0126] The judgment logic of the above formula is as follows: when the required number of inductor winding turns exceeds the maximum number of winding turns that the magnetic core can withstand, the design scheme cannot be implemented; the window filling coefficient is generally in the range of 0.3-0.4 based on experience. If the window filling coefficient exceeds the range of 0.3-0.4, the design scheme cannot be implemented; for the design scheme that cannot be implemented, it is labeled with a value of 0, and the design scheme that can be implemented is labeled with a value of 1 to distinguish it.

[0127] After obtaining the feasibility label, the inductor cost and loss of different design solutions are calculated;

[0128] Since the same type of cores produced by different manufacturers are slightly different, the average cost of the same type of cores produced by different manufacturers is calculated as the inductor cost. The calculation formula is as follows:

[0129]

[0130] Among them, F L is the inductor cost, F1, F2, ..., F n is the cost corresponding to the production of inductors by different manufacturers, and n is the total number of manufacturers of different manufacturers;

[0131] After calculating the cost, we start to calculate the inductor loss. For the resonant converter, the total inductor loss includes the coil DC loss, coil AC loss and core loss. Among them, the coil AC loss and core loss are more difficult to calculate. Different manufacturers provide their own core inductor loss calculators for the convenience of design. You only need to input the inductance value and other data to output the estimated inductor loss. Therefore, the average value of the inductor loss calculated by the calculators of different manufacturers is used as the inductor loss. The calculation formula is as follows:

[0132]

[0133] Among them, P L is the inductor loss, P1, P2, ..., P n is the inductor loss calculated using calculators from different manufacturers, and n is the total number of manufacturers from different manufacturers.

[0134] According to the obtained resonant capacitor value and the resonant capacitor voltage obtained by simulation, the resonant capacitor is designed in combination with the capacitance data of different manufacturers, and the feasibility label, cost and loss are calculated. The specific method is as follows:

[0135] T capacitor data are selected from the capacitor database, and resonant capacitor designs are performed respectively to obtain different design schemes of the resonant capacitor.

[0136] Calculate the number of different capacitors in series, the calculation formula is as follows:

[0137]

[0138] Among them, n2 is the number of capacitors required to be connected in series, C i is the capacitance value corresponding to different capacitors, C r is the capacitance value of the resonant capacitor, and round() is the rounding function.

[0139] After calculating the number of different capacitors in series, the feasibility label is calculated, and the calculation formula is as follows:

[0140]

[0141] Where n is the number of capacitors that need to be connected in series, N cmax is the maximum number of stacks (usually set to 5), U ci is the voltage value that a single capacitor can withstand, U cmax The maximum voltage that the total capacitor can withstand. Label is the feasibility label.

[0142] After obtaining the feasibility label, the capacitor cost is calculated, and the calculation formula is as follows:

[0143] F c =Fci ×n i (13)

[0144] Among them, F c is the total price required for different capacitors, Fc i The price of a single capacitor with different capacitance values, n i The number of capacitors required for different capacitance values.

[0145] After calculating the cost of different capacitors, the equivalent series resistance and loss are calculated. The calculation formula is as follows:

[0146]

[0147] Among them, ESR is the equivalent series resistance of the capacitor, tanδ is the loss angle of the capacitor, and f s is the switching frequency corresponding to the output of rated voltage from the low-voltage DC side to the high-voltage DC side, and C is the capacitance of the capacitor.

[0148] P C =I C 2 ×ESR (15)

[0149] Among them, I C is the effective value of the current flowing through the capacitor, P C is the loss of the capacitor.

[0150] Based on the previous step, this step aims to use the previously obtained inductance test data to design the resonant inductance and excitation inductance, and then calculate the feasibility label, cost, and power consumption to provide a part of the data set for the initial input data set, the first output data set, and the second output data set.

[0151] This step also aims to use the previously obtained capacitor test data to design the resonant capacitor, and then calculate the feasibility label, cost, and power consumption to provide another part of the data set for the initial input data set, the first output data set, and the second output data set.

[0152] S3. Integrate the design data related to inductance and capacitance into the initial input data set, use the comprehensive index data calculated using the cost and loss data as the first output data set, use the feasibility label as the second output data set, calculate the maximum mutual information coefficient between each feature of the input data set and the first output data set and the second output data set, and use the maximum mutual information coefficient to screen out the input data features with high correlation with the first output data set and the second output data set, and form two new input data sets respectively. The specific implementation method is as follows.

