Method, device and equipment for selecting temperature monitoring points of oil-immersed transformer winding

Abstract

The invention provides an oil-immersed transformer winding temperature monitoring point selection method, device and equipment, and relates to the technical field of transformer temperature monitoring. The method comprises the following steps: based on a pre-constructed simulation model of the oil-immersed transformer winding, performing simulation calculation on a fluid-solid coupling heat transfer process of the oil-immersed transformer winding to obtain a snapshot matrix of the oil-immersed transformer winding; according to the snapshot matrix, determining a fitness function in an improved genetic algorithm, and screening out a plurality of target temperature monitoring points of the oil-immersed transformer winding from the plurality of target nodes based on the improved genetic algorithm; wherein the target node is a node, on which a temperature monitoring facility can be arranged, on the oil-immersed transformer winding. According to the invention, it can be ensured that the screened target temperature monitoring points can meet the arrangement requirement and the temperature monitoring requirement of the oil-immersed transformer winding.

Description

Technical Field

[0001] This invention relates to the field of transformer temperature monitoring technology, and in particular to a method, device and equipment for selecting temperature monitoring points of oil-immersed transformer windings. Background Technology

[0002] As a core component of power conversion in power systems, the winding temperature of oil-immersed transformers directly affects insulation life and operational safety. Rapidly and accurately acquiring winding temperature distribution and hotspot locations is of great significance for preventing thermal aging faults and improving power grid reliability.

[0003] In existing technologies for measuring the winding temperature of oil-immersed transformers, reduced-order models (ROMs) are introduced to quickly solve the temperature field and overcome computational bottlenecks. Intrusive ROMs, such as the POD-DEIM method, reduce the dimensionality of the equations through basis projection, but residual computational overhead such as nonlinear interpolation and high-dimensional basis operations still limits the efficiency of transient solutions. Non-intrusive ROMs, combining sensor data, construct a computational flow from monitoring point data to reduced-order modes and then to the overall temperature field, avoiding the solution of governing equations and becoming a feasible path for efficiently reconstructing the temperature field. In this type of method, the selection of monitoring points significantly affects the computational accuracy.

[0004] However, existing monitoring point placement methods have fundamental limitations. Engineering point placement relies on subjective experience and lacks quantitative evaluation standards. Selecting monitoring points across the entire field usually yields a local optimum, making it impossible to optimize monitoring points within the pre-selected node range. This can easily lead to situations where the selected monitoring point locations are actually not feasible. Summary of the Invention

[0005] This invention provides a method, apparatus, and equipment for selecting temperature monitoring points in oil-immersed transformer windings, to solve the problem that existing point placement methods cannot optimize the selection of monitoring points within a pre-selected node range, and that the selected monitoring point locations are often not feasible for actual placement.

[0006] In a first aspect, embodiments of the present invention provide a method for selecting temperature monitoring points of an oil-immersed transformer winding, including: Based on a pre-built simulation model of the oil-immersed transformer winding, the fluid-structure interaction heat transfer process of the oil-immersed transformer winding is simulated and calculated to obtain the snapshot matrix of the oil-immersed transformer winding. Based on the snapshot matrix, the fitness function in the improved genetic algorithm is determined, and based on the improved genetic algorithm, multiple target temperature monitoring points of the oil-immersed transformer winding are selected from multiple target nodes; wherein, the target node is the node on the oil-immersed transformer winding that can be equipped with temperature monitoring facilities.

[0007] Secondly, embodiments of the present invention provide a device for selecting temperature monitoring points of an oil-immersed transformer winding, comprising: The simulation module is used to simulate and calculate the fluid-structure interaction heat transfer process of the oil-immersed transformer winding based on a pre-built simulation model of the oil-immersed transformer winding, and obtain the snapshot matrix of the oil-immersed transformer winding. The screening module is used to determine the fitness function in the improved genetic algorithm based on the snapshot matrix, and to screen multiple target temperature monitoring points of the oil-immersed transformer winding from multiple target nodes based on the improved genetic algorithm; wherein, the target node is the node on the oil-immersed transformer winding that can be equipped with temperature monitoring facilities.

