Switch cabinet hotspot analysis method, heat dissipation design method and electronic equipment

Abstract

This invention relates to the field of switchgear technology, and provides a switchgear hotspot analysis method, heat dissipation design method, and electronic device, including: calculating the fitness of each sample individual based on an objective function, storing them in a database in descending order, wherein the objective function is set based on the hotspot temperature of the switchgear; and sorting the top N... P Using sample individuals as the parent population, a child population is generated based on an improved mountain gazelle optimizer, an incremental kriging model, and the parent population. The improved mountain gazelle optimizer generates the child population based on a sinusoidal scaling factor and an adaptive distribution. The optimal candidate individuals in the child population are determined based on the incremental kriging model and the expected improvement criterion, and the evaluation count is incremented by one. When the maximum number of evaluations is reached, the optimal solution for the objective variable is determined based on the corresponding optimal candidate individuals, which is used to determine the optimal location of the heat dissipation components. This approach addresses the difficulty in balancing accuracy and efficiency when performing hotspot analysis on switchgear in related technologies.

Description

Technical Field

[0001] This invention relates to the field of switchgear technology, and in particular to a switchgear hotspot analysis method, heat dissipation design method, and electronic device. Background Technology

[0002] Switchgear is widely used in power transmission, serving functions such as control and protection. However, due to its compact internal structure, the heat generated during operation is difficult to dissipate effectively, especially in high-current switchgear with a rated operating current ≥630A, which is prone to overheating, potentially leading to major accidents. Therefore, conducting hotspot analysis on switchgear for heat dissipation design, thereby reducing the risks during switchgear operation, is a crucial foundation for the stable operation of power systems.

[0003] Relevant hotspot analysis methods include experimental measurement, model simulation analysis, and empirical formula estimation.

[0004] (1) The experimental measurement method generally involves placing contact sensors at key points on the contacts and cable busbars to directly obtain temperature data of key parts of the cable busbars. However, contact sensors are located in harsh environments with high potential and strong vibration, such as the contacts. They are affected by strong electromagnetic waves generated by the switching action, which often leads to abnormal data or even equipment failure. At the same time, their lifespan is much shorter than that of the switch cabinet itself, which can easily cause the switch cabinet to discharge or arc, leading to accidents.

[0005] (2) Model simulation analysis is a multiphysics coupling analysis method based on computer simulation. It uses the finite element method and combines it with fluid dynamics to perform fully coupled analysis of electricity, magnetism, heat and flow, which can obtain detailed field distribution data and conduct parametric studies. However, the models of model simulation analysis are usually more complex, and the quality requirements of mesh generation are high. It cannot meet the balance between computational efficiency and accuracy, resulting in high computational costs for modeling and simulation.

[0006] (3) Empirical formula estimation is an estimation method based on industry standards and empirical formulas. This method mainly uses empirical formulas to derive, construct equivalent thermal resistance network models for calculation, or conduct analogy analysis with similar products. It is widely used in engineering. However, the estimation accuracy of empirical formula estimation is limited, the estimated value is conservative, it is not conducive to hot spot analysis and temperature rise prediction, and it cannot take into account complex internal structural details. Therefore, it has poor adaptability when applied to switchgear.

[0007] There is currently no effective solution to the problem of balancing accuracy and efficiency when performing hotspot analysis on switchgear in related technologies. Summary of the Invention

[0008] The present invention provides a method for hotspot analysis of switchgear, a method for heat dissipation design, and an electronic device, which at least solves the problem of difficulty in balancing accuracy and efficiency when performing hotspot analysis of switchgear in related technologies.

[0009] This invention provides a method for analyzing hotspots in switchgear, comprising: calculating the fitness of each sample individual based on an objective function; storing the sample individuals and their corresponding fitness in a database in descending order of fitness; wherein the objective function is set based on the hotspot temperature of the switchgear, and the sample individuals are obtained by sampling the objective variable within a sampling range, the upper and lower limits of which are determined based on the parameters of the switchgear and the heat dissipation components; and sorting the top N samples in the database. P Using a sample population as the parent population, an offspring population is generated based on an improved mountain gazelle optimizer, an incremental kriging model, and the parent population. The improved mountain gazelle optimizer generates the offspring population based on a sinusoidal scaling factor and an adaptive distribution. The initial model of the incremental kriging model is constructed based on the sample individuals and their fitness, and is used to predict the fitness of the offspring individuals. The optimal candidate individuals in the offspring population are determined based on the incremental kriging model and the expected improvement criterion, and the evaluation count is incremented by one, with an initial value of zero. If the evaluation count has not reached the preset maximum, the database, parent population, and incremental kriging model are updated based on the optimal candidate individuals, and new optimal candidate individuals are determined. If the evaluation count has reached the maximum, the optimal solution for the objective variable is determined based on the new optimal candidate individuals, which is used to determine the optimal location of the heat dissipation component.

[0010] Preferably, the offspring population is generated based on an improved mountain gazelle optimizer, an incremental kriging model, and a parent population, including: calculating a sinusoidal scaling factor based on the current generation count of an individual and a preset total generation count; generating territorial solitary males, females, single males, and migrating foraging individuals based on the sinusoidal scaling factor and the parent population, incrementing the current generation count of each individual by one; after each generation of territorial solitary males, females, single males, and migrating foraging individuals, perturbing the current optimal individual based on an adaptive distribution to obtain updated individuals, where the current optimal individual refers to the parent population. The population, including all solitary males, females, single males, and migratory foraging individuals currently identified, is used to determine the individual with the highest fitness. If the current generation count of an individual is less than the total generation count, the sine scaling factor is updated based on the current generation count, and new solitary males, females, single males, migratory foraging individuals, and new individuals are identified. If the current generation count of an individual equals the total generation count, all solitary males, females, single males, migratory foraging individuals, and new individuals are considered as the offspring population.