[0153] Select the inductor and capacitor related design data, including: resonant inductance value, resonant capacitance value, excitation inductance value, core inner diameter, maximum current density effective value, core effective flux cross-sectional area, core window area, actual capacitance value, etc., and integrate them into the initial input data set.

[0154] After obtaining the initial input data set, the total cost and total loss data are used to calculate the comprehensive index, and it is used as the first output data set. The feasibility labels corresponding to each input data are used as the second output data set. The calculation formulas for total cost, total loss and comprehensive index are as follows:

[0155] F i总 =F iLr +F iLm +F iCr (16)

[0156] P i总 =P iLr +P iLm +P iCr (17)

[0157]

[0158] Among them, t i is the comprehensive index of cost and loss of the i-th design solution, F i总 is the total design cost of the ith design solution, F max is the maximum design cost among all design options, P i总 is the total loss of the ith design solution, P max is the maximum loss among all design solutions, F iLr The total cost of the resonant inductor designed for the ith design, F iLm The total cost of the magnetizing inductance designed for the ith design, F iCr The total cost of the resonant capacitor designed for the ith design, P iLr is the total loss of the resonant inductor, P iLm is the total loss of the excitation inductance, P iCr is the total loss of the resonant capacitor.

[0159] After the first output data set and the second output data set are obtained, the maximum mutual information coefficients between each feature of the input data set and the first output data set and the second output data set are calculated.

[0160] The calculation process of the maximum mutual information coefficient between the features in the input dataset and a single output dataset is as follows:

[0161] (1) The data of one feature in the input data set and the output data set are selected in sequence to form a two-dimensional data set, and are plotted in the form of coordinate points in a two-dimensional coordinate plane.

[0162] (2) The dividing lines of the divided areas are calculated according to the precision traversal, where the calculation formulas for the dividing lines of the x-axis and y-axis are as follows:

[0163] x 分割线 = k × d x (19)

[0164] y 分割线 = k × d y (20)

[0165] Among them, x 分割线 is the x-coordinate corresponding to the current x-axis dividing line, y 分割线 is the y coordinate corresponding to the current y-axis dividing line, k is the number of current traversals, d x The step size used to traverse the x-axis, d y The step size to use for traversing the y-axis.

[0166] (3) After the segmentation line is calculated, it is segmented into four regions: "upper left", "lower left", "upper right", and "lower right". At this time, X has two values: "left" and "right", and Y has two values: "up" and "down". At this time, the coordinate points in different regions are counted respectively, and the probability of the coordinate point in each region is calculated. The calculation formula is as follows:

[0167]

[0168] Among them, p is the probability of the coordinate point appearing in the area, n i is the number of coordinate points in the area, n 总 is the total number of coordinate points.

[0169] (4) After calculating the probability of the coordinate point in each area, the mutual information value between the feature and the output data set under the current traversal is calculated according to the mutual information calculation formula. The calculation formula is as follows:

[0170]

[0171] Among them, I(x, y) is the mutual information value of the two variables x and y, p(X=x i , Y=y j ) is in (x i ,y j ) region, p(x=x i ) is X in x i The probability of the region, p(Y = y j ) is Y in y j The probability of the region; when x i When the subscript i=1, it represents the left side, and when i=2, it represents the right side; when y jSubscript j=1 represents the upper side, and subscript j=2 represents the lower side.

[0172] (5) After changing the segmentation line by traversing and calculating the mutual information value of each traversal feature and the output data set, when the segmentation line meets the traversal condition, stop traversing and calculate the maximum mutual information coefficient. The calculation formula is as follows:

[0173]

[0174] Among them, mic(x, y) is the maximum mutual information coefficient between the feature and the output data set, max() is the maximum value function, I(x, y) is the mutual information coefficient between the feature and the output data set, a is the number of regions divided in the x direction, and b is the number of regions divided in the y direction.

[0175] After calculating the maximum mutual information coefficient between each feature and the output data set, the features in the input data set are sorted according to the maximum mutual information coefficient, and some features with large maximum mutual information coefficients are selected as the features of the new input data set according to the required data dimension.

[0176] This step mainly uses the maximum mutual information coefficient between each feature of the initial input data set and the first output data set and the second output data set to screen out the features in the initial input data set that are highly correlated with the first output data set and the second output data set, aiming to provide a higher quality data set for the next step of SVM training and improve the accuracy of the trained model.