[0008] Thirdly, embodiments of the present invention provide an electronic device, including a memory and a processor, wherein the memory stores a computer program, and the processor executes the computer program to implement the method described in the first aspect or any possible implementation thereof.

[0009] In this embodiment of the invention, a simulation model of the oil-immersed transformer winding is pre-constructed to simulate the fluid-structure interaction heat transfer process of the oil-immersed transformer winding, obtaining a snapshot matrix of the oil-immersed transformer winding. Based on the snapshot matrix, the fitness function in the improved genetic algorithm is determined. Furthermore, based on the improved genetic algorithm, multiple target temperature monitoring points of the oil-immersed transformer winding are selected from multiple target nodes. The target nodes are nodes on the oil-immersed transformer winding where temperature monitoring facilities can be installed. This setting ensures that all target temperature monitoring points can be arranged. By combining the improved genetic algorithm with the snapshot matrix to optimize the target nodes, it can be ensured that the finally selected target temperature monitoring points can meet the temperature monitoring requirements of the oil-immersed transformer winding to the greatest extent. Attached Figure Description

[0010] Figure 1 This is a flowchart illustrating the implementation of the method for selecting temperature monitoring points of an oil-immersed transformer winding provided in this embodiment of the invention. Figure 2a This is a simulation model diagram of the 110kV oil-immersed transformer winding for the method of selecting temperature monitoring points for oil-immersed transformer windings provided in this embodiment of the invention. Figure 2b This is a partial enlarged view of section 2 of the winding of a 110kV oil-immersed transformer, which is the method for selecting temperature monitoring points of the oil-immersed transformer winding provided in the embodiment of the present invention. Figure 2c This is a partial enlarged view of each coil of a 110kV oil-immersed transformer winding, as provided in the embodiment of the present invention for the method of selecting temperature monitoring points for oil-immersed transformer windings. Figure 2d This is a partial enlarged view of the copper conductor of each turn of the winding of a 110kV oil-immersed transformer, which is the method for selecting temperature monitoring points of the oil-immersed transformer winding provided in the embodiment of the present invention. Figure 3 This is a flowchart illustrating the implementation of step S120 of the method for selecting temperature monitoring points of an oil-immersed transformer winding provided in an embodiment of the present invention. Figure 4 This is a flowchart illustrating the implementation of step S1205 of the method for selecting temperature monitoring points of an oil-immersed transformer winding provided in an embodiment of the present invention. Figure 5 This is a flowchart illustrating the implementation of step S12052 of the method for selecting temperature monitoring points of an oil-immersed transformer winding provided in this embodiment of the invention. Figure 6 This is a cross diagram illustrating the method for selecting temperature monitoring points of an oil-immersed transformer winding provided in an embodiment of the present invention. Figure 7 This invention provides a method for selecting temperature monitoring points of oil-immersed transformer windings, which uses a snapshot matrix to solve for the cumulative energy percentage of different modes. Figure 8 This is a graph showing the root mean square error and average relative error at each time step of the method for selecting temperature monitoring points of oil-immersed transformer windings provided in this embodiment of the invention. Figure 9 These are the transient temperature rise curves and error curves of the method for selecting temperature monitoring points of oil-immersed transformer windings provided in this embodiment of the invention. Figure 10 This is a schematic diagram of the structure of the oil-immersed transformer winding temperature monitoring point selection device provided in an embodiment of the present invention; Figure 11 This is a schematic diagram of an electronic device provided in an embodiment of the present invention. Detailed Implementation

[0011] The embodiments of the present invention will now be described in detail with reference to the accompanying drawings.

[0012] See Figure 1 The document illustrates a flowchart of the implementation of the method for selecting temperature monitoring points of an oil-immersed transformer winding provided in an embodiment of the present invention, which is described in detail below: Step S110: Based on the pre-built simulation model of the oil-immersed transformer winding, the fluid-structure interaction heat transfer process of the oil-immersed transformer winding is simulated and calculated to obtain the snapshot matrix of the oil-immersed transformer winding.