[0011] Preferably, the formula for calculating the sinusoidal scaling factor based on the current number of times an individual is generated and the preset total number of times an individual is generated is as follows: ; In the formula, This represents the sine scaling factor. This indicates the current number of times an individual has been generated. This represents the total number of times an individual is generated. The formula for generating territorial solitary males, females, single males, and migrating foraging individuals based on the sine scaling factor and the parent population is as follows: ; In the formula, This indicates a solitary male individual residing in a territory. Indicates a female individual. Indicates a single male individual. Indicates migrating individuals in search of food. This represents the current optimal individual during the calculation process. , , , , , , Both represent random integers 1 or 2. This represents the coefficient vector of young male individuals. Indicates the first The parent population of the next iteration. This represents a perturbation vector that changes nonlinearly. , , and Both represent randomly selected coefficient vectors. This refers to a parent individual randomly selected from the parent population. express Dimensional decision variables, This indicates the upper limit of the computational problem. This indicates the lower bound of the computation problem.

[0012] Preferably, the formula for perturbing the current optimal individual based on the adaptive distribution to obtain the updated individual is expressed as follows: ; In the formula, Indicates updating the individual. This represents the current optimal individual. Describing the degrees of freedom as Adaptive distribution, This represents a row vector where all dimensions are 1; where the degrees of freedom are... It is determined using a piecewise function; adaptive distribution. probability density function for: ; In the formula, Represents the gamma function. Represents an adaptive distribution A random variable.

[0013] Preferably, degrees of freedom The calculation formula is: ; In the formula, Indicates the number of iterations. Indicates the threshold number of iterations. This indicates the maximum number of iterations.

[0014] Preferably, the optimal candidate individual in the offspring population is determined based on the incremental kriging model and the expected improvement criterion, and the number of evaluations is incremented by one. This includes: predicting the fitness of each offspring individual in the offspring population based on the incremental kriging model; determining the optimal candidate individual based on the fitness of each offspring individual in the offspring population and the expected improvement criterion; and performing a true evaluation on the optimal candidate individual based on the objective function to obtain the true fitness of the optimal candidate individual, and the number of evaluations is incremented by one.

[0015] Preferably, the number of optimal candidate individuals is greater than or equal to 1 and less than or equal to N. P The database, parent population, and incremental Kriging model are updated based on the optimal candidate individuals, including: comparing the true fitness of the optimal candidate individuals with the fitness of the sample individuals; storing the optimal candidate individuals and their corresponding true fitness in the database in order; incrementing the number of sample individuals in the database to obtain the updated database; and sorting the top N individuals in the updated database. P The sample individuals are used as the updated parent population; the incremental Kriging model is updated based on the best candidate individuals and their corresponding true fitness.

[0016] Preferably, updating the incremental kriging model based on the optimal candidate individual and the corresponding true fitness includes: reusing the historical parameters of the incremental kriging model; and updating the incremental kriging model by adding an extension term to the Gaussian correlation function matrix of the incremental kriging model based on the historical parameters, the optimal candidate individual and the corresponding true fitness, using the method of partition matrix inversion.

[0017] The present invention provides a heat dissipation design method for switchgear, which determines the optimal solution of the target variable based on any of the above methods; and designs the installation position of the heat dissipation component on the switchgear based on the optimal solution for heat dissipation of the switchgear.

[0018] An embodiment of the present invention provides an electronic device comprising: a processor, and a memory storing a program, the program including instructions which, when executed by the processor, cause the processor to perform any of the methods described above.

[0019] This invention provides a method for hotspot analysis of switchgear, a heat dissipation design method, and an electronic device. By improving the mountain gazelle optimizer through a sinusoidal scaling factor and adaptive distribution, the generated offspring individuals can effectively diffuse within the search space, contributing to improved analysis accuracy. Based on an incremental kriging model, the fitness of offspring individuals is predicted during the generation of the offspring population and the selection of optimal candidate individuals, significantly improving analysis efficiency. Even when the number of evaluations does not reach the preset maximum, updating the database, parent population, and incremental kriging model based on the optimal candidate individuals does not require global re-optimization; instead, historical information is reused along with incremental expansion, i.e., local parameter fine-tuning, further improving analysis efficiency. This addresses the difficulty in balancing accuracy and efficiency in switchgear hotspot analysis using related technologies. Attached Figure Description

[0020] To more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the accompanying drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are merely some embodiments of the present invention, and those skilled in the art can obtain other embodiments based on these drawings without creative effort.

[0021] Figure 1 This is a flowchart illustrating the steps of a switchgear hotspot analysis method in an embodiment of the present invention.

[0022] Figure 2 This is a schematic diagram of the variation curve of the sinusoidal scaling factor in an embodiment of the present invention.

[0023] Figure 3 This is a schematic diagram comparing the distribution of basic offspring individuals with and without the introduction of a sinusoidal scaling factor in an embodiment of the present invention.

[0024] Figure 4 This is a schematic diagram of the probability density curves of the adaptive distribution under different degrees of freedom in an embodiment of the present invention.

[0025] Figure 5 This is a schematic diagram of the overall model of the switch cabinet in an embodiment of the present invention.

[0026] Figure 6 This is a schematic diagram of the optimizable domain of the fan on the back of the switch cabinet in an embodiment of the present invention.

[0027] Figure 7This is a schematic diagram of the algorithm search history in an embodiment of the present invention.

[0028] Figure 8 This is a schematic diagram of the contour temperature distribution of historical search points in an embodiment of the present invention.

[0029] Figure 9 This is a schematic diagram of the average temperature change curves corresponding to the parent population individuals in an embodiment of the present invention.

[0030] Figure 10 This is a schematic diagram of the local optimization convergence curve in an embodiment of the present invention.

[0031] Figure 11 This is a schematic diagram of the structure of an electronic device according to an embodiment of the present invention. Detailed Implementation

[0032] Embodiments of the present invention will now be described in more detail with reference to the accompanying drawings. While some embodiments of the present invention are shown in the drawings, it should be understood that the present invention can be implemented in various forms and should not be construed as limited to the embodiments set forth herein. Rather, these embodiments are provided to provide a more thorough and complete understanding of the present invention. It should be understood that the drawings and embodiments of the present invention are for illustrative purposes only and are not intended to limit the scope of protection of the present invention.

[0033] Switchgear is widely used in power transmission, serving functions such as control and protection. However, due to its compact internal structure, the heat generated during operation is difficult to dissipate effectively, especially in high-current switchgear with a rated operating current ≥630A, which is prone to overheating, potentially leading to major accidents. Therefore, conducting hotspot analysis on switchgear for heat dissipation design, thereby reducing the risks during switchgear operation, is a crucial foundation for the stable operation of power systems.