[0177] S4, integrate the two new input data sets to obtain the final input data set, integrate the first output data set and the second output data set to obtain the final output data set, and input the final input data and the final output data into the SVM for training, and then randomly generate a large amount of input data with the same format as the final input data, input it into the SVM for output, and after eliminating infeasible solutions, select the optimal design solution according to the comprehensive indicators. The specific implementation method is as follows.

[0178] The new input data set obtained has two rows, which are the data set after the input data set and the first output data set filter features and the data set after the input data set and the second output data set filter features. These two rows of data are combined according to the upper and lower rows of data sequence numbers to obtain the final input data set; the first output data set and the second output data set are combined according to the upper and lower rows of data sequence numbers to obtain the final output data set.

[0179] The final input data and the final output data are input into SVM for model training. After the model training is completed, M input data are randomly generated according to the data format of the final input data, where each input data represents a design scheme. These randomly generated input data are input into SVM to make it output the corresponding comprehensive index and the feasibility label of each design scheme. The input data, i.e., the design scheme, whose output feasibility label is 0 is eliminated, and the comprehensive indexes corresponding to the remaining input data are sorted, and the input data, i.e., the design scheme, with the largest comprehensive index is selected as the optimal design scheme after optimization.

[0180] This step mainly uses the two new input data sets and two output data sets in the previous step to integrate to obtain the final input data set and the final output data set. The final input data set and the final output data set are put into SVM for training. After the training is completed, input data in the same format as the final input data, i.e., the design scheme, is randomly generated and put into SVM for comprehensive index and feasibility label output. After the feasibility label is used to filter out the infeasible scheme, the optimal input data in the remaining data is selected according to the comprehensive index, i.e., the optimized design scheme. The purpose is to optimize the design of the resonant cavity through the algorithm model to achieve the purpose of reducing the design cost and power consumption of the resonant cavity.

[0181] In summary, this embodiment obtains relevant data for subsequent inductor and capacitor design by preliminarily selecting N resonant inductors, resonant capacitors and excitation inductance values ​​and performing simulation operations; uses the previously obtained inductor test data and the core data manual to design the resonant inductor and excitation inductor, and performs feasibility labels and cost loss calculations, providing a portion of the data set for the initial input data set, the first output data set and the second output data set; uses the previously obtained capacitor test data and the capacitor data manual to design the resonant capacitor, and performs feasibility labels and cost loss calculations, providing another portion of the data set for the initial input data set, the first output data set and the second output data set; calculates a comprehensive index, integrates the data related to the inductor and capacitor designs in the first two steps into an initial input data set, uses the comprehensive index as the first output data set, and uses the feasibility label as the second output data set; and screens out features with a high correlation between the initial input data set and the output data set based on the maximum mutual information coefficient between the initial data set and the first output data set and the second output data set, forming two new input data sets, preparing for the final input data set in the next step; finally, integrates the two new input data sets obtained in the previous step into the final input data set, and integrates the two output data sets into the final data set. The final input data set and the final output data set are used to train the SVM model. After the SVM model training is completed, a large number of input data with the same format as the final input data, i.e., design schemes, are randomly generated and put into the SVM model to output comprehensive indicators and feasibility labels. After eliminating the unfeasible design schemes, the optimal results of the feasible design schemes are screened out according to the comprehensive indicators of cost loss. This design method combines the engineering experience design method with the algorithm design method and uses the maximum mutual information coefficient to perform feature screening on the input data obtained using engineering experience, thereby providing a higher quality data set for intelligent algorithm training, so that the trained algorithm model can perform parameter optimization design more efficiently. The parameters optimized by this method can reduce the resonant cavity power consumption of the LLC circuit of the energy storage bidirectional DC converter, improve the energy conversion efficiency of the energy storage bidirectional DC converter, and reduce the hardware cost.

[0182] The above is only a preferred embodiment of the present invention, and does not limit the present invention in other forms. Any technician familiar with the profession may use the above disclosed technical content to change or modify it into an equivalent embodiment with equivalent changes. However, any simple modification, equivalent change and modification made to the above embodiment according to the technical essence of the present invention without departing from the technical solution of the present invention still belongs to the protection scope of the technical solution of the present invention.