[0013] In some embodiments, taking a 110kV oil-immersed transformer winding as an example, its simulation model is as follows: Figure 2a , Figure 2b , Figure 2c , Figure 2d As shown, the simulation model includes main structures such as baffles, insulating paper, coils, oil channels, and oil cylinders. (See also...) Figure 2aThe simulation model has an overall width of 155.5 mm and a height of 1125.5 mm. All windings and oil channels between two adjacent baffles form a partition, resulting in a total of 8 partitions. Oil flow within each partition carries away heat from the windings through S-shaped oil channels. The entire model contains 66 disc-shaped windings, which are then meshed to complete the full-order model construction. (See also...) Figure 2b The left side of the vertical oil passage is the inner axial side, with a dimension of 8mm. The right side is the outer axial side, with a dimension of 10mm. The oil inlet is at the lower end of the outer axial side, and the oil outlet is at the upper end of the inner axial side. The distance between the oil cakes is 6mm, and the distance between the oil cake and the baffle is 3mm. (See also...) Figure 2c Each coil consists of 30 turns of copper wire, and each coil is 138mm wide. See also Figure 2d Each coil is 10.8mm high. The copper wire turns of the coil are separated by insulating paper with a thickness of 0.7mm. The total width of the copper wire and the insulating paper on both sides is 4.6mm.

[0014] Based on this model, the fluid-structure interaction heat transfer process of the winding is simulated. During the simulation, the total simulation time can be set to 2000s, and the time step can be 10s. The Fluent simulation results can be obtained through the simulation model.

[0015] It should be noted that at each time step, the simulation model can calculate the temperature of each target node. These temperatures are arranged into a column vector, and the column vectors of all time steps form a snapshot matrix. The snapshot matrix is ​​as follows:

[0016] in, Each column vector represents a snapshot of the winding temperature field, that is, the temperature value of each target node in a time step. This refers to the number of snapshots, i.e., the number of time steps. This is a column vector (temperature field snapshot) consisting of the temperatures of all target nodes at time step k.

[0017] Step S120: Based on the snapshot matrix, determine the fitness function in the improved genetic algorithm, and based on the improved genetic algorithm, select multiple target temperature monitoring points of the oil-immersed transformer winding from multiple target nodes; wherein, the target node is a node on the oil-immersed transformer winding that can be equipped with temperature monitoring facilities.

[0018] In some embodiments, the fitness function is used to evaluate the fitness of each individual in the population obtained in each iteration of the improved genetic algorithm. The higher the fitness, the better the monitoring effect when using multiple target nodes contained in that individual for temperature monitoring. For a 110kV oil-immersed transformer winding, five target temperature monitoring points can be selected, and the number of target temperature monitoring points can also be other than that.

[0019] See Figure 3 The specific processing method of step S120 above includes steps S1201-S1205, and the specific content is as follows: Step S1201: Based on the simulation model, perform reduced-order calculation on the fluid-structure interaction heat transfer process of the oil-immersed transformer winding to obtain the reduced-order calculation results of each node of the oil-immersed transformer winding at each time step.

[0020] In some embodiments, order reduction computation is a computational method that transforms a complex high-order problem into a low-order problem for solution by simplifying the model dimensions, ignoring minor factors, or using approximation methods.

[0021] Step S1202: Calculate the root mean square error of the reduced-order calculation results and the full-order calculation results for all nodes at each time step.

[0022] In some embodiments, the root mean square error is calculated using the following formula:

[0023] in, This represents the full-order computation result of the i-th node in the j-th time step. This represents the order reduction calculation result for the i-th node at the j-th time step. This represents the total number of nodes in the model.

[0024] Step S1203: The mean of the root mean square error of all time steps is determined as the fitness function in the improved genetic algorithm.

[0025] In some embodiments, the fitness of each individual in the population is calculated in each iteration to serve as the selection criterion for superior individuals. The physical meaning of the fitness function is the reciprocal of the mean root mean square error of each time step in the snapshot matrix. The individual with the highest fitness selected through the fitness function is the one with the best accuracy in full-field temperature inversion, i.e., the most accurate individual for full-field temperature monitoring. The advantage of this fitness function is that it facilitates the selection of monitoring points capable of achieving high-precision full-field temperature inversion. The fitness function is:

[0026] in, Let be the fitness function of the y-th individual. This represents the number of time steps.

[0027] Step S1204: Based on the target node, the predetermined population size, and the individual gene length, construct the initial population in the improved genetic algorithm; each individual in the initial population is a candidate monitoring point selection scheme; each monitoring point selection scheme includes N target nodes at different locations; where N is a preset value.