[0034] Related hot topic analysis methods, such as experimental measurement, model simulation analysis, and empirical formula estimation, all struggle to balance analytical accuracy and efficiency.

[0035] This invention addresses the complexities of hotspot analysis in high-current switchgear, noting that existing methods are inaccurate, time-consuming, and computationally expensive. Furthermore, inaccurate hotspot analysis can easily lead to overestimation of the cooling capacity of components like fans during heat dissipation design. Since hotspot analysis (temperature analysis) and heat dissipation design in high-current switchgear fall under the category of expensive optimization problems (where the objective function calculation relies on computer simulation models or physical experiments and incurs high computational costs), efficiently solving these expensive optimization problems using optimization algorithms can reduce the cycle and cost of switchgear simulation design, effectively lowering the internal temperature of high-current switchgear.

[0036] Next, we consider using evolutionary algorithms and Kriging surrogate model-assisted evolutionary algorithms to solve expensive optimization problems. Evolutionary algorithms simulate the process of natural selection, generating offspring populations through the reproduction of parent populations, calculating the fitness of individuals in the offspring population, and retaining individuals with high fitness for the next generation. This utilizes population diversity to perform global exploration of computationally expensive engineering optimization problems. Kriging surrogate model-assisted evolutionary algorithms incorporate a Kriging surrogate model into the evolutionary algorithm. This surrogate model assists in calculating candidate solutions for offspring, selects the optimal candidate solution, adds it to a historical database, and updates the surrogate model, thus reducing computational costs while optimizing the problem.

[0037] However, while evolutionary algorithms are simple and easy to implement, requiring only input variables and constraints to solve problems, they are prone to the curse of dimensionality when dealing with high-dimensional optimization problems. Furthermore, using the true function to solve for each individual also incurs high computational costs. While Kriging surrogate models assist evolutionary algorithms by using model management strategies to select the optimal candidate individuals and reduce computational costs, the modeling time cost gradually increases with the number of historical samples, and the time required for later optimization modeling becomes excessive.

[0038] Therefore, please refer to Figure 1 As shown, the present invention provides a switchgear hotspot analysis method by improving the mountain gazelle optimization algorithm (mountain gazelle optimizer) in evolutionary algorithms and adjusting the update mechanism of the Kriging agent auxiliary model (hereinafter referred to as Kriging model) (to obtain an incremental Kriging model), including steps S101 to S105, which can solve the problem of difficulty in balancing accuracy and efficiency when performing switchgear hotspot analysis in related technologies.

[0039] Step S101: Calculate the fitness of each sample individual based on the objective function, and store the sample individuals and their corresponding fitness in the database in descending order of fitness. The objective function is set based on the hot spot temperature of the switch cabinet, and the sample individuals are obtained by sampling the target variable within the sampling range. The upper and lower limits of the sampling range are determined based on the parameters of the switch cabinet and the heat dissipation components.

[0040] Step S102: Sort the database by the first N items. P Using a sample of individuals as the parent population, an offspring population is generated based on an improved mountain gazelle optimizer, an incremental kriging model, and the parent population. The improved mountain gazelle optimizer is used to generate the offspring population based on a sinusoidal scaling factor and an adaptive distribution. The initial model of the incremental kriging model is constructed based on the sample individuals and their fitness. The incremental kriging model is used to predict the fitness of the offspring individuals.

[0041] Step S103: Based on the incremental Kriging model and the expected improvement criterion, determine the optimal candidate individuals in the offspring population, and increment the evaluation count by one, where the initial value of the evaluation count is zero.

[0042] Step S104: If the number of evaluations has not reached the preset maximum number of evaluations, update the database, parent population and incremental Kriging model based on the optimal candidate individual, and determine a new optimal candidate individual.

[0043] Step S105: When the number of evaluations reaches the maximum number of evaluations, determine the optimal solution of the target variable based on the new optimal candidate individuals, which is used to determine the optimal position of the heat dissipation component.

[0044] The objective function is set based on the hot spot temperature of the switchgear. The hot spot temperature can be, but is not limited to, the maximum temperature, temperature standard deviation, temperature margin, temperature-energy consumption composite function, or average steady-state temperature. This embodiment will subsequently use the average steady-state temperature of the hot spot area of ​​the busbar inside the switchgear as an example to illustrate the objective function.

[0045] The target variable can be, but is not limited to, the installation location of the heat dissipation component on the switch cabinet. This embodiment will subsequently use the coordinates of the center point of a fan mounted on the back of the switch cabinet as an example to illustrate the specifics.

[0046] The methods for sampling the target variable within the sampling range can include, but are not limited to, Latin hypercube sampling, low-discrepancy sequence sampling, and orthogonal experimental design sampling based on orthogonal arrays. In this embodiment, the target variable is sampled within the sampling range using Latin hypercube sampling.

[0047] The fitness of each sample individual is calculated based on the objective function. The sample individuals and their corresponding fitness values ​​are then stored in a database in descending order of fitness, thus constructing the initial database. The fitness values ​​calculated in this process represent the true values, laying the foundation for selecting high-quality parent populations and constructing an accurate and effective initial model for the incremental kriging model. In this process, sample individuals with higher fitness represent lower hotspot temperatures.

[0048] Incremental Kriging models predict the fitness of offspring individuals during the generation of offspring populations and the selection of optimal candidate individuals. In other words, the fitness of offspring individuals calculated in this process is a predicted value, which can greatly improve the efficiency of analysis.

[0049] When the number of evaluations does not reach the preset maximum number of evaluations, when updating the database, parent population, and incremental Kriging model based on the best candidate individuals, there is no need to perform global optimization again. Instead, historical information is reused and incremental expansion is used to fine-tune the parameters locally, which can further improve the analysis efficiency.

[0050] The fitness of the best candidate individual used to update the database is based on the true value calculated by the objective function, which is the same as the fitness value used when storing sample individuals in the database, thus helping to improve the accuracy of the analysis.

[0051] Furthermore, the completion of one evaluation is equivalent to the completion of one global iteration. In the subsequent description of the algorithm in this embodiment, we will primarily use the term "number of iterations" rather than "number of evaluations," but in this embodiment, the two can be understood as equivalent. The specific value of the maximum number of evaluations (maximum number of iterations) should be adjusted according to the actual application scenario and requirements, and can be determined by those skilled in the art through a limited number of experiments.