Claims

1. A resonant cavity parameter optimization design method for an energy storage bidirectional DC converter LLC circuit, characterized in that: include: Select N inductance values ​​as resonant inductance values, and calculate corresponding excitation inductance values ​​and resonant capacitance values ​​based on the relationship between the resonant inductance and the excitation inductance and the resonant capacitance, so as to obtain N groups of data consisting of the resonant inductance values ​​and the corresponding excitation inductance values ​​and resonant capacitance values; use the N groups of data to calculate the corresponding resonant inductance current, excitation inductance current, and resonant capacitance voltage data in turn; Based on the obtained resonant inductance value, excitation inductance value, simulated resonant inductance current and excitation inductance current data, the resonant inductance and excitation inductance are designed in combination with the core data in the core database, and the feasibility label, cost and loss are calculated; based on the obtained resonant capacitance value and the resonant capacitance voltage obtained by simulation, the resonant capacitance is designed in combination with the capacitance data of different manufacturers, and the feasibility label, cost and loss are calculated; Integrate the inductor and capacitor related design data into the initial input data set, use the comprehensive index data calculated using the cost and loss data as the first output data set, use the feasibility label as the second output data set, calculate the maximum mutual information coefficient between each feature of the input data set and the first output data set and the second output data set, and use the maximum mutual information coefficient to screen out the input data features with high correlation with the first output data set and the second output data set, respectively, to form two new input data sets; The two new input data sets are integrated to obtain the final input data set, and the first output data set and the second output data set are integrated to obtain the final output data set. The final input data and the final output data are input into the SVM for training, and then a large amount of input data with the same format as the final input data is randomly generated and input into the SVM for output. After eliminating infeasible solutions, the optimal design solution is selected according to the comprehensive indicators.

2. The resonant cavity parameter optimization design method for the energy storage bidirectional DC converter LLC circuit according to claim 1 is characterized in that: Select N inductance values ​​as resonant inductance values, and calculate the corresponding excitation inductance values ​​and resonant capacitance values ​​based on the relationship between the resonant inductance and the excitation inductance and the resonant capacitance, and obtain N groups of data consisting of the resonant inductance values ​​and the corresponding excitation inductance values ​​and resonant capacitance values; use the N groups of data to calculate the corresponding resonant inductance current, excitation inductance current, and resonant capacitance voltage data in turn, and the specific method is as follows: Select N inductance values ​​from the inductance database as the resonant inductance value, and the corresponding resonant capacitance value and excitation inductance value are calculated as follows: Lm=k×Lr (2) Wherein, Cr is the resonant capacitance value corresponding to the resonant inductance value, fr is the resonant frequency, Lr is the resonant inductance value, Lm is the excitation inductance value corresponding to the resonant inductance value, and k is the proportional coefficient between the excitation inductance and the resonant inductance; After calculating the N resonant capacitance values ​​and N excitation inductance values ​​corresponding to the N resonant inductance values, they are combined into N groups of data consisting of the resonant inductance value and the corresponding excitation inductance value and the resonant capacitance value, and the N groups of data are input into the simulation model in turn for simulation, and the maximum current I1max and the effective value I1 flowing through the resonant inductance and the resonant capacitor, the switching frequency fs corresponding to the rated voltage output from the low-voltage DC side to the high-voltage DC side, the maximum voltage Ucmax borne at the resonant capacitor, and the maximum current I2max and the effective value I2 flowing through the excitation inductance are calculated.

3. The resonant cavity parameter optimization design method for the energy storage bidirectional DC converter LLC circuit according to claim 1 is characterized in that: According to the obtained resonant inductance value, excitation inductance value, simulated resonant inductance current and excitation inductance current data, the resonant inductance and excitation inductance are designed in combination with the core data in the core database, and the feasibility label, cost and loss are calculated. The specific method is as follows: Select U magnetic core data from the magnetic core database, and perform resonant inductance or excitation inductance design respectively to obtain different design schemes of resonant inductance or excitation inductance: First, the winding conductor is designed. The calculation formula for the diameter of the enameled wire conductor of the inductor winding is as follows: Among them, d w is the diameter of the enameled wire conductor, I3 is the effective value of the inductor current, J max is the maximum current density effective value, h ins is the thickness of the insulation layer of the winding wire; After calculating the diameter of the enameled wire conductor, the maximum number of turns of the core is calculated based on the core data in the core database. The calculation formula is as follows: Among them, N max is the maximum number of turns of the magnetic core, d core is the inner diameter of the core, h ins is the thickness of the insulation layer of the winding wire, d w is the diameter of the enameled wire, S wire is the winding wire spacing; After calculating the maximum number of turns of the magnetic core, the number of turns of the inductor winding, the cross-sectional area of ​​the winding wire and the filling factor of the inductor window are calculated. The calculation formula is as follows: Among them, N Lr is the number of turns of the inductor winding, L3 is the inductance value, I3max is the maximum current flowing through the inductor, ΔB is the magnetic flux swing, A e is the effective magnetic flux cross-sectional area; Among them, A1 is the cross-sectional area of ​​the winding wire, I3 is the effective value of the inductor current, J max is the maximum effective value of current density; Among them, K L is the inductor window filling factor, A w is the core window area; After calculating the design parameters of the inductor, the feasibility label is calculated based on the design parameters. The calculation formula is as follows: Among them, Label is the feasibility label, N Lr is the number of turns of the inductor winding, N max K is the maximum number of turns of the core, L is the inductance window filling factor; After obtaining the feasibility label, calculate the inductor cost and loss of different design solutions; The inductance cost is calculated by calculating the average cost of the same type of core produced by different manufacturers. The calculation formula is as follows: Among them, F L is the inductor cost, F1, F2, ..., F n is the cost corresponding to the production of inductors by different manufacturers, and n is the total number of manufacturers of different manufacturers; The average value of the inductance loss calculated by calculators from different manufacturers is used as the inductance loss. The calculation formula is as follows: Among them, P L is the inductor loss, P1, P2, ..., P n is the inductor loss calculated using calculators from different manufacturers, and n is the total number of manufacturers from different manufacturers.

4. The resonant cavity parameter optimization design method for the energy storage bidirectional DC converter LLC circuit according to claim 1 is characterized in that: According to the obtained resonant capacitor value and the resonant capacitor voltage obtained by simulation, the resonant capacitor is designed in combination with the capacitance data of different manufacturers, and the feasibility label, cost and loss are calculated. The specific method is as follows: Select T capacitor data from the capacitor database and design the resonant capacitors respectively to obtain different design schemes of the resonant capacitors: Calculate the number of different capacitors in series, the calculation formula is as follows: Among them, n2 is the number of capacitors required to be connected in series, C i is the capacitance value corresponding to different capacitors, C r is the capacitance value of the resonant capacitor, and round() is the rounding function; After calculating the number of different capacitors in series, the feasibility label is calculated, and the calculation formula is as follows: Where n is the number of capacitors that need to be connected in series, N cmax is the maximum number of stacks, U ci is the voltage value that a single capacitor can withstand, u cmax The maximum voltage that the total capacitor can withstand, Label is the feasibility label; After obtaining the feasibility label, the capacitor cost is calculated, and the calculation formula is as follows: F c =F ci ×n i (13) Among them, F c is the total price required for different capacitors, F ci The price of a single capacitor with different capacitance values, n i The number of capacitors required for different capacitance values; After calculating the cost of different capacitors, the equivalent series resistance and loss are calculated. The calculation formula is as follows: Among them, ESR is the equivalent series resistance of the capacitor, tanδ is the loss angle of the capacitor, and f s is the switching frequency corresponding to the output of rated voltage from the low-voltage DC side to the high-voltage DC side, and C is the capacitance value of the capacitor; P C =I C 2 ×ESR (15) Among them, I C is the effective value of the current flowing through the capacitor, P C is the loss of the capacitor.

5. The resonant cavity parameter optimization design method for the energy storage bidirectional DC converter LLC circuit according to claim 1 is characterized in that: Integrate the inductor and capacitor related design data into the initial input data set, use the comprehensive index data calculated using the cost and loss data as the first output data set, and use the feasibility label as the second output data set. Calculate the maximum mutual information coefficient between each feature of the input data set and the first output data set and the second output data set, and use the maximum mutual information coefficient to screen out the input data features with high correlation with the first output data set and the second output data set, and form two new input data sets respectively. The specific method is as follows: Select the design data related to inductance and capacitance, including resonant inductance value, resonant capacitance value, excitation inductance value, magnetic core inner diameter, maximum current density effective value, magnetic core effective flux cross-sectional area, magnetic core window area, and actual capacitance value, and integrate them into the initial input data set; After obtaining the initial input data set, the total cost and total loss data are used to calculate the comprehensive index, and it is used as the first output data set. The feasibility labels corresponding to each input data are used as the second output data set. The calculation formulas for total cost, total loss and comprehensive index are as follows: F i总 =F iLr +F iLm +F iCr (16) P i总 =P iLr +P iLm +P iCr (17) Among them, t i is the comprehensive index of cost and loss of the i-th design solution, F i总 is the total design cost of the ith design solution, F max is the maximum design cost among all design options, P i总 is the total loss of the ith design solution, P max is the maximum loss among all design solutions, F iLr The total cost of the resonant inductor designed for the ith design, F iLm The total cost of the magnetizing inductance designed for the ith design, F iCr The total cost of the resonant capacitor designed for the ith design, P iLr is the total loss of the resonant inductor, P iLm is the total loss of the excitation inductance, P iCr is the total loss of the resonant capacitor; After obtaining the first output data set and the second output data set, calculating the maximum mutual information coefficient between each feature of the input data set and the first output data set and the second output data set; After calculating the maximum mutual information coefficient between each feature and the output data set, the features in the input data set are sorted according to the maximum mutual information coefficient, and some features with large maximum mutual information coefficients are selected as the features of the new input data set according to the required data dimension.

6. The resonant cavity parameter optimization design method for the energy storage bidirectional DC converter LLC circuit according to claim 5 is characterized in that: The calculation process of the maximum mutual information coefficient between the features in the input dataset and a single output dataset is as follows: (1) Selecting data of one feature in the input data set and the output data set in turn to form a two-dimensional data set, and plotting them in the form of coordinate points in a two-dimensional coordinate plane; (2) The dividing lines of the divided areas are calculated according to the precision traversal, where the calculation formulas for the dividing lines of the x-axis and y-axis are as follows: x 分割线 =k×d x (19) and 分割线 =k×d y (20) Among them, x 分割线 is the x-coordinate corresponding to the current x-axis dividing line, y 分割线 is the y coordinate corresponding to the current y-axis dividing line, k is the number of current traversals, d x The step size used to traverse the x-axis, d y The step size used to traverse the y axis; (3) After the dividing line is calculated, it divides the image into four regions: "upper left", "lower left", "upper right", and "lower right". At this time, the coordinate points in different regions are counted respectively, and the probability of the coordinate points in each region is calculated. The calculation formula is as follows: Among them, p is the probability of the coordinate point appearing in the area, n i is the number of coordinate points in the area, n 总 is the total number of coordinate points; (4) After calculating the probability of the coordinate point in each area, the mutual information value between the feature and the output data set under the current traversal is calculated according to the mutual information calculation formula. The calculation formula is as follows: Among them, I(x,y) is the mutual information value of the two variables x and y, p(X=x i ,Y=y j ) is in (x i ,y j ) region, p(x=x i ) is X in x i The probability of the region, p(Y = y j ) is Y in y j The probability of the region; when x i When the subscript i=1, it represents the left side, and when i=2, it represents the right side; when y j When the subscript j=1, it represents the upper side, and when j=2, it represents the lower side; (5) After changing the segmentation line by traversing and calculating the mutual information value of each traversal feature and the output data set, when the segmentation line meets the traversal condition, stop traversing and calculate the maximum mutual information coefficient. The calculation formula is as follows: Among them, mic(x,y) is the maximum mutual information coefficient between the feature and the output data set, max() is the maximum value function, I(x,y) is the mutual information coefficient between the feature and the output data set, a is the number of regions divided in the x direction, and b is the number of regions divided in the y direction.

7. The resonant cavity parameter optimization design method for the energy storage bidirectional DC converter LLC circuit according to claim 1 is characterized in that: The new input data set obtained has two rows, which are the data set after the input data set and the first output data set filter features and the data set after the input data set and the second output data set filter features. These two rows of data are combined according to the upper and lower rows of data sequence numbers to obtain the final input data set; the first output data set and the second output data set are combined according to the upper and lower rows of data sequence numbers to obtain the final output data set.

8. The resonant cavity parameter optimization design method for the energy storage bidirectional DC converter LLC circuit according to claim 1 is characterized in that: The final input data and the final output data are input into SVM for model training; after the model training is completed, M input data are randomly generated according to the data format of the final input data, where each input data represents a design scheme; these randomly generated input data are input into SVM to make it output the corresponding comprehensive index and the feasibility label of each design scheme; The input data, i.e., the design scheme, whose output feasibility label is 0 is eliminated, and the comprehensive indicators corresponding to the remaining input data are sorted, and the input data, i.e., the design scheme, with the largest comprehensive indicator is selected as the optimal design scheme after optimization.

Images