[0028] In some embodiments, to reduce computational complexity, narrow the search range, and ensure that temperature monitoring facilities can be installed at the locations of the finally selected temperature monitoring points, an initial population needs to be set based on the target nodes. The population size can be set to 50, the maximum number of iterations to 100, or other values; no specific restrictions are imposed here. When the population size is set to 50, the initial population contains 50 individuals. Individual encoding uses real-number encoding, and the gene length of each individual is the pre-set number of target temperature monitoring points to be determined. The gene code is the node number in the simulation model. N is pre-set, and its value is related to the individual gene length; for example, when the individual gene length is 5, the value of N is also 5. The monitoring point selection scheme includes multiple target nodes located at different locations.

[0029] Step S1205: Based on the predetermined maximum number of iterations and fitness function, iterate the initial population to obtain the target population, and determine the multiple target nodes including the individuals with the highest fitness in the target population as multiple target temperature monitoring points of the oil-immersed transformer winding.

[0030] In some embodiments, the predetermined maximum number of iterations can be set to 100, or other numbers. In the improved genetic algorithm, the final population is determined by continuously iterating over the initial population. When the maximum number of iterations is reached, iteration stops, and the population obtained from the last iteration is taken as the final population, i.e., the target population. During the iteration process, individuals are selected using a fitness function. When the population size is 50, the target population will contain 50 individuals, and each individual will contain a predetermined number of target nodes.

[0031] See Figure 4 In each iteration, the specific processing method of step S1205 includes steps S12051-S12054, as detailed below: Step S12051: Calculate the fitness of each individual in the population obtained in the previous iteration according to the fitness function.

[0032] In some embodiments, the population can be updated during the current iteration by calculating the fitness of each individual in the population obtained in the previous iteration.

[0033] Step S12052: Update the population obtained in the previous iteration based on the fitness of each individual to obtain the population for the current iteration.

[0034] In some embodiments, based on the fitness of each individual, the crossover probability and mutation probability of each individual can be calculated, and crossover can be performed according to the crossover probability. Then, the individuals after crossover can be mutated according to the mutation probability to obtain the individuals of the population in this iteration process.

[0035] See Figure 5 The specific processing method of the above step S12052 includes steps S120521-S120524, the specific contents of which are as follows: Step S120521: Using the binary tournament selection operator, determine the parent individuals in the current iteration process based on the fitness function of each individual and the individuals in the population whose fitness is not the highest obtained in the previous iteration.

[0036] In some embodiments, the improved genetic algorithm requires applying a selection operator to the entire population to select individuals with high fitness, ensuring these individuals have a greater chance of passing on their genes to the next generation. This application employs a binary tournament selection operator, utilizing a random competition mechanism where two individuals are randomly selected from the population each time to form a group; the individual with higher fitness wins. This method balances selection pressure and computational efficiency, retaining highly fit individuals while maintaining population diversity and effectively avoiding local optima. The parent individual refers to an individual in the population after the previous iteration.

[0037] Step S120522: Based on the fitness of each individual, crossover is performed on all parent individuals to obtain multiple offspring individuals.

[0038] In some embodiments, this scheme employs a sequential crossover operator. The sequential crossover operator requires a combination of two parent individuals to exchange partial gene segments to obtain two offspring individuals. Therefore, before crossover, the combination of parent individuals needs to be determined. To preserve individuals with high fitness and improve individuals with low fitness through crossover, the crossover probability needs to be calculated based on fitness to control the crossover of parent individuals. Using the sequential crossover operator can introduce new gene combinations while maintaining the superior genetic characteristics of the parents, thereby increasing population diversity and gradually optimizing the population's genetic makeup.

[0039] In one possible implementation, step S120522 is specifically processed as follows: based on the fitness of each parent individual, the crossover probability of each parent individual combination is determined, and based on the crossover probability of each parent individual combination, the parent individual combination that needs to be crossed is determined; wherein, the parent individual combination contains any two parent individuals; the two parent individuals in the parent individual combination that does not need to be crossed are determined as the two child individuals of the parent individual combination; the two parent individuals in the parent individual combination are crossed to obtain two child individuals.