[0052] In one embodiment of this invention, step S102, generating a progeny population based on an improved mountain gazelle optimizer, an incremental kriging model, and a parent population, includes: calculating a sinusoidal scaling factor based on the current generation count of an individual and a preset total generation count of individuals. Setting the sinusoidal scaling factor is to limit the diffusion of the progeny population during generation, reduce the number of progeny individuals at the variable boundary, and help improve the accuracy of the analysis.

[0053] Based on the sinusoidal scaling factor and the parent population, solitary male individuals, female individuals, single male individuals, and migrating foraging individuals are generated, with the current generation count of each individual incremented by one.

[0054] After each generation of solitary males, females, lone males, and migratory foraging individuals, the current best individual is perturbed based on an adaptive distribution to obtain updated individuals. The current best individual refers to the individual with the highest fitness among all solitary males, females, lone males, and migratory foraging individuals obtained up to the current time, from the parent population. The adaptive distribution perturbation is used to increase the diversity of offspring individuals, which helps improve the accuracy of the analysis.

[0055] If the current number of times an individual is generated is less than the total number of times an individual is generated, the sine scaling factor is updated based on the current number of times an individual is generated, and new territorial solitary male individuals, female individuals, single male individuals, migrating foraging individuals, and updated individuals are determined.

[0056] When the current generation count of an individual equals the total generation count, all territorial solitary males, females, single males, migrating foraging individuals, and newly generated individuals are considered as the offspring population. Therefore, the offspring population includes territorial solitary males, females, single males, migrating foraging individuals, and newly generated individuals.

[0057] Preferably, the formula for calculating the sinusoidal scaling factor based on the current number of times an individual is generated and the preset total number of times an individual is generated is as follows: ; In the formula, This represents the sine scaling factor. This indicates the current number of times an individual has been generated. This represents the total number of times an individual has been generated.

[0058] For example, when the total number of individual generation times is 50, the curve of the change of the sinusoidal scaling factor is as follows: Figure 2 As shown. Figure 2 In the equation, a value greater than 0 indicates the degree to which the offspring generation area shrinks, while a value less than 0 indicates the degree to which the offspring generation area expands. Both shrinkage and expansion refer to the extension range around the optimal solution. The sinusoidal scaling factor, with a non-linear decreasing and increasing trend, restricts the four types of basic offspring individuals (territorial solitary males, females, single males, and migratory foraging individuals) to the search range and allows them to cover a wider area in space.

[0059] The formula for generating territorial solitary males, females, single males, and migratory foraging individuals based on a sinusoidal scaling factor and the parent population is as follows: ; In the formula, This indicates a solitary male individual residing in a territory. Indicates a female individual. Indicates a single male individual. Indicates migrating individuals in search of food. This represents the current optimal individual during the calculation process. , , , , , , Both represent random integers 1 or 2. This represents the coefficient vector of young male individuals. Indicates the first The parent population of the next iteration. This represents a perturbation vector that changes nonlinearly. , , and Both represent randomly selected coefficient vectors. This refers to a parent individual randomly selected from the parent population. express Dimensional decision variables, This indicates the upper limit of the computational problem. This represents the lower bound of the computation problem. (Except for the sine scaling factor.) In addition, the specific settings of the other variables in the above formula can be referred to the settings of the mountain gazelle optimizer, which will not be repeated here in this embodiment.

[0060] For example, please refer to Figure 3 As shown, the blue solid circle represents the basic offspring distribution generated by the mountain gazelle optimizer without introducing a sinusoidal scaling factor, while the red solid circle represents the basic offspring distribution generated by the mountain gazelle optimizer after introducing a sinusoidal scaling factor.

[0061] For example, the average distribution after 30 population updates (each update generates a total of 50 offspring, meaning one parent individual generates 200 basic offspring individuals) is as follows: without the sinusoidal scaling factor, the number of basic offspring individuals generated on the boundary is 79, and the number of basic offspring individuals generated in the interior space is 121; after the sinusoidal scaling factor is introduced, the number of basic offspring individuals generated on the boundary is 53, and the number of basic offspring individuals generated in the interior space is 147.

[0062] It can be seen that after introducing the sinusoidal scaling factor, the number of basic offspring individuals generated on the boundary decreased from 79 to 53, which improved the sample distribution within the design space.

[0063] Preferably, the formula for perturbing the current optimal individual based on the adaptive distribution to obtain the updated individual is expressed as follows: ; In the formula, Indicates updating the individual. This represents the current optimal individual. Describing the degrees of freedom as Adaptive distribution, This represents a row vector where all dimensions are 1. The degrees of freedom are... It is determined using a piecewise function.

[0064] In this embodiment, a probability density distribution within the interval [-1, 1] is used to perturb the current optimal individual in order to quickly find the updated individual obtained after perturbation, i.e., an adaptive distribution. The probability of conforming to the corresponding distribution is obtained by taking random values ​​in the range [-1, 1], which is used to represent the amplitude of the disturbance.

[0065] Adaptive distribution probability density function for: ; In the formula, Represents the gamma function. Represents an adaptive distribution A random variable.

[0066] Specifically, adaptive distribution under different degrees of freedom The probability density curve is as follows Figure 4As shown. By Figure 4 It can be seen that the probability density is relatively high in the interval [-1,1], which represents the local development capability; the probability density outside the interval [-1,1] represents the global search capability.

[0067] Furthermore, to prevent getting trapped in local optima during the iteration process, this embodiment uses the following piecewise function to calculate the degrees of freedom. : ; In the formula, Indicates the number of iterations. Indicates the threshold number of iterations. This indicates the maximum number of iterations. The iteration threshold is used to distinguish between the early and late stages of iteration. Both the iteration threshold and the maximum number of iterations are preset and can be determined by those skilled in the art based on actual circumstances and a limited number of experiments.

[0068] In the early stages of iteration, the degrees of freedom increase linearly, while in the later stages, a logarithmic function is used for non-linear increases. This allows for a significant perturbation of the generated current best individual, helping it escape local optima. Here, "current best individual" refers to the current state at each iteration.

[0069] In one embodiment of the present invention, step S103, determining the optimal candidate individual in the offspring population based on the incremental kriging model and the expected improvement criterion, and incrementing the evaluation count by one, includes: predicting the fitness of each offspring individual in the offspring population based on the incremental kriging model; determining the optimal candidate individual based on the fitness of each offspring individual in the offspring population and the expected improvement criterion; performing a true evaluation on the optimal candidate individual based on the objective function to obtain the true fitness of the optimal candidate individual, and incrementing the evaluation count by one.

[0070] The expression for the expected improvement criterion is: ; In the formula, Indicates the expected value. This represents the current minimum objective function value for all samples in the database. The cumulative distribution function represents the standard normal distribution. The density function representing the standard normal distribution. This represents the optimal linear unbiased prediction (fitness prediction) of the incremental Kriging model. This represents the square root of the estimation error obtained from the incremental Kriging model prediction.

[0071] The expected value of each offspring individual is calculated using the above-mentioned expected improvement criterion expression, and the offspring individual with the largest expected value is selected as the optimal candidate individual.

[0072] Specifically, the number of optimal candidate individuals is greater than or equal to 1 and less than or equal to N. P .

[0073] The database, parent population, and incremental Kriging model are updated based on the optimal candidate individuals, including: comparing the true fitness of the optimal candidate individuals with the fitness of the sample individuals; storing the optimal candidate individuals and their corresponding true fitness values ​​in the database in order; incrementing the number of sample individuals in the database to obtain the updated database; and sorting the top N individuals in the updated database. P The sample individuals are used as the updated parent population; the incremental Kriging model is updated based on the best candidate individuals and their corresponding true fitness.

[0074] Specifically, the incremental kriging model is updated based on the optimal candidate individuals and their corresponding true fitness, including: reusing the historical parameters of the incremental kriging model. Based on the historical parameters, the optimal candidate individuals, and their corresponding true fitness, the incremental kriging model is updated by adding an extension term to the Gaussian correlation function matrix of the incremental kriging model using the partition matrix inversion method.

[0075] This embodiment provides a detailed comparison between the relevant Kriging model and the aforementioned incremental Kriging model.

[0076] The relevant Kriging model approximates a single objective function as: ; In the formula, It is an input variable. It is the average value of the Gaussian process. It has a mean of 0 and a variance of And related functions exist and A Gaussian distribution between them. Where, Indicates the first The Gaussian process response function values ​​corresponding to each sample. Indicates the first The Gaussian process response function values ​​corresponding to each sample. , , (yes The number of variables in a problem is called a hyperparameter.

[0077] Related functions Calculated using the Gaussian correlation function: ; Assuming there are n sampling points and their objective function values The maximum likelihood function is used to estimate the hyperparameters: ; ; ; In the formula, express 3D real space, It is The correlation matrix, yes unit vector, This indicates transpose.

[0078] Using the estimated hyperparameters, the optimal linear unbiased prediction is obtained as follows: ; The estimation error obtained from the prediction is: ; The biggest problem with the correlation kriging model is that it is very time-consuming to build, with the time mainly spent on the correlation matrix. The inverse of the expression has a time complexity of O(n^2). 3 Using Cholesky decomposition generally yields numerically more accurate results than directly inverting the correlation matrix, and its time complexity can be reduced from O(n^2). 3 Reduced to O(n) 2 However, as historical data increases, the relevant Kriging models remain very time-consuming.

[0079] ; Suppose we have already built a correlation kriging model with n samples, and then we take N samples from these n samples. P To construct a parent population from 1 sample and then update the Kriging model with q optimal candidate individuals obtained after simulating natural selection, an incremental expression needs to be constructed.

[0080] In the incremental kriging model provided in this embodiment, The increment expression is: ; The increment expression is: ; The increment expression is: ; In the formula, These are the hyperparameters of the relevant kriging model (initial model), which will be used throughout the optimization process. The new correlation matrix is ​​represented as: ; In the formula, This is the correlation matrix of the relevant Kriging model (initial model).

[0081] The inverse operation of the partition matrix is: ; In the formula, , It was computed and stored before updating the incremental Kriging model. It is Matrix and The time complexity of the calculation is O(q) 3 ).in, (1≤ ≤N P The time complexity is O(n) because the number of training samples is much smaller than the number of training samples n. 2 Compared to the previous example, the time complexity of updating the incremental Kriging model in this embodiment is O(q). 3 (This can be ignored.)

[0082] In other words, compared to the relevant kriging model, the incremental kriging model provided in this embodiment has lower time complexity, and only q new samples are needed each time the model is updated. As historical data increases (the number of individual samples and their fitness in the database increases), the time spent updating the incremental kriging model is significantly reduced compared to rebuilding the relevant kriging model. Simply put, for problems with large datasets, the incremental kriging model provided in this embodiment requires less time.

[0083] This invention also provides a heat dissipation design method for switchgear, which determines the optimal solution of the target variable based on any of the aforementioned switchgear hotspot analysis methods; and designs the installation position of the heat dissipation components on the switchgear based on the optimal solution for heat dissipation. This method solves the problem of balancing accuracy and efficiency when performing hotspot analysis and heat dissipation design for high-current switchgear in related technologies.

[0084] By employing an incremental Kriging agent-assisted evolutionary algorithm based on an improved mountain gazelle optimizer, this method efficiently solves the expensive optimization problems of hotspot analysis and heat dissipation design for high-current switchgear. It significantly reduces reliance on high-cost simulation models while maintaining powerful optimization capabilities, thereby effectively reducing time and computational costs and shortening the design cycle. It can also provide important technical support for similar engineering simulation optimization challenges.

[0085] For example, the embodiments of the present invention also verify the technical effect of the above-mentioned solution through specific simulation experiments, taking the optimization of the position of the switch cabinet fan as an example of the corresponding expensive optimization problem, as follows.

[0086] In this experiment, the switchgear cooling system adopts a forced air cooling scheme. The cooling component is an axial fan, whose structure mainly consists of fan blades, a drive motor, and a housing. During actual operation, the motor drives the fan blades to rotate, generating thrust to make air flow axially, thereby expelling the heat accumulated inside the switchgear. The design preferably adopts an exhaust layout, with the fan installed at the rear air outlet of the switchgear. When the temperature inside the cabinet rises, the fan starts to extract the hot air, while external cool air enters through the air inlet under negative pressure, forming a convection channel.

[0087] The main heat source inside the switchgear is Joule heat generated by the contact resistance of the busbars and contacts. The total power consumption of the system is... Calculated using the following formula: ; In the formula, Indicates the rated operating current of the switchgear. This represents the total resistance of the circuit.

[0088] According to the heat balance equation, under the set allowable temperature rise Under the premise that the required cooling air volume Q is calculated, the formula is: ; In the formula, This indicates the specific heat capacity of air. This indicates air density.

[0089] After completing the above fan selection, the maximum airflow of the selected fan and The -Q characteristic curve is used as a boundary condition input into the COMSOL finite element simulation model to ensure that the simulation environment can realistically reproduce the internal flow field distribution of the switchgear and verify whether the maximum temperature at the contact and busbar overlap is lower than the national standard under the Q airflow. By optimizing the switchgear structure, the aim is to solve the pain points of high-current switchgear with compact structure and high air resistance.

[0090] The mathematical model for optimizing fan position is as follows: ; In the formula, The cost function is represented by the average steady-state temperature of the heating point area of ​​the busbar inside the switch cabinet in this experiment. This represents the optimization variable (target variable). Represents the two-dimensional plane mounting coordinates of the fan. and These represent the first and second coordinates of the fan center, respectively.

[0091] Specifically, optimizing variables Subject to the physical boundaries of the installation space Constraints, among which, This represents the lower limit of the coordinates for permissible fan installation. This represents the upper limit of the coordinates at which the fan is allowed to be installed.

[0092] objective function for: ; In the formula, It is an implicit function representing a complex calculation process, with fan coordinates as input. The output is the volume average temperature value of the heating point region of the busbar, calculated through COMSOL electro-thermal-fluid coupling analysis. The purpose of this optimization is to find a set of coordinates while satisfying the boundary condition constraints. This makes the average steady-state temperature value Minimize and install the high static pressure axial fan at coordinates determined by the optimization algorithm. This allows for the construction of a switch cabinet entity with an optimal heat dissipation path.

[0093] The algorithm relies on COMSOL to build fan and switch cabinet models with real physical data. Using the COMSOLMultiphysics 6.3 with MATLAB interface, a mapping between structural parameters and physical field data is established between the algorithm and the model. COMSOL, as the core of the multiphysics solution, can accurately calculate the coupling effect of complex flow and temperature fields caused by changes in fan coordinates. This data interaction not only improves the efficiency of optimization iterations but also ensures that the fan coordinates remain consistent in each optimization round. All adjustments are strictly based on the physical laws governing the relationship between real fluid dynamics and heat transfer.

[0094] Overall model of the switch cabinet as follows Figure 5 As shown, the square part on the back of the switch cabinet, which is the bold black part, is the target fan for optimization. The remaining part includes the switch cabinet body and busbars.

[0095] from Figure 5 Viewing the switch cabinet from the back, and rotating the switch cabinet 90 degrees counterclockwise, we obtain... Figure 6 .

[0096] exist Figure 6 In the diagram, the center point coordinates of the fan are set as constraints, with the upper and lower limits of the X coordinate being [400, 1800] and the upper and lower limits of the Y coordinate being [200, 600]. The red line shows the optimizable domain of the fan on the back of the switch cabinet.

[0097] By driving COMSOL to perform automated iterative simulation in the algorithm program, the hot spot temperature of the busbar obtained from each generation of simulation is used to generate the fan coordinates of new shape parameters after the algorithm learns. The optimal solution of the fan position coordinates is found through a certain number of iterations.

[0098] The experiment verified the method provided by the embodiments of the present invention by optimizing the fan position of the switch cabinet. The optimization results are as follows: Figures 7 to 10 As shown.

[0099] from Figure 7 Looking at the search history graph, the points evaluated by the algorithm, i.e., the blue points, are relatively evenly distributed within the effective search space of [400, 1800]. The optimal solution, i.e., the red pentagram, is displayed near [X=1150, Y=410]. This indicates that within a limited number of iterations, the algorithm not only calculated the coordinates of the points within the effective search space relatively evenly, but also explored the vicinity of the optimal solution multiple times (there are multiple blue points near the red pentagram), which is consistent with the algorithm design process and experimental requirements.

[0100] Figure 8 The contour temperature distribution of historical search points is shown. Black dots represent successful sample points, black crosses represent simulation failure points, and white pentagrams represent optimal points. It can be seen that the algorithm samples globally and performs more local optimizations in areas with lower temperatures, ultimately obtaining the optimal point with the minimum temperature.

[0101] Figure 9 The graph shows the average temperature change of the parent population individuals in the algorithm. The overall trend of the curve is downward. The average temperature of the first 10 initial parent population individuals is 49.6 degrees Celsius. Then, starting from the first iteration of local optimization, the average temperature curve of the first ten parent population individuals is plotted each time until the average temperature of the parent population individuals reaches 48.4 degrees Celsius at the end of the 22nd local optimization. This shows that the algorithm is not only looking for a minimum point, but also continuously replacing the old individuals in the parent population with new individuals with lower temperatures (better solutions replace old solutions in the parent population), so that the average temperature of the entire parent population is steadily decreasing (the level of the solution of the entire parent population is steadily improving).

[0102] For example, the algorithm presets a total number of iterations of 42 and a local optimization iteration number of 22. The total number of iterations is the number of complete closed-loop iterations of the global optimization, which is equal to the maximum number of evaluations. The number of local optimization iterations is the number of times the real model based on the expected improvement criterion is invoked within the local space. The real model refers to the model generated by COMSOL; that is, for each local iteration, COMSOL reconstructs the model based on the expected improvement criterion.

[0103] Figure 10To optimize the convergence curve locally, the method described in this embodiment can find a smaller average temperature value after 5 iterations of local optimization, and find the point with the lowest average temperature in the domain after a total of 38 iterations. The convergence speed is relatively fast, which shows that its efficient search performance can effectively solve the expensive optimization problem of COMSOL simulation.

[0104] The present invention also provides a non-transitory machine-readable medium storing a computer program, wherein the computer program, when executed by a computer's processor, is used to cause the computer to perform a method according to an embodiment of the present invention.

[0105] The present invention also provides a computer program product, including a computer program, wherein the computer program, when executed by a computer processor, is used to cause the computer to perform the switchgear hotspot analysis method provided in the present invention.

[0106] This invention also provides an electronic device, including: at least one processor; and a memory communicatively connected to the at least one processor. The memory stores a computer program executable by the at least one processor, which, when executed by the at least one processor, causes the electronic device to perform the switchgear hotspot analysis method provided in this invention.

[0107] refer to Figure 11 This is a structural block diagram of an electronic device, either a server or a client, according to an embodiment of the present invention. It is an example of a hardware device that can be applied to various aspects of the present invention. The electronic device is intended to represent various forms of digital electronic computer devices, such as laptop computers, desktop computers, workstations, personal digital assistants, servers, blade servers, mainframe computers, and other suitable computers. The electronic device can also represent various forms of mobile devices, such as personal digital processors, cellular phones, smartphones, wearable devices, and other similar computing devices. The components shown herein, their connections and relationships, and their functions are merely illustrative and are not intended to limit the implementation of the present invention described and / or claimed herein.

[0108] like Figure 11As shown, the electronic device includes a computing unit 1101, which can perform various appropriate actions and processes according to a computer program stored in a read-only memory (ROM) 1102 or a computer program loaded into a random access memory (RAM) 1103 from a storage unit 1108. The RAM 1103 may also store various programs and data required for the operation of the electronic device. The computing unit 1101, the ROM 1102, and the RAM 1103 are interconnected via a bus 1104. An input / output (I / O) interface 1105 is also connected to the bus 1104.

[0109] Multiple components in the electronic device are connected to I / O interface 1105, including: input unit 1106, output unit 1107, storage unit 1108, and communication unit 1109. Input unit 1106 can be any type of device capable of inputting information into the electronic device. Input unit 1106 can receive input digital or character information and generate key signal inputs related to user settings and / or function control of the electronic device. Output unit 1107 can be any type of device capable of presenting information and may include, but is not limited to, a display, speaker, video / audio output terminal, vibrator, and / or printer. Storage unit 1108 may include, but is not limited to, disks and optical discs. Communication unit 1109 allows the electronic device to exchange information / data with other devices through computer networks such as the Internet and / or various telecommunications networks, and may include, but is not limited to, modems, network cards, infrared communication devices, and / or wireless communication transceivers, such as Bluetooth devices, WiFi devices, WiMax devices, cellular communication devices, and / or the like.

[0110] The computing unit 1101 can be a variety of general-purpose and / or special-purpose processing components with processing and computing capabilities. Some examples of the computing unit 1101 include, but are not limited to, CPUs, graphics processing units (GPUs), various special-purpose artificial intelligence (AI) computing units, various computing units running machine learning model algorithms, digital signal processors (DSPs), and any suitable processor, controller, microcontroller, etc. The computing unit 1101 performs the various methods and processes described above. For example, in some embodiments, the method embodiments of the present invention can be implemented as computer programs tangibly contained in a machine-readable medium, such as storage unit 1108. In some embodiments, part or all of the computer program can be loaded and / or installed on an electronic device via ROM 1102 and / or communication unit 1109. In some embodiments, the computing unit 1101 can be configured by any other suitable means (e.g., by means of firmware) to perform the switchgear hotspot analysis method described above.

[0111] Computer programs for implementing the methods of embodiments of the present invention may be written in any combination of one or more programming languages. These computer programs may be provided to a processor or controller of a general-purpose computer, special-purpose computer, or other programmable data processing apparatus, such that when executed by the processor or controller, the computer programs cause the functions / operations specified in the flowcharts and / or block diagrams to be performed. The computer programs may be executed entirely on a machine, partially on a machine, or as a standalone software package, partially on a machine and partially on a remote machine, or entirely on a remote machine or server.

[0112] In the context of embodiments of this invention, a machine-readable medium can be a tangible medium that may contain or store a program for use by or in conjunction with an instruction execution system, apparatus, or device. A machine-readable medium can be a machine-readable signal medium or a machine-readable storage medium. A machine-readable signal medium may include, but is not limited to, electronic, magnetic, optical, electromagnetic, or infrared systems, apparatus, or devices, or any suitable combination of the foregoing. More specific examples of machine-readable storage media include electrical connections based on one or more wires, portable computer disks, hard disks, random access memory (RAM), read-only memory (ROM), erasable programmable read-only memory (EPROM or flash memory), optical fibers, portable compact disk read-only memory (CD-ROM), optical storage devices, magnetic storage devices, or any suitable combination of the foregoing.

[0113] It should be noted that the term "comprising" and its variations used in the embodiments of this invention are open-ended, meaning "including but not limited to". The term "based on" means "at least partially based on". The term "one embodiment" means "at least one embodiment"; the term "another embodiment" means "at least one additional embodiment"; the term "some embodiments" means "at least some embodiments". The modifications of "one" and "a plurality" mentioned in the embodiments of this invention are illustrative and not restrictive, and those skilled in the art should understand that unless explicitly indicated otherwise in the context, they should be understood as "one or more". The descriptions of terms such as "first", "second", etc., are for descriptive purposes only and should not be construed as indicating or implying their relative importance or implicitly specifying the number of indicated technical features.

[0114] The user information (including but not limited to user device information, user personal information, etc.) and data (including but not limited to data used for analysis, stored data, displayed data, etc.) involved in the embodiments of this invention are all information and data authorized by the user or fully authorized by all parties.

[0115] The steps described in the method embodiments provided by the present invention can be performed in different orders and / or in parallel. Furthermore, the method embodiments may include additional steps and / or omit the steps shown. The scope of protection of the present invention is not limited in this respect.

[0116] The term "embodiment" in this specification refers to a specific feature, structure, or characteristic described in connection with an embodiment that may be included in at least one embodiment of the invention. The appearance of this phrase in various places throughout the specification does not necessarily imply the same embodiment, nor does it imply independence or alternativeity from other embodiments. The various embodiments in this specification are described in a related manner, with reference to each other for similar or identical parts. In particular, for apparatus, device, and system embodiments, since they are substantially similar to method embodiments, the description is relatively simple, and relevant details are referred to in the description of the method embodiments.

[0117] The above-described embodiments are merely illustrative of several implementations of the present invention, and while the descriptions are specific and detailed, they should not be construed as limiting the scope of protection. It should be noted that those skilled in the art can make various modifications and improvements without departing from the inventive concept of the present invention, and these modifications and improvements all fall within the scope of protection of the present invention. Therefore, the scope of protection of the present invention should be determined by the appended claims.

Claims

1. A method for analyzing hotspots in switchgear, characterized in that, include: The fitness of each sample individual is calculated based on the objective function. The sample individuals and their corresponding fitness are stored in the database in descending order of fitness. The objective function is set based on the hot spot temperature of the switch cabinet. The sample individuals are obtained by sampling the target variable within the sampling range. The upper and lower limits of the sampling range are determined based on the parameters of the switch cabinet and the heat dissipation components. Sort the database into the top N P A sample of individuals is used as the parent population. An offspring population is generated based on the improved mountain gazelle optimizer, the incremental kriging model, and the parent population. The improved mountain gazelle optimizer is used to generate the offspring population based on a sinusoidal scaling factor and an adaptive distribution. The initial model of the incremental kriging model is constructed based on the sample individuals and their fitness. The incremental kriging model is used to predict the fitness of the offspring individuals. Based on the incremental Kriging model and the expected improvement criterion, the optimal candidate individual in the offspring population is determined, and the number of evaluations is incremented by one, wherein the initial value of the number of evaluations is zero; If the number of evaluations does not reach the preset maximum number of evaluations, the database, the parent population, and the incremental Kriging model are updated based on the optimal candidate individual, and a new optimal candidate individual is determined. If the number of evaluations reaches the maximum number of evaluations, the optimal solution of the target variable is determined based on the new optimal candidate individuals, which is used to determine the optimal position of the heat dissipation component.

2. The method according to claim 1, characterized in that, The offspring population is generated based on the improved mountain gazelle optimizer, the incremental kriging model, and the parent population, including: The sinusoidal scaling factor is calculated based on the current number of times an individual is generated and the preset total number of times an individual is generated; Based on the sinusoidal scaling factor and the parent population, solitary male individuals, female individuals, single male individuals, and migrating foraging individuals are generated, and the current generation count of each individual is incremented by one. After each generation of the solitary male, female, single male, and migratory foraging individual, the current best individual is perturbed based on the adaptive distribution to obtain an updated individual. The current best individual refers to the individual with the highest fitness among all solitary males, females, single males, and migratory foraging individuals obtained up to the present time in the parent population. If the current number of times an individual is generated is less than the total number of times an individual is generated, the sinusoidal scaling factor is updated based on the current number of times an individual is generated, and new territorial solitary male individuals, female individuals, single male individuals, migrating foraging individuals, and updated individuals are determined; If the current generation count of an individual is equal to the total generation count of that individual, all territorial solitary males, females, single males, migrating foraging individuals, and renewing individuals are considered as the offspring population.

3. The method according to claim 2, characterized in that, The formula for calculating the sinusoidal scaling factor based on the current number of times an individual is generated and the preset total number of times an individual is generated is as follows: ； In the formula, Indicates the sine scaling factor. Indicates the current number of times an individual has been generated. This represents the total number of times an individual was generated. The formula for generating territorial solitary males, females, single males, and migratory foraging individuals based on the sinusoidal scaling factor and the parent population is as follows: ； In the formula, This indicates a solitary male individual residing in a territory. Indicates a female individual. Indicates a single male individual. Indicates migrating individuals in search of food. This represents the current optimal individual during the calculation process. , , , , , , Both represent random integers 1 or 2. This represents the coefficient vector of young male individuals. Indicates the first The parent population of the next iteration. This represents a perturbation vector that changes nonlinearly. , , and Both represent randomly selected coefficient vectors. This refers to a parent individual randomly selected from the parent population. express Dimensional decision variables, This indicates the upper limit of the computational problem. This indicates the lower bound of the computation problem.

4. The method according to claim 2, characterized in that, The formula for perturbing the current optimal individual based on the adaptive distribution to obtain the updated individual is expressed as follows: ； In the formula, Indicates updating an individual. This represents the current optimal individual. Describing the degrees of freedom as Adaptive distribution, This represents a row vector where all dimensions are 1; Among them, degrees of freedom It is determined using a piecewise function. Adaptive distribution probability density function for: ； In the formula, Represents the gamma function. Represents an adaptive distribution A random variable.

5. The method according to claim 4, characterized in that, Degrees of freedom The calculation formula is: ； In the formula, Indicates the number of iterations. Indicates the threshold number of iterations. This indicates the maximum number of iterations.

6. The method according to claim 1, characterized in that, Based on the incremental Kriging model and the expected improvement criterion, the optimal candidate individuals in the offspring population are determined, and the number of evaluations is incremented by one, including: The fitness of each offspring individual in the offspring population is predicted based on the incremental kriging model. The optimal candidate individual is determined based on the fitness of each offspring individual in the offspring population and the expected improvement criterion; The optimal candidate individual is evaluated based on the objective function to obtain the true fitness of the optimal candidate individual, and the evaluation count is incremented by one.

7. The method according to claim 6, characterized in that, The number of optimal candidate individuals is greater than or equal to 1 and less than or equal to N. P ; The database, the parent population, and the incremental kriging model are updated based on the optimal candidate individuals, including: The true fitness of the optimal candidate individual is compared with the fitness of the sample individual. The optimal candidate individual and its corresponding true fitness are stored in the database in order. The number of sample individuals in the database is incremented by one to obtain an updated database. Sort the updated database in the top N order. P One sample individual is used as the updated parent population; The incremental Kriging model is updated based on the optimal candidate individual and its corresponding true fitness.

8. The method according to claim 7, characterized in that, The incremental Kriging model is updated based on the optimal candidate individuals and their corresponding true fitness, including: Reuse the historical parameters of the incremental kriging model; Based on the historical parameters, the optimal candidate individuals, and their corresponding true fitness, the incremental kriging model is updated by adding an extension term to the Gaussian correlation function matrix of the incremental kriging model using the partition matrix inversion method.

9. A heat dissipation design method for switchgear, characterized in that, The optimal solution for the target variable is determined based on the method of any one of claims 1 to 8; Based on the optimal solution, the installation position of the heat dissipation component on the switch cabinet is designed to dissipate heat from the switch cabinet.

10. An electronic device, comprising: A processor and a memory storing a program, characterized in that the program includes instructions that, when executed by the processor, cause the processor to perform the method according to any one of claims 1 to 8.

Images