A method for predicting transformer design parameters

Abstract

The present invention provides a method for predicting transformer design parameters, comprising the following steps: setting control variables, constraint conditions, and objective functions of the algorithm; initially generating a superior population by means of randomly initializing the population. After the optimization program starts, an initial population is generated by using real-integer coding, and the fitness function is used to evaluate the advantages and disadvantages of each individual, thereby determining the size of its genetic opportunity; putting the obtained superior population into the transformer magnetic field cloud map rapid generation module to calculate the magnetic field cloud map under each structural parameter; putting the obtained magnetic field cloud map into the superior solution screening module to calculate the mean value and standard deviation, and selecting the structural parameters with smaller mean value and standard deviation of the magnetic field distribution as the guidance. The beneficial effects of the present invention: A method for predicting transformer design parameters, with the particle swarm optimization algorithm as the basic framework, has a very fast calculation speed compared with traditional algorithms, strong global search ability, is easy to converge to the global optimal solution, and the convergence speed is less affected by the population.

Description

Technical Field

[0001] The present invention belongs to the field of transformer structure and magnetic field, and in particular relates to a method for predicting transformer design parameters. Background Art

[0002] In the power loss of our country, the loss of the distribution network accounts for about 70% of the total network loss, becoming the main part of the entire power grid loss. The loads of the distribution network are mostly seasonal or daily loads. There are a large number of distribution transformers that are lightly loaded or unloaded for a long time, and their no-load losses account for a large part of the distribution network loss. Reducing the no-load loss of transformers can bring huge energy-saving and environmental protection benefits. Summary of the Invention

[0003] In view of this, the present invention aims to propose a method for predicting transformer design parameters, which can quickly calculate the design parameters that meet the conditions of transformer volume, capacity, and magnetic induction intensity.

[0004] To achieve the above object, the technical solution of the present invention is realized as follows:

[0005] A method for predicting transformer design parameters includes the following steps:

[0006] S1. Set the control variables, constraint conditions, and objective function of the algorithm;

[0007] S2. Initially generate a superior population by using the method of randomly initializing the population. After the optimization program starts, use the real-integer coding method to generate the initial population, and evaluate the quality of each individual through the fitness function to determine the size of its genetic opportunity;

[0008] S3. Put the superior population obtained in step S2 into the fast generation module of the transformer magnetic field cloud map to calculate the magnetic field cloud map under each structural parameter;

[0009] S4. Put the magnetic field cloud map obtained in step S3 into the superior solution screening module to calculate the mean value and standard deviation, and select the structural parameters with smaller mean value and standard deviation of the magnetic field distribution as the guide;

[0010] S5. Repeat the above steps, and conduct overall and local analyses with the volume and capacity of the transformer, magnetic induction intensity and volume as the optimization objectives until the Pareto optimal solution set is selected.

[0011] Further, in step S1, the volume and capacity of the transformer, magnetic induction intensity and volume are respectively used as the optimization objectives, and the window width B, window height C, core height D, and core width E are used as the control variables;

[0012] Optimization Objectives:

[0013]

[0014] Constraints:

[0015]

[0016] Objective function:

[0017] Volume

[0018]

[0019] Magnetic induction intensity

[0020] Where U is the voltage and W is the number of turns of the winding;

[0021] Volume

[0022] Where 0.9 is the effective cross-sectional area coefficient of the iron core.

[0023] Furthermore, a method of randomly initializing the population is used to initially generate a better population. After the optimization program starts, the real-integer coding method is used to generate the initial population. Each individual's quality is evaluated through the fitness function, thereby determining the size of its genetic opportunity.

[0024] Furthermore, in step S3, the obtained better population is put into the transformer magnetic field cloud map rapid generation module to calculate the magnetic field cloud map under each structural parameter. The magnetic field cloud map is calculated using a magnetic field prediction model based on deep learning methods. Using digital twin technology, the function values of other points are predicted based on the function values of typical sampling points, the field value distribution of the magnetic field is converted into an image, and finally the corresponding field values are inverted through the pixels of the predicted image;

[0025] Furthermore, in step S4, the obtained magnetic field cloud map is put into the better solution screening module to calculate the mean and standard deviation, and the structural parameters with smaller mean and standard deviation of the magnetic field distribution are selected as the guidance.

[0026] By changing the structure of the transformer iron core, the finite element magnetic field simulation results of the iron core magnetic field distribution are initially obtained. In the solution of the magnetic field mean and standard deviation, the magnetic field is first grayscale processed;

[0027] Formula for the mean of magnetic induction intensity:

[0028]

[0029] Formula for the standard deviation of magnetic induction intensity:

[0030]

[0031] An electronic device includes a processor and a memory communicatively connected to the processor and configured to store executable instructions of the processor, and the processor is configured to execute a method for predicting transformer design parameters.

[0032] A server includes at least one processor and a memory communicatively connected to the processor, the memory storing instructions executable by the at least one processor, and the instructions being executed by the processor to cause the at least one processor to execute a method for predicting transformer design parameters.

[0033] A computer-readable storage medium stores a computer program, and when the computer program is executed by a processor, a method for predicting transformer design parameters is implemented.

[0034] Compared with the prior art, a method for predicting transformer design parameters according to the present invention has the following beneficial effects:

[0035] (1) For the method for predicting transformer design parameters according to the present invention, after adding prior knowledge, the optimization result is relatively stable and the solution set distribution is relatively uniform;

[0036] (2) For the method for predicting transformer design parameters according to the present invention, with the particle swarm optimization algorithm as the basic framework, compared with traditional algorithms, the calculation speed is very fast, the global search ability is strong, it is easy to converge to the global optimal solution, and the convergence speed is less affected by the population. BRIEF DESCRIPTION OF THE DRAWINGS

[0037] The drawings constituting a part of the present invention are used to provide a further understanding of the present invention. The schematic embodiments of the present invention and their descriptions are used to explain the present invention and do not constitute an improper limitation to the present invention. In the drawings:

[0038] Figure 1 Flowchart of the combined guidance optimization algorithm;

[0039] Figure 2 Finite element magnetic field simulation results;

[0040] Figure 3 Schematic diagram of grayscale processing. DETAILED DESCRIPTION OF THE EMBODIMENTS

[0041] It should be noted that, without conflict, the embodiments in the present invention and the features in the embodiments may be combined with each other.

[0042] This solution discloses a method for predicting the design parameters of a transformer. Based on the particle swarm optimization algorithm as the basic framework, the volume and capacity of the transformer, and the magnetic induction intensity and volume are used as optimization objectives respectively. The window width B, window height C, core height D, and core width E are used as control variables (design parameters). First, a relatively optimal population is initially generated by using the method of randomly initializing the population. Further, the magnetic field cloud map under each structural parameter of the relatively optimal population is calculated. Then, the magnetic field distribution mean value and standard deviation of the obtained magnetic field cloud map are calculated, and the structural parameters with smaller mean value and standard deviation are selected as the guidance. Finally, the above steps are repeated until the Pareto optimal solution set is obtained. Ensure that the optimized result can make the magnetic field distribution of the transformer as uniform as possible while meeting the optimization objectives.

[0043] The present invention will be described in detail below with reference to the accompanying drawings and in conjunction with embodiments.

[0044] S1: Respectively use the volume and capacity of the transformer, and the magnetic induction intensity and volume as optimization objectives, and use the window width B, window height C, core height D, and core width E as control variables (design parameters).

[0045] Optimization objectives:

[0046]

[0047] Constraints:

[0048]

[0049] Objective function:

[0050] Volume

[0051]

[0052] Magnetic induction intensity

[0053] In the formula, U is the voltage and W is the number of winding turns.

[0054] Volume

[0055] In the formula, 0.9 is the core effective cross-sectional area coefficient;

[0056] S2: Use the method of randomly initializing the population to initially generate a relatively optimal population. After the optimization program starts, the initial population is generated by using the real-integer coding method, and the fitness function is used to evaluate the quality of each individual, so as to determine the size of its genetic opportunity.

[0057] Since the objective function in this solution always takes non - negative values and the optimization goal is to find the maximum or minimum of the function, the value of the objective function can be directly used as the fitness of the individual. Then, through selection, crossover, and mutation in turn, a better population is obtained, and the parameter settings are shown in Table 1.

[0058] Table 1 Parameter values

[0059]

[0060] S3: Put the obtained better population into the transformer magnetic field cloud map rapid generation module to calculate the magnetic field cloud map under each structural parameter. When calculating the magnetic field cloud map, several magnetic field prediction models based on deep learning methods are used. Using digital twin technology, the function values at other points can be predicted according to the function values at typical sampling points. It converts the field value distribution (color map) of the magnetic field into an image, predicts the pixels of the image, and finally inversely calculates the corresponding field values;

[0061] S4: Put the obtained magnetic field cloud map into the better solution screening module to calculate the mean and standard deviation, and select the structural parameters with smaller mean and standard deviation of the magnetic field distribution as the guidance.

[0062] By changing the transformer core structure, the finite - element magnetic field simulation results of the core magnetic field distribution are initially obtained as Figure 2 shown; in the solution of the magnetic field mean and standard deviation, first, the magnetic field is grayscale - processed, as Figure 3 shown.

[0063] Formula for the mean of magnetic induction intensity:

[0064]

[0065] Formula for the standard deviation of magnetic induction intensity:

[0066]

[0067] S5: Repeat the above steps, taking the volume and capacity of the transformer, magnetic induction intensity and volume as the optimization goals for overall and local analysis until the Pareto optimal solution set is selected.

[0068] Taking the transformer volume and capacity as the optimization goal, first run the optimization algorithm on the whole to obtain a preliminary better solution set. The transformer structure diagrams in the obtained better solution set are input into the transformer magnetic field rapid generation module to further obtain the corresponding magnetic field cloud maps. Solve the mean and standard deviation for each magnetic field cloud map. Select the structural parameters with small mean and standard deviation as the guidance, and finally obtain the Pareto optimal solution set. Next, focus on the analysis of the local magnetic field distribution, and obtain the Pareto solution set with prior knowledge added from the mean and standard deviation obtained during the optimization process.

[0069] Taking the magnetic induction intensity and volume of the transformer as the optimization objectives, perform the same steps above to optimize the overall average value and standard deviation to obtain the Pareto solution set.

[0070] Matters not covered by this invention are well-known technologies.

[0071] Those of ordinary skill in the art can realize that the units and method steps of each example described in combination with the embodiments disclosed herein can be implemented by electronic hardware, computer software, or a combination of the two. To clearly illustrate the interchangeability of hardware and software, the composition and steps of each example have been generally described according to functions in the above description. Whether these functions are executed in a hardware or software manner depends on the specific application and design constraints of the technical solution. Professional technicians can use different methods to implement the described functions for each specific application, but such implementation should not be considered to exceed the scope of this invention.

[0072] In several embodiments provided in this application, it should be understood that the disclosed methods and systems can be implemented in other ways. For example, the division of the above-mentioned units is only a logical function division, and there can be other division methods in actual implementation. For example, multiple units or components can be combined or integrated into another system, or some features can be ignored or not executed. The above-mentioned units may or may not be physically separated, and the components shown as units may or may not be physical units, that is, they can be located in one place or distributed to multiple network units. Some or all of the units can be selected according to actual needs to achieve the purpose of the solution of the embodiments of this invention.

[0073] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of this invention and are not intended to limit them. Although this invention has been described in detail with reference to the foregoing embodiments, those of ordinary skill in the art should understand that they can still modify the technical solutions described in the foregoing embodiments, or perform equivalent replacements for some or all of the technical features; and these modifications or replacements do not cause the essence of the corresponding technical solutions to deviate from the scope of the technical solutions of the embodiments of this invention, and they should all be covered within the scope of the claims and the description of this invention.

[0074] The above is only a preferred embodiment of this invention and is not intended to limit this invention. Any modifications, equivalent replacements, improvements, etc. made within the spirit and principle of this invention shall be included within the protection scope of this invention.

Claims

1. A method for predicting transformer design parameters, characterized in that, It includes the following steps: S1. Set the control variables, constraint conditions, and objective function of the algorithm; S2. Initially generate a superior population by using the method of randomly initializing the population. After the optimization program starts, use the real-integer coding method to generate the initial population. Evaluate the quality of each individual through the fitness function, thereby determining the size of its genetic opportunity; S3. Put the superior population obtained in step S2 into the transformer magnetic field cloud map rapid generation module to calculate the magnetic field cloud map under each structural parameter; S4. Put the magnetic field cloud map obtained in step S3 into the superior solution screening module to calculate the mean value and standard deviation, and select the structural parameters with smaller mean value and standard deviation of the magnetic field distribution as the guide; S5. Repeat the above steps, take the volume and capacity of the transformer, magnetic induction intensity and volume as the optimization objectives for overall and local analysis until the Pareto optimal solution set is selected; In step S1, take the volume and capacity of the transformer, magnetic induction intensity and volume as the optimization objectives, and the window width A, window height C, core height D, and core width E as the control variables; Optimization objective: Constraint condition: Objective function: Volume In the formula, U is the voltage and W is the number of turns of the winding; In the formula, 0.9 is the core effective cross-sectional area coefficient.

2. The method for predicting transformer design parameters according to claim 1, wherein: In step S3, put the obtained superior population into the transformer magnetic field cloud map rapid generation module to calculate the magnetic field cloud map under each structural parameter. Use the magnetic field prediction model based on the deep learning method to calculate the magnetic field cloud map. Use the digital twin technology to predict the function values of other points according to the function values of typical sampling points, convert the field value distribution of the magnetic field into an image, and finally invert the corresponding field value by predicting the pixels of the image.

3. The method for predicting design parameters of a transformer according to claim 1, wherein In step S4, put the obtained magnetic field cloud map into the superior solution screening module to calculate the mean value and standard deviation, and select the structural parameters with smaller mean value and standard deviation of the magnetic field distribution as the guide; By changing the transformer core structure, initially obtain the finite element magnetic field simulation results of the core magnetic field distribution. In the solution of the magnetic field mean value and standard deviation, first perform gray-scale processing on the magnetic field; Mean formula of magnetic induction intensity: Standard deviation formula of magnetic induction intensity:

4. An electronic device, comprising a processor and a memory communicatively connected to the processor and configured to store instructions executable by the processor, wherein: The processor is used to execute a transformer design parameter prediction method according to any one of claims 1-3 above.

5. A server, characterized in that: It includes at least one processor and a memory communicatively connected to the processor. The memory stores instructions executable by the at least one processor. The instructions are executed by the processor so that the at least one processor executes a transformer design parameter prediction method according to any one of claims 1-3 above.

6. A computer-readable storage medium stores a computer program, characterized in that: When the computer program is executed by the processor, it implements a transformer design parameter prediction method according to any one of claims 1-3.

Images