Optimization method and system for hybrid distribution transformer

Abstract

The invention relates to an optimization method and system for a hybrid distribution transformer. The method comprises the following steps: acquiring operation data of the hybrid distribution transformer; constructing a magnetic circuit topology matrix, and calculating a reconstruction path to obtain a reconstruction decision matrix; performing feasibility verification and reliability analysis on the reconstruction decision matrix to obtain a feasible reconstruction scheme set and a reliability evaluation matrix; calculating an integration level index of each feasible reconstruction scheme to obtain an integration level evaluation matrix; performing multi-objective optimization calculation to obtain a balance optimization scheme; calculating the dynamic response of the multi-physics field coupling model by using a finite element discretization algorithm to obtain a dynamic response sequence; solving a control strategy optimization problem by using an interior point method to obtain a control strategy; optimizing by utilizing a self-adaptive parameter adjustment algorithm to obtain a parameter correction matrix; obtaining an optimal control strategy based on the parameter correction matrix; and the hybrid distribution transformer completes optimization based on the optimal control strategy.

Description

Technical Field

[0001] This application relates to the field of hybrid distribution transformer optimization, and mainly relates to an optimization method and system for hybrid distribution transformers. Background Art

[0002] As a core component of the new generation of distribution equipment, the hybrid distribution transformer integrates the advantages of traditional transformers and power electronic converters, and can achieve flexible conversion of voltage levels and active control of power quality. With the rapid development of smart grids and distributed energy systems, higher requirements are put forward for the performance of hybrid distribution transformers, especially in terms of power density, dynamic response, and system reliability.

[0003] Currently, the design optimization of hybrid distribution transformers mainly adopts a method combining empirical design and single physical field simulation. In terms of magnetic field analysis, a two-dimensional finite element model is mainly used to calculate the magnetic flux density distribution and eddy current loss; in terms of thermal field analysis, a simplified thermal network method is adopted to estimate the temperature rise; in terms of mechanical stress analysis, it is mainly checked according to empirical formulas. Although these methods have been applied to a certain extent in engineering practice, due to ignoring the coupling effect between physical fields, it is difficult to accurately predict the actual performance of the system. At the same time, the existing optimization methods mostly adopt single-objective optimization strategies, and it is difficult to balance multiple performance indicators such as power density, efficiency, and reliability.

[0004] For example, the Chinese invention patent with the publication number "CN116029067A" discloses a "Transformer Heat Dissipation Structure Optimization Design Method and Device", which specifically discloses "obtaining each design object corresponding to the transformer heat dissipation structure and the initial design parameters corresponding to each design object; based on each initial design parameter, obtaining the parameters of the heat dissipation structure to be verified; based on the finite element analysis model, simulating the parameters of the heat dissipation structure to be verified to obtain the target temperature distribution image corresponding to the heat dissipation structure to be verified; obtaining the historical temperature distribution image corresponding to the actual heat dissipation structure; comparing the similarity between the target temperature distribution image and the historical temperature distribution image to obtain a comparison result; where the actual heat dissipation structure is the heat dissipation structure with temperature anomalies at historical time points; according to the comparison result, determining whether the heat dissipation structure to be verified meets the optimization conditions", but this method only compares based on the data of the heat dissipation structure with historical temperature anomalies, and may not cover the optimization requirements under normal operating conditions, resulting in the deviation of the design optimization direction from the actual scenario; in addition, this method does not conduct simulation in combination with the electro-multiphysical field coupling effect, and it is easy to ignore the influence of the interaction between the heat dissipation structure and other components of the transformer on the temperature distribution. Summary of the Invention

[0005] In order to solve the above problems existing in the prior art, this application provides an optimization method and system for hybrid distribution transformers.

[0006] The technical solution of this application is as follows:

[0007] On the one hand, the present invention proposes an optimization method for a hybrid distribution transformer, and the method includes:

[0008] Obtain the operation data of the hybrid distribution transformer and perform data cleaning on the operation data;

[0009] Construct a magnetic circuit topology matrix, calculate the reconstruction path of the magnetic circuit topology matrix using the minimum spanning tree to obtain a reconstruction decision matrix; perform feasibility verification and reliability analysis on the reconstruction decision matrix to obtain a set of feasible reconstruction schemes and a reliability evaluation matrix; calculate the integration index of each feasible reconstruction scheme based on the set of feasible reconstruction schemes to obtain an integration evaluation matrix; perform multi-objective optimization calculation on the integration evaluation matrix and the reliability evaluation matrix to obtain a balanced optimization scheme; construct a multi-physical field coupling model, and calculate the dynamic response using the finite element discretization algorithm to obtain a dynamic response sequence;

[0010] Construct a control strategy optimization problem, where the control strategy optimization problem includes an objective function constructed from the dynamic response sequence and the balanced optimization scheme, and constraint conditions set based on the cleaned operation data; use the interior point method to solve the control strategy optimization problem to obtain a control strategy; and optimize the control strategy parameters using an adaptive parameter adjustment algorithm to obtain a parameter correction matrix; based on the parameter correction matrix, obtain an optimal control strategy;

[0011] Optimize the hybrid distribution transformer based on the optimal control strategy.

[0012] Preferably, calculating the reconstruction path of the magnetic circuit topology matrix using the minimum spanning tree to obtain a reconstruction decision matrix specifically includes:

[0013] Construct a magnetic circuit topology matrix, where the edge weight set of the magnetic circuit topology matrix is expressed by the formula:

[0014]

[0015] W = ∑W(i,j);

[0016] In the formula, W(i,j) represents the edge weight between the i-th node and the j-th node; F i represents the eigenvector of the i-th node; F j represents the eigenvector of the j-th node; P i represents the spatial position vector of the i-th node; P j represents the spatial position vector of the j-th node; φ i represents the magnetic potential of the i-th node; φ j represents the magnetic potential of the j-th node; α F represents the characteristic weight coefficient; αP represents the spatial position weight coefficient; α φ represents the magnetic potential weight coefficient; σ1 represents the characteristic scale parameter; σ2 represents the spatial position scale parameter; σ3 represents the magnetic potential scale parameter; i represents the index value of the i-th node; j represents the index value of the j-th node;

[0017] Initialize the candidate spanning trees of the preset trees, and take the minimum total cost of the minimum spanning tree as the objective function to calculate the reconstruction path of the magnetic circuit topology matrix, where the total cost objective function is expressed by the formula:

[0018] f = -∑W(i,j) + μ∑D(i) + η∑R(i,j);

[0019]

[0020] R(i,j) = 1 - Π(1 - r(i,j));

[0021] In the formula, D(i) represents the metric function of the i-th node; R(i,j) represents the reliability function between the i-th node and the j-th node; μ represents the first balance coefficient; η represents the second balance coefficient; d(i,j) represents the distance between the i-th node and the j-th node; ρ represents the attenuation coefficient; exp() represents the exponential function; r(i,j) represents the reliability of the corresponding component of the edge between the i-th node and the j-th node; f represents the total cost objective function;

[0022] Iteratively solve the optimal spanning tree until the maximum number of iterations is reached or the total cost objective function converges to obtain the optimal spanning tree;

[0023] Arrange the edge set of the optimal spanning tree in the order of iterative solution of the optimal spanning tree to obtain the reconstruction decision matrix denoted as G ref = {g o (i,j)}, where g o (i,j) represents that the edge selected for the o-th reconstruction is (i,j), that is, the o-th reconstruction decision scheme, and o represents the index value of the o-th reconstruction.

[0024] Preferably, calculate the integration index of each feasible reconstruction scheme based on the set of feasible reconstruction schemes, which is expressed by the formula:

[0025]

[0026] In the formula, I(k) represents the integration index of the k-th feasible reconstruction scheme; V(k) represents the volume utilization rate of the k-th feasible reconstruction scheme; V eff (k) represents the effective volume of the k-th feasible reconstruction scheme; V tot(k) represents the total volume of the k-th feasible reconstruction solution; T(k) represents the temperature distribution function of the k-th feasible reconstruction solution; ΔT(k) represents the temperature change of the k-th feasible reconstruction solution; ΔT max represents the temperature change threshold of the k-th feasible reconstruction solution; P(k) represents the power density of the k-th feasible reconstruction solution; P(k) represents the transmission power of the k-th feasible reconstruction solution; B(k) represents the magnetic flux density distribution of the k-th feasible reconstruction solution; ω v represents the volume weight coefficient; λ represents the temperature influence coefficient; ε represents the magnetic field uniformity coefficient; δ(B(k)) represents the standard deviation of the magnetic flux density of the k-th feasible reconstruction solution; μ B represents the mean value of the magnetic flux density; B z represents the magnetic flux density at the z-th sampling point; z represents the index value of the z-th sampling point; Z represents the total number of sampling points;

[0027] The obtained integration evaluation matrix is expressed as where K represents the number of feasible reconstruction solutions.

[0028] Preferably, the integration evaluation matrix and the reliability evaluation matrix are subjected to multi-objective optimization calculation, specifically:

[0029] The multi-objective optimization function is expressed by the formula:

[0030] F(k) = max{ω I ·(I * - I(k)), ω R ·(R * - R(k))};

[0031] I * = max(I(k));

[0032] R * = max(R(k));

[0033] In the formula, R(k) represents the reliability index of the k-th feasible reconstruction solution; I * represents the candidate optimal solution of the integration index; R * represents the candidate optimal solution of the reliability index; ω I represents the weight of the integration index; ω R represents the weight of the reliability index;

[0034] Using the non-dominated sorting genetic algorithm to solve the multi-objective optimization function is specifically to initialize the population, where the population includes K individuals, and each individual corresponds to a feasible reconstruction solution; iterate the population evolution process until the preset maximum number of evolutionary iterations is reached, and generate the Pareto optimal solution set;

[0035] The comprehensive membership of the integration index and reliability index is calculated using the fuzzy membership function, which can be expressed as follows:

[0036]

[0037] ζ(k)=a·ζ I (k)+(1-a)·ζ R (k);

[0038]

[0039] Where, ζ I (k) represents the degree of satisfaction of the integration index of the kth feasible reconstruction scheme; ζ R (k) represents the degree of satisfaction of the reliability index of the kth feasible reconstruction scheme; I min Indicates the minimum value of the integration index; I max Indicates the maximum value of the integration index; R min Represents the minimum value of the reliability index; R max Represents the maximum value of the reliability index; ζ(k) represents the comprehensive membership of the kth feasible reconstruction scheme; a represents the membership weight; k * Indicates the index value of the feasible reconstruction scheme corresponding to the maximum comprehensive membership; argmax indicates the maximum comprehensive membership function;

[0040] The obtained balanced optimization solution set is expressed as {I(k * ),R(k * )}.

[0041] Preferably, the dynamic response is calculated using a finite element discretization algorithm to obtain a dynamic response sequence, specifically:

[0042] The multi-physics coupling model is expressed as follows:

[0043]

[0044] Wherein, Φ represents the generalized field quantity; S represents the diffusion coefficient matrix; K represents the coupling coefficient matrix; A represents the magnetic vector potential; B represents the magnetic induction intensity; Q represents the temperature field; U represents the stress tensor; γ1 represents the thermoelectric coupling coefficient; γ2 represents the electromechanical coupling coefficient; curl() represents the analytical vector field rotation characteristic function; t represents the time; Represents the Nabla operator; grad() represents the gradient function; div() represents division;

[0045] The dynamic response is calculated using the finite element discretization algorithm, and the finite element discretization equation is obtained, which can be expressed as follows:

[0046]

[0047] m ij = ∫ Ω ρX i X j dΩ;

[0048] In the formula, N represents the stiffness matrix; M represents the mass matrix; J represents the damping matrix; c represents the nodal displacement vector; Y represents the external load vector; τ represents the relaxation factor; O(c) represents the non - linear nodal displacement vector term; l ij represents the element of the stiffness matrix between the i - th node and the j - th node; m ij represents the element of the mass matrix between the i - th node and the j - th node; X i represents the shape function of the i - th node; X j represents the shape function of the j - th node; Ω represents the preset finite - element calculation domain;

[0049] Iteratively solve the finite - element discretized equation until the maximum number of equation - solving iterations is reached or the finite - element discretized equation converges, then stop the equation - solving iteration to obtain the dynamic response sequence.

[0050] Preferably, construct an optimization problem for the control strategy, specifically:

[0051] Construct the objective function of the optimization problem for the control strategy, which is expressed by the formula:

[0052]

[0053] In the formula, CF represents the objective function of the optimization problem for the control strategy; y E represents the preset dynamic tracking coefficient; E q represents the dynamic response at the q - th time step; EX q represents the expected value at the q - th time step; Q represents the number of time steps; q represents the index value of the q - th time step; y C represents the preset control smoothing coefficient; u q represents the control input at the q - th time step; y I represents the preset integration - degree optimization coefficient; y R represents the preset reliability - optimization coefficient; I * represents the candidate optimal solution of the integration - degree index; R * represents the candidate optimal solution of the reliability index; I(k) represents the integration - degree index of the k - th feasible reconstruction scheme; R(k) represents the reliability index of the k - th feasible reconstruction scheme;

[0054] Construct the constraint conditions of the optimization problem for the control strategy, which are expressed by the formula:

[0055]

[0056] where, u min represents the minimum value of the control input; u max represents the maximum value of the control input; T q represents the temperature distribution function at the q-th time step; T max represents the preset maximum temperature rise threshold; B q represents the magnetic flux density distribution function at the q-th time step; B sat represents the saturation magnetic flux density of the magnetic core material; represents the magnetic potential at the z-th sampling point; is the rated magnetic potential;

[0057] The interior point method is used to solve the control strategy optimization problem. The constraints of the control strategy optimization problem are transformed into penalty terms, and a Lagrangian function is constructed, which is expressed by the formula:

[0058]

[0059] where, L represents the Lagrangian function; χ represents the penalty term factor;

[0060] The Lagrangian function is solved, which is expressed by the formula:

[0061]

[0062] where, H represents the Hessian matrix; h represents the index value of the h-th iteration to solve the Lagrangian function; sp represents the preset step size;

[0063] The process of iteratively solving the Lagrangian function is carried out until the maximum number of iterations for solving the Lagrangian function is reached or the objective function of the control strategy optimization problem converges, and the iteration is stopped to obtain the control strategy.

[0064] Preferably, an adaptive parameter adjustment algorithm is used to optimize the control strategy parameters to obtain a parameter correction matrix, specifically:

[0065] The performance index of the control strategy parameters is calculated, which is expressed by the formula:

[0066]

[0067] θ = {y E , y C , y I , y R};

[0068] where, PI(θ) represents the performance index of the control strategy parameters; b r represents the efficiency weight coefficient of the r-th control strategy parameter; C r (θ) represents; β rThe temperature rise influence coefficient representing the r-th control strategy parameter; U r (θ) represents the temperature rise function of the r-th control strategy parameter; represents the reliability compensation coefficient; θ represents the control strategy parameter; r represents the index value of the r-th control strategy parameter;

[0069] The sensitivity analysis of the control strategy parameters is carried out by using the Morris screening method, which is expressed by the formula:

[0070]

[0071] In the formula, S(θ) represents the control strategy parameter sensitivity index; s1 represents the second-order sensitivity weight; s2 represents the indirect effect weight; w r represents the state variable weight coefficient of the r-th control strategy parameter; θ r represents the r-th control strategy parameter;

[0072] The control strategy parameters are optimized by using an adaptive parameter adjustment algorithm, which is expressed by the formula:

[0073]

[0074] In the formula, represents the r-th control strategy parameter of the n-th parameter correction iteration; (lr) n represents the adaptive learning rate of the n-th parameter correction iteration; (lr) 0 represents the preset initial adaptive learning rate; n represents the index value of the n-th parameter correction iteration;

[0075] Calculate the adjustment amount of the control strategy parameters for each iteration Generate a parameter correction matrix.

[0076] On the other hand, the present invention also proposes an optimization system for a hybrid distribution transformer, and the system includes a data acquisition module, an optimization module and a result output module, wherein:

[0077] The data acquisition module is used to acquire the operation data of the hybrid distribution transformer, and perform data cleaning on the operation data; and transmit the cleaned operation data to the optimization module;

[0078] The optimization module is used to construct a magnetic circuit topology matrix, calculate the reconstruction path of the magnetic circuit topology matrix by using the minimum spanning tree, and obtain a reconstruction decision matrix; perform feasibility verification and reliability analysis on the reconstruction decision matrix to obtain a set of feasible reconstruction schemes and a reliability evaluation matrix; calculate the integration index of each feasible reconstruction scheme based on the set of feasible reconstruction schemes to obtain an integration evaluation matrix; perform multi-objective optimization calculation on the integration evaluation matrix and the reliability evaluation matrix to obtain a balanced optimization scheme; construct a multi-physical field coupling model, and calculate the dynamic response by using a finite element discretization algorithm to obtain a dynamic response sequence;

[0079] Construct a control strategy optimization problem, where the control strategy optimization problem includes an objective function constructed by the dynamic response sequence and the balanced optimization scheme, and constraint conditions set based on the cleaned operation data; use the interior point method to solve the control strategy optimization problem to obtain a control strategy; and optimize the control strategy parameters by using an adaptive parameter adjustment algorithm to obtain a parameter correction matrix; based on the parameter correction matrix, obtain an optimal control strategy;

[0080] Optimize the hybrid distribution transformer based on the optimal control strategy;

[0081] The result output module is used to display the optimal control strategy.

[0082] On the other hand, the present invention also proposes an electronic device, including a memory, a processor, and a computer program stored on the memory and executable on the processor. When the processor executes the program, it implements an optimization method for a hybrid distribution transformer as described in any embodiment of the present invention.

[0083] On the other hand, the present invention also proposes a computer-readable storage medium, on which a computer program is stored. When the program is executed by a processor, it implements an optimization method for a hybrid distribution transformer as described in any embodiment of the present invention.

[0084] Compared with the prior art, the beneficial effects of the present invention are:

[0085] 1) The present invention provides an optimization method and system for a hybrid distribution transformer. Through operation data cleaning, magnetic circuit reconstruction, and integration evaluation, the collaborative optimization level of transformer efficiency, reliability, and economy is improved, and the dynamic response accuracy under complex working conditions is enhanced;

[0086] 2) The present invention provides an optimization method and system for a hybrid distribution transformer. By combining a multi-physical field coupling model and dynamic response sequence analysis, the adaptive ability of the system to environmental changes and load fluctuations is improved, and the long-term operation and maintenance costs and failure risks are reduced; among them, through the finite element discretization algorithm and dynamic response sequence calculation, the dynamic adaptation ability to multi-physical field coupling effects is strengthened;

[0087] 3) The present invention provides an optimization method and system for a hybrid distribution transformer. Through multi-objective optimization and interior point method solution, the convergence efficiency of the global optimal solution under complex constraints is improved; combined with an adaptive parameter adjustment algorithm and a parameter correction matrix, the real-time response ability of the control strategy to operating condition fluctuations is enhanced. BRIEF DESCRIPTION OF THE DRAWINGS

[0088] Figure 1 is a flowchart of the method of an embodiment of the present invention. DETAILED DESCRIPTION OF THE EMBODIMENTS

[0089] The following describes the specific embodiments of the present invention to facilitate those skilled in the art of the present technology to understand the present invention. However, it should be clear that the present invention is not limited to the scope of the specific embodiments. For those of ordinary skill in the art of the present technology, as long as various changes are within the spirit and scope of the present invention defined and determined by the appended claims, these changes are obvious, and all inventions and creations using the concept of the present invention are within the scope of protection.

[0090] The present invention provides the following technical solutions: an optimization method and system for a hybrid distribution transformer.

[0091] Embodiment 1

[0092] Specifically refer to Figure 1 , this embodiment provides an optimization method for a hybrid distribution transformer, and the specific steps include:

[0093] S1. Obtain the operation data of the hybrid distribution transformer and perform data cleaning on the operation data;

[0094] The operation data includes core parameters, winding parameters, rated parameters, operation parameters, electromagnetic field parameters, thermal field parameters, and loss parameters;

[0095] The core parameters include core column diameter, core window height, core material, core window width, core lamination thickness, eddy current loss coefficient, hysteresis loss coefficient, and lamination coefficient;

[0096] The winding parameters include high-voltage turns, high-voltage wire cross-sectional area, high-voltage number of layers, high-voltage inner and outer diameters, high-voltage height, low-voltage turns, low-voltage wire cross-sectional area, low-voltage number of layers, low-voltage inner and outer diameters, low-voltage height, winding material, winding material conductivity, winding filling factor, winding pitch, and oil duct width;

[0097] The rated parameters include rated capacity, high-voltage rated voltage, low-voltage rated voltage, high-voltage side rated current, low-voltage side rated current, rated frequency, connection group, and impedance voltage;

[0098] The operation parameters include no-load current, load factor, power factor, voltage regulation rate, and cooling mode parameters;

[0099] The electromagnetic field parameters include permeability, magnetic induction intensity, magnetic field intensity, current density, and electric field intensity;

[0100] The thermal field parameters include thermal conductivity, specific heat capacity, density, temperature distribution, and heat dissipation coefficient;

[0101] The loss parameters include hysteresis loss, eddy current loss, additional loss, and winding loss;

[0102] S2. Specifically constructing the magnetic circuit topology matrix means converting the connection relationship between magnetic circuit nodes into a weighted directed graph; calculating the reconstruction path of the magnetic circuit topology matrix using the minimum spanning tree to obtain a reconstruction decision matrix;

[0103] Construct the magnetic circuit topology matrix, where the edge weight set of the magnetic circuit topology matrix is expressed by the formula:

[0104]

[0105] W = ∑W(i,j);

[0106] In the formula, W(i,j) represents the edge weight between the i-th node and the j-th node; F i represents the eigenvector of the i-th node; F j represents the eigenvector of the j-th node; P i represents the spatial position vector of the i-th node; P j represents the spatial position vector of the j-th node; φ i represents the magnetic potential of the i-th node; φ j represents the magnetic potential of the j-th node; α F represents the characteristic weight coefficient; α P represents the spatial position weight coefficient; α φ represents the magnetic potential weight coefficient; σ1 represents the characteristic scale parameter; σ2 represents the spatial position scale parameter; σ3 represents the magnetic potential scale parameter; i represents the index value of the i-th node; j represents the index value of the j-th node;

[0107] Initialize the candidate spanning trees of the preset number of trees, and calculate the reconstruction path of the magnetic circuit topology matrix with the minimum total cost of the minimum spanning tree as the objective function, where the total cost objective function is expressed by the formula:

[0108] f = -∑W(i,j) + μ∑D(i) + η∑R(i,j);

[0109]

[0110] R(i,j) = 1 - Π(1 - r(i,j));

[0111] Wherein, D(i) represents the metric function of the i-th node; R(i,j) represents the reliability function between the i-th node and the j-th node; μ represents the first balance coefficient; η represents the second balance coefficient; d(i,j) represents the distance between the i-th node and the j-th node; ρ represents the attenuation coefficient; exp() represents the exponential function; r(i,j) represents the reliability of the corresponding component of the edge between the i-th node and the j-th node; f represents the total cost objective function;

[0112] Iteratively solve the optimal spanning tree until the maximum number of iterations is reached or the total cost objective function converges to obtain the optimal spanning tree;

[0113] Arrange the edge set of the optimal spanning tree in the order of iterative solution of the optimal spanning tree to obtain a reconstructed decision matrix denoted as G ref ={g o (i,j)}, where g o (i,j) indicates that the edge selected for the o-th reconstruction is (i,j), that is, the o-th reconstructed decision scheme, and o represents the index value of the o-th reconstruction;

[0114] In this embodiment, the reconstruction decision method includes: Scheme 1: Concentric cylinder arrangement; Scheme 2: Staggered arrangement; Scheme 3: Flat arrangement;

[0115] S3. Perform feasibility verification and reliability analysis on the reconstructed decision matrix to obtain a set of feasible reconstruction schemes and a reliability evaluation matrix; Calculate the integration index of each feasible reconstruction scheme based on the set of feasible reconstruction schemes, which is expressed by the formula:

[0116]

[0117] Wherein, I(k) represents the integration index of the k-th feasible reconstruction scheme; V(k) represents the volume utilization rate of the k-th feasible reconstruction scheme; V eff (k) represents the effective volume of the k-th feasible reconstruction scheme; V tot (k) represents the total volume of the k-th feasible reconstruction scheme; T(k) represents the temperature distribution function of the k-th feasible reconstruction scheme; ΔT(k) represents the temperature change of the k-th feasible reconstruction scheme; ΔT max represents the temperature change threshold of the k-th feasible reconstruction scheme; P(k) represents the power density of the k-th feasible reconstruction scheme; P(k) represents the transmission power of the k-th feasible reconstruction scheme; B(k) represents the magnetic flux density distribution of the k-th feasible reconstruction scheme; ω v represents the volume weight coefficient; λ represents the temperature influence coefficient; ε represents the magnetic field uniformity coefficient; δ(B(k)) represents the standard deviation of the magnetic flux density of the k-th feasible reconstruction scheme; μ B represents the mean value of the magnetic flux density; B zIt represents the magnetic flux density at the z-th sampling point; z represents the index value of the z-th sampling point; Z represents the total number of sampling points;

[0118] The obtained integration degree evaluation matrix is expressed as where K represents the number of feasible reconstruction schemes;

[0119] In this embodiment, the integration degree evaluation results corresponding to the three schemes are specifically as follows: the volume utilization rate of Scheme 1 is 82%, and the power density is 2.8 kVA / dm 3 ; the volume utilization rate of Scheme 2 is 85%, and the power density is 3.0 kVA / dm 3 ; the volume utilization rate of Scheme 3 is 80%, and the power density is 2.6 kVA / dm 3 ;

[0120] S4. Perform multi-objective optimization calculation on the integration degree evaluation matrix and the reliability evaluation matrix. The multi-objective optimization function is expressed by the formula as:

[0121] F(k) = max{ω I ·(I * - I(k)), ω R ·(R * - R(k))};

[0122] I * = max(I(k));

[0123] R * = max(R(k));

[0124] In the formula, R(k) represents the reliability index of the k-th feasible reconstruction scheme; I * represents the candidate optimal solution of the integration degree index; R * represents the candidate optimal solution of the reliability index; ω I represents the weight of the integration degree index; ω R represents the weight of the reliability index;

[0125] Using the non-dominated sorting genetic algorithm to solve the multi-objective optimization function is specifically to initialize the population. The population includes K individuals, and each individual corresponds to a feasible reconstruction scheme; iterate the population evolution process until the preset maximum evolution iteration number is reached to generate the Pareto optimal solution set;

[0126] Using the fuzzy membership function to calculate the comprehensive membership degree of the integration degree index and the reliability index, which is expressed by the formula as:

[0127]

[0128] ζ(k) = a·ζ I (k) + (1 - a)·ζR (k);

[0129]

[0130] where ζ I (k) represents the degree of satisfaction of the integration index of the k-th feasible reconstruction scheme; ζ R (k) represents the degree of satisfaction of the reliability index of the k-th feasible reconstruction scheme; I min represents the minimum value of the integration index; I max represents the maximum value of the integration index; R min represents the minimum value of the reliability index; R max represents the maximum value of the reliability index; ζ(k) represents the comprehensive membership degree of the k-th feasible reconstruction scheme; a represents the membership degree weight; k * represents the index value of the feasible reconstruction scheme corresponding to the maximum comprehensive membership degree; argmax represents the function to obtain the maximum comprehensive membership degree;

[0131] The balanced optimization scheme set is obtained and expressed as {I(k * ), R(k * )};

[0132] S5. Construct a multi-physical field coupling model, where the multi-physical fields are specifically the electric field, magnetic field, thermal field, and mechanical field; use the finite element discretization algorithm to calculate the dynamic response and obtain the dynamic response sequence;

[0133] The multi-physical field coupling model is expressed by the formula:

[0134]

[0135] where Φ represents the generalized field quantity; S represents the diffusion coefficient matrix; K represents the coupling coefficient matrix; A represents the magnetic vector potential; B represents the magnetic induction intensity; Q represents the temperature field; U represents the stress tensor; γ1 represents the thermoelectric coupling coefficient; γ2 represents the electromechanical coupling coefficient; curl() represents the function to analyze the rotation characteristics of the vector field; t represents the time; represents the Nabla operator; grad() represents the gradient function; div() represents the division;

[0136] Use the finite element discretization algorithm to calculate the dynamic response and obtain the finite element discretization equation, which is expressed by the formula:

[0137]

[0138] m ij = ∫ Ω ρX i X j dΩ;

[0139] Wherein, N represents the stiffness matrix; M represents the mass matrix; J represents the damping matrix; c represents the nodal displacement vector; Y represents the external load vector; τ represents the relaxation factor; O(c) represents the non-linear nodal displacement vector term; l ij represents the element of the stiffness matrix between the i-th node and the j-th node; m ij represents the element of the mass matrix between the i-th node and the j-th node; X i represents the shape function of the i-th node; X j represents the shape function of the j-th node; Ω represents the preset finite element calculation domain;

[0140] Iteratively solve the finite element discretized equation until the maximum number of equation solving iterations is reached or the finite element discretized equation converges, then stop the equation solving iteration to obtain the dynamic response sequence;

[0141] S6. Construct an optimization problem for the control strategy, where the optimization problem for the control strategy includes an objective function constructed from the dynamic response sequence and the balance optimization scheme, and constraint conditions set based on the cleaned operation data;

[0142] Construct the objective function of the optimization problem for the control strategy, which is expressed by the formula:

[0143]

[0144] Wherein, CF represents the objective function of the optimization problem for the control strategy; y E represents the preset dynamic tracking coefficient; E q represents the dynamic response at the q-th time step; EX q represents the expected value at the q-th time step; Q represents the number of time steps; q represents the index value of the q-th time step; y C represents the preset control smoothing coefficient; u q represents the control input at the q-th time step; y I represents the preset integration optimization coefficient; y R represents the preset reliability optimization coefficient;

[0145] Among them, minimizing the first term is used to ensure that the operating state of the hybrid distribution transformer is maintained within the expected value; minimizing the second term is used to ensure that the frequent fluctuations of the control instructions are maintained within the preset range; the third term y I ·(I * -I(k)) is used to optimize the structural space of the hybrid distribution transformer; the fourth term y R ·(R * -R(k)) is used to stabilize the reliability of the hybrid distribution transformer; the priority of each term of the objective function can be adjusted by adjusting the corresponding coefficients;

[0146] Construct the constraint conditions for the control strategy optimization problem, which are expressed by the formula as:

[0147]

[0148] In the formula, u min represents the minimum value of the control input; u max represents the maximum value of the control input; T q represents the temperature distribution function at the q-th time step; T max represents the preset maximum temperature rise threshold; B q represents the magnetic flux density distribution function at the q-th time step; B sat represents the saturation magnetic flux density of the magnetic core material; represents the magnetic potential at the z-th sampling point; is the rated magnetic potential; I * represents the candidate optimal solution of the integration index; R * represents the candidate optimal solution of the reliability index; I(k) represents the integration index of the k-th feasible reconstruction scheme; R(k) represents the reliability index of the k-th feasible reconstruction scheme;

[0149] S7. Solve the control strategy optimization problem using the interior point method, transform the constraint conditions of the control strategy optimization problem into penalty terms, and construct the Lagrangian function, which is expressed by the formula as:

[0150]

[0151] In the formula, L represents the Lagrangian function; χ represents the penalty term factor;

[0152] Solve the Lagrangian function, which is expressed by the formula as:

[0153]

[0154] In the formula, H represents the Hessian matrix; h represents the index value of the h-th iteration for solving the Lagrangian function; sp represents the preset step size;

[0155] Iteratively solve the Lagrangian function until the maximum number of iterations for solving the Lagrangian function is reached or the objective function of the control strategy optimization problem converges, then stop the iteration and obtain the control strategy;

[0156] S8. And use the adaptive parameter adjustment algorithm to optimize the control strategy parameters to obtain the parameter correction matrix, specifically:

[0157] Calculate the performance index of the control strategy parameters, which is expressed by the formula as:

[0158]

[0159] θ = {yE , y C , y I , y R};

[0160] In the formula, PI(θ) represents the performance index of the control strategy parameters; b r represents the efficiency weight coefficient of the r-th control strategy parameter; C r (θ) represents; β r represents the temperature rise influence coefficient of the r-th control strategy parameter; U r (θ) represents the temperature rise function of the r-th control strategy parameter; represents the reliability compensation coefficient; θ represents the control strategy parameter; r represents the index value of the r-th control strategy parameter;

[0161] The sensitivity analysis of the control strategy parameters is carried out by using the Morris screening method, which is expressed by the formula:

[0162]

[0163] In the formula, S(θ) represents the control strategy parameter sensitivity index; s1 represents the second-order sensitivity weight; s2 represents the indirect effect weight; w r represents the state variable weight coefficient of the r-th control strategy parameter; θ r represents the r-th control strategy parameter;

[0164] The control strategy parameters are optimized by using the adaptive parameter adjustment algorithm, which is expressed by the formula:

[0165]

[0166] In the formula, represents the r-th control strategy parameter in the n-th parameter correction iteration; (lr) n represents the adaptive learning rate in the n-th parameter correction iteration; (lr) 0 represents the preset initial adaptive learning rate; n represents the index value of the n-th parameter correction iteration;

[0167] Calculate the adjustment amount of the control strategy parameters for each iteration Generate a parameter correction matrix;

[0168] S9. Based on the parameter correction matrix, obtain the optimal control strategy; optimize the hybrid distribution transformer based on the optimal control strategy.

[0169] Embodiment 2

[0170] This embodiment provides an optimization system for a hybrid distribution transformer. The system includes a data acquisition module, an optimization module, and a result output module, where:

[0171] The data acquisition module is used to acquire the operation data of the hybrid distribution transformer, clean the operation data, and transmit the cleaned operation data to the optimization module.

[0172] The optimization module is used to construct a magnetic circuit topology matrix, calculate the reconstruction path of the magnetic circuit topology matrix using the minimum spanning tree to obtain a reconstruction decision matrix, conduct feasibility verification and reliability analysis on the reconstruction decision matrix to obtain a set of feasible reconstruction schemes and a reliability evaluation matrix, calculate the integration index of each feasible reconstruction scheme based on the set of feasible reconstruction schemes to obtain an integration evaluation matrix, perform multi-objective optimization calculation on the integration evaluation matrix and the reliability evaluation matrix to obtain a balanced optimization scheme, construct a multi-physical field coupling model, and calculate the dynamic response using the finite element discretization algorithm to obtain a dynamic response sequence.

[0173] Construct a control strategy optimization problem, where the control strategy optimization problem includes an objective function constructed from the dynamic response sequence and the balanced optimization scheme, and constraint conditions set based on the cleaned operation data. Solve the control strategy optimization problem using the interior point method to obtain a control strategy, and optimize the control strategy parameters using an adaptive parameter adjustment algorithm to obtain a parameter correction matrix. Based on the parameter correction matrix, obtain an optimal control strategy.

[0174] Optimize the hybrid distribution transformer based on the optimal control strategy.

[0175] The result output module is used to display the optimal control strategy.

[0176] Embodiment 3

[0177] This embodiment provides an electronic device, including a memory, a processor, and a computer program stored on the memory and executable on the processor. When the processor executes the program, it implements an optimization method for a hybrid distribution transformer as described in any embodiment of the present invention.

[0178] Embodiment 4

[0179] This embodiment provides a computer-readable storage medium, on which a computer program is stored. When the program is executed by a processor, it implements an optimization method for a hybrid distribution transformer as described in any embodiment of the present invention.

[0180] It should be noted that the systems, electronic devices, and computer-readable storage media described in the present invention are all based on the same principle as the method described in Embodiment 1, and will not be elaborated here.

[0181] The above are only embodiments of the present invention, and do not thus limit the patent scope of the present invention. Any equivalent structure or equivalent process transformation made by using the content of the specification and drawings of the present invention, or directly or indirectly applied in other related technical fields, shall similarly be included within the patent protection scope of the present invention.

Claims

1. An optimization method for a hybrid distribution transformer, characterized in that, The method includes: Obtaining the operation data of the hybrid distribution transformer and performing data cleaning on the operation data; Constructing a magnetic circuit topology matrix, calculating the reconstruction path of the magnetic circuit topology matrix using the minimum spanning tree to obtain a reconstruction decision matrix; performing feasibility verification and reliability analysis on the reconstruction decision matrix to obtain a set of feasible reconstruction schemes and a reliability evaluation matrix; calculating the integration degree index of each feasible reconstruction scheme based on the set of feasible reconstruction schemes to obtain an integration degree evaluation matrix; performing multi-objective optimization calculation on the integration degree evaluation matrix and the reliability evaluation matrix to obtain a balanced optimization scheme; constructing a multi-physical field coupling model and calculating the dynamic response using the finite element discretization algorithm to obtain a dynamic response sequence; Constructing a control strategy optimization problem, where the control strategy optimization problem includes an objective function constructed from the dynamic response sequence and the balanced optimization scheme, and constraint conditions set based on the cleaned operation data; solving the control strategy optimization problem using the interior point method to obtain a control strategy; and optimizing the control strategy parameters using an adaptive parameter adjustment algorithm to obtain a parameter correction matrix; obtaining the optimal control strategy based on the parameter correction matrix; Optimizing the hybrid distribution transformer based on the optimal control strategy.

2. The optimization method of a hybrid distribution transformer according to claim 1, characterized in that Calculating the reconstruction path of the magnetic circuit topology matrix using the minimum spanning tree to obtain a reconstruction decision matrix, specifically: Constructing a magnetic circuit topology matrix, where the edge weight set of the magnetic circuit topology matrix is expressed by the formula: W = ∑W(i,j); Where, W(i,j) represents the edge weight between the i-th node and the j-th node; F i represents the feature vector of the i-th node; F j represents the feature vector of the j-th node; P i represents the spatial position vector of the i-th node; P j represents the spatial position vector of the j-th node; φ i represents the magnetic potential of the i-th node; φ j represents the magnetic potential of the j-th node; α F represents the feature weight coefficient; α P represents the spatial position weight coefficient; α φ represents the magnetic potential weight coefficient; σ1 represents the characteristic scale parameter; σ2 represents the spatial position scale parameter; σ3 represents the magnetic potential scale parameter; i represents the index value of the i-th node; j represents the index value of the j-th node; Initializing the candidate spanning trees of the preset number of trees, taking the minimum total cost of the minimum spanning tree as the objective function, and calculating the reconstruction path of the magnetic circuit topology matrix, where the total cost objective function is expressed by the formula: f = -∑W(i,j) + μ∑D(i) + η∑R(i,j); R(i,j) = 1 - Π(1 - r(i,j)); In the formula, D(i) represents the metric function of the i-th node; R(i,j) represents the reliability function between the i-th node and the j-th node; μ represents the first balance coefficient; η represents the second balance coefficient; d(i,j) represents the distance between the i-th node and the j-th node; ρ represents the attenuation coefficient; exp() represents the exponential function; r(i,j) represents the reliability of the corresponding component of the edge between the i-th node and the j-th node; f represents the total cost objective function; Iteratively solving the optimal spanning tree until the maximum number of iterations is reached or the total cost objective function converges to obtain the optimal spanning tree; Arrange the edge set of the optimal spanning tree in the order of iterative solution of the optimal spanning tree to obtain a reconstructed decision matrix denoted as \(G\). ref = \(\{g\) o (i, j)\}, where \(g\) o (i, j) indicates that the edge selected in the \(o\)-th reconstruction is \((i, j)\), that is, the \(o\)-th reconstruction decision scheme, and \(o\) represents the index value of the \(o\)-th reconstruction.

3. The optimization method of a hybrid distribution transformer according to claim 1, characterized in that Calculating the integration degree index of each feasible reconstruction scheme based on the set of feasible reconstruction schemes, which is expressed by the formula: Where, I(k) represents the integration index of the k-th feasible reconstruction scheme; V(k) represents the volume utilization rate of the k-th feasible reconstruction scheme; V eff (k) represents the effective volume of the k-th feasible reconstruction scheme; V tot (k) represents the total volume of the k-th feasible reconstruction scheme; T(k) represents the temperature distribution function of the k-th feasible reconstruction scheme; ΔT(k) represents the temperature change of the k-th feasible reconstruction scheme; ΔT max represents the temperature change threshold of the k-th feasible reconstruction scheme; P(k) represents the power density of the k-th feasible reconstruction scheme; P(k) represents the transmission power of the k-th feasible reconstruction scheme; B(k) represents the magnetic flux density distribution of the k-th feasible reconstruction scheme; ω v represents the volume weight coefficient; λ represents the temperature influence coefficient; ε represents the magnetic field uniformity coefficient; δ(B(k)) represents the standard deviation of the magnetic flux density of the k-th feasible reconstruction scheme; μ B represents the mean value of the magnetic flux density; B z represents the magnetic flux density at the z-th sampling point; z represents the index value of the z-th sampling point; Z represents the total number of sampling points; The integrated degree evaluation matrix is obtained and expressed as where K represents the number of feasible reconstruction schemes.

4. An optimization method for a hybrid distribution transformer according to claim 3, characterized in that Performing multi-objective optimization calculation on the integration degree evaluation matrix and the reliability evaluation matrix, specifically: The multi-objective optimization function is expressed by the formula: F(k) = max{ω I ·(I * - I(k)), ω R ·(R * - R(k))}; I * = max(I(k)); R * = max(R(k)); where \(R(k)\) represents the reliability index of the \(k\)-th feasible reconstruction scheme; \(I\) * represents the candidate optimal solution of the integration index; \(R\) * represents the candidate optimal solution of the reliability index; \(\omega\) I represents the weight of the integration index; \(\omega\) R represents the weight of the reliability index; Solving the multi-objective optimization function using the non-dominated sorting genetic algorithm, specifically initializing the population, where the population includes K individuals, and each individual corresponds to a feasible reconstruction scheme; iterating the population evolution process until the preset maximum number of evolutionary iterations is reached to generate a Pareto optimal solution set; Calculate the comprehensive membership degree of the integration index and the reliability index by using the fuzzy membership function, which is expressed by the formula as follows: ζ(k) = a·ζ I (k) + (1 - a)·ζ R (k); where ζ I (k) represents the degree of satisfaction of the integration index for the k-th feasible reconstruction solution; ζ R (k) represents the degree of satisfaction of the reliability index for the k-th feasible reconstruction solution; I min represents the minimum value of the integration index; I max represents the maximum value of the integration index; R min represents the minimum value of the reliability index; R max represents the maximum value of the reliability index; ζ(k) represents the comprehensive membership degree of the k-th feasible reconstruction solution; a represents the membership degree weight; k * represents the index value of the feasible reconstruction solution corresponding to the maximum comprehensive membership degree; argmax represents obtaining the maximum comprehensive membership degree function; Obtain the balanced optimization solution set, denoted as {I(k * ), R(k * )}.

5. The optimization method of a hybrid distribution transformer according to claim 1, wherein Construct a multi-physical-field coupling model, and use the finite element discretization algorithm to calculate the dynamic response to obtain a dynamic response sequence, specifically as follows: The multi-physical-field coupling model is expressed by the formula as follows: In the formula, Φ represents the generalized field quantity; S represents the diffusion coefficient matrix; K represents the coupling coefficient matrix; A represents the magnetic vector potential; B represents the magnetic induction intensity; Q represents the temperature field; U represents the stress tensor; γ1 represents the thermoelectric coupling coefficient; γ2 represents the electromechanical coupling coefficient; curl() represents the function for analyzing the rotation characteristics of a vector field; t represents the time; represents the Nabla operator; grad() represents the gradient function; div() represents the division; Use the finite element discretization algorithm to calculate the dynamic response to obtain a finite element discretization equation, which is expressed by the formula as follows: m ij = ∫ Ω ρX i X j dΩ; Where, N represents the stiffness matrix; M represents the mass matrix; J represents the damping matrix; c represents the nodal displacement vector; Y represents the external load vector; τ represents the relaxation factor; O(c) represents the non-linear nodal displacement vector term; l ij represents the element of the stiffness matrix between the i-th node and the j-th node; m ij represents the element of the mass matrix between the i-th node and the j-th node; X i represents the shape function of the i-th node; X j represents the shape function of the j-th node; Ω represents the preset finite element calculation domain; Iteratively solve the finite element discretization equation until the maximum number of equation-solving iterations is reached or the finite element discretization equation converges, and then stop the equation-solving iteration to obtain a dynamic response sequence.

6. The optimization method of a hybrid distribution transformer according to claim 5, characterized in that Construct an optimization problem for the control strategy, specifically as follows: Construct the objective function of the optimization problem for the control strategy, which is expressed by the formula as follows: where CF represents the objective function of the control strategy optimization problem; y E represents the preset dynamic tracking coefficient; E q represents the dynamic response at the q-th time step; EX q represents the expected value at the q-th time step; Q represents the number of time steps; q represents the index value of the q-th time step; y C represents the preset control smoothing coefficient; u q represents the control input at the q-th time step; y I represents the preset integration optimization coefficient; y R represents the preset reliability optimization coefficient; I * represents the candidate optimal solution of the integration index; R * represents the candidate optimal solution of the reliability index; I(k) represents the integration index of the k-th feasible reconstruction scheme; R(k) represents the reliability index of the k-th feasible reconstruction scheme; Construct the constraint conditions of the optimization problem for the control strategy, which is expressed by the formula as follows: where \(u\) min represents the minimum value of the control input; \(u\) max represents the maximum value of the control input; \(T\) q represents the temperature distribution function at the \(q\)-th time step; \(T\) max represents the preset maximum temperature rise threshold; \(B\) q represents the magnetic flux density distribution function at the \(q\)-th time step; \(B\) sat represents the saturation magnetic flux density of the magnetic core material; represents the magnetic potential at the \(z\)-th sampling point; is the rated magnetic potential; Use the interior point method to solve the optimization problem for the control strategy, transform the constraint conditions of the optimization problem for the control strategy into penalty terms, and construct a Lagrangian function, which is expressed by the formula as follows: In the formula, L represents the Lagrangian function; χ represents the penalty term factor; Solve the Lagrangian function, which is expressed by the formula as follows: In the formula, H represents the Hessian matrix; h represents the index value of the h-th iteration for solving the Lagrangian function; sp represents the preset step size; Iteratively solve the Lagrangian function until the maximum number of iterations for solving the Lagrangian function is reached or the objective function of the optimization problem for the control strategy converges, and then stop the iteration to obtain a control strategy.

7. An optimization method for a hybrid distribution transformer according to claim 6, characterized in that And use the adaptive parameter adjustment algorithm to optimize the control strategy parameters to obtain a parameter correction matrix, specifically as follows: Calculate the performance index of the control strategy parameters, which is expressed by the formula as follows: θ = {y E , y C , y I , y R}; Where, PI(θ) represents the performance index of the control strategy parameters; b r represents the efficiency weight coefficient of the r-th control strategy parameter; C r (θ) represents; β r represents the temperature rise influence coefficient of the r-th control strategy parameter; U r (θ) represents the temperature rise function of the r-th control strategy parameter; represents the reliability compensation coefficient; θ represents the control strategy parameter; r represents the index value of the r-th control strategy parameter; Use the Morris screening method to perform sensitivity analysis on the control strategy parameters, which is expressed by the formula as follows: In the formula, S(θ) represents the control strategy parameter sensitivity index; s1 represents the second-order sensitivity weight; s2 represents the indirect effect weight; w r The state variable weight coefficient representing the r-th control strategy parameter; θ r represents the r-th control strategy parameter; Use the adaptive parameter adjustment algorithm to optimize the control strategy parameters, which is expressed by the formula as follows: In the formula, represents the r-th control strategy parameter of the n-th parameter correction iteration; (lr) n represents the adaptive learning rate of the n-th parameter correction iteration; (lr) 0 represents the preset initial adaptive learning rate; n represents the index value of the n-th parameter correction iteration; Calculate the adjustment amount of the control strategy parameters for each iteration Generate a parameter correction matrix.

8. An optimization system for a hybrid distribution transformer, characterized in that, The system includes a data acquisition module, an optimization module, and a result output module, where: The data acquisition module is used to acquire the operation data of the hybrid distribution transformer, perform data cleaning on the operation data, and transmit the cleaned operation data to the optimization module; The optimization module is used to construct a magnetic circuit topology matrix, use the minimum spanning tree to calculate the reconstruction path of the magnetic circuit topology matrix to obtain a reconstruction decision matrix, perform feasibility verification and reliability analysis on the reconstruction decision matrix to obtain a set of feasible reconstruction schemes and a reliability evaluation matrix, calculate the integration index of each feasible reconstruction scheme based on the set of feasible reconstruction schemes to obtain an integration evaluation matrix, perform multi-objective optimization calculation on the integration evaluation matrix and the reliability evaluation matrix to obtain a balanced optimization scheme, construct a multi-physical-field coupling model, and use the finite element discretization algorithm to calculate the dynamic response to obtain a dynamic response sequence; Construct a control strategy optimization problem, where the control strategy optimization problem includes an objective function constructed from a dynamic response sequence and a balance optimization scheme, and constraint conditions set based on the cleaned operation data; solve the control strategy optimization problem using the interior point method to obtain a control strategy; and optimize the control strategy parameters using an adaptive parameter adjustment algorithm to obtain a parameter correction matrix; based on the parameter correction matrix, obtain an optimal control strategy; Optimize the hybrid distribution transformer based on the optimal control strategy; The result output module is used to display the optimal control strategy.

9. An electronic device, comprising a memory, a processor, and a computer program stored on the memory and executable on the processor, characterized in that, When the processor executes the program, it implements an optimization method for a hybrid distribution transformer according to any one of claims 1 to 7.

10. A computer-readable storage medium having a computer program stored thereon, characterized in that, When the program is executed by the processor, it implements an optimization method for a hybrid distribution transformer according to any one of claims 1 to 7.

Images