[0040] In some embodiments, after determining the crossover probability for each parent generation combination, it is necessary to use the crossover probability to screen out the parent individuals that need to undergo crossover and those that do not. Parent individuals that do not need to undergo crossover will pass on all their genes to their offspring. See also Figure 6 During crossover, the start and end positions of the crossover segment are randomly selected from the two parent individuals p1 and p2. The gene between the start and end positions in parent individual p1 is copied to the same position in offspring individual c1. Then, the missing gene in c1 is filled into parent individual p2 in order to obtain offspring gene c1. Similarly, the gene between the start and end positions in parent individual p2 is copied to the same position in offspring individual c2. Then, the missing gene in c2 is filled into parent individual p2 in order to obtain offspring gene c2.

[0041] In one possible implementation, the specific processing method for "determining the crossover probability of each parent individual combination based on the fitness of each parent individual" is as follows: The fitness of the parent individuals with higher fitness in each parent individual combination is determined as the fitness of the corresponding parent individual combination; the fitness of each parent individual combination is input into the crossover probability calculation expression to obtain the crossover probability of each parent individual combination; the crossover probability calculation expression is:

[0042] in, For crossover probability, This represents the maximum crossover probability. This represents the minimum crossover probability. For the fitness of the parent generation's individual combinations, This represents the average fitness of all individuals in the population obtained from the previous iteration. This represents the current iteration number. This represents the maximum number of iterations.

[0043] In some embodiments, for individuals participating in crossover, the initial crossover probability can be set to 0.9, and a dual adaptive mechanism combining generation decay and fitness awareness can be introduced to adjust the crossover probability, resulting in a crossover probability calculation expression. According to the crossover probability calculation expression, it can be seen that for individuals with high fitness, the adaptive crossover probability that decays with generation can protect the superior genes of individuals with high fitness. For ordinary individuals, the basic crossover probability can be maintained, enhancing the search capability.

[0044] Step S120523: Based on the fitness of each offspring individual, calculate the mutation probability of each offspring individual, and perform single-point mutation on the corresponding offspring individual according to the mutation probability.

[0045] In some embodiments, a single-point mutation operator is used to mutate a single target node within an individual according to the mutation rate. An adaptive mutation probability is set through a dual adaptive mechanism of generation decay and fitness awareness, which can perform mutation operations on offspring individuals to obtain new offspring individuals.

[0046] In one possible implementation, step S120523 is specifically processed as follows: the fitness of each offspring individual is input into the mutation probability calculation formula to calculate the mutation probability of each offspring individual; the mutation probability calculation formula is:

[0047] in, The mutation probability, This represents the maximum mutation probability. This represents the minimum probability of mutation. For offspring individuals The degree of adaptability.

[0048] In some embodiments, the mutation probability of each offspring individual is determined according to the mutation probability calculation formula, and mutation operations are performed on the offspring individuals to obtain new offspring individuals. This method can preserve the superior genes of individuals with high fitness, maintain the basic mutation rate of individuals with insufficient fitness, and enhance exploration capabilities. The maximum mutation probability can be set to 0.1, and the minimum mutation probability can be set to 0.05.

[0049] Step S120524: Combine all offspring individuals after single-point mutation with the individual with the highest fitness from the population obtained in the previous iteration to obtain the population in this iteration.

[0050] In some embodiments, the individual with the highest fitness from the previous iteration is directly preserved as an individual in the population of the current iteration without undergoing crossover and mutation. Individuals from the previous iteration whose fitness is not the highest undergo crossover and mutation, and all resulting offspring constitute the population of the current iteration.

[0051] Step S12053: Determine whether the number of iterations corresponding to this iteration process is the maximum number of iterations.

[0052] In some embodiments, the number of iterations is used to determine whether to stop the iteration. The number of iterations can be set by the user and is not limited here.

[0053] In step S12054, if the number of iterations is not the maximum, the iteration continues; if the number of iterations is the maximum, the population obtained in this iteration is determined as the target population of the oil-immersed transformer winding.

[0054] In some embodiments, when the number of iterations reaches the maximum number of iterations, the iteration is stopped, the population obtained in this iteration is taken as the target population, and the target nodes contained in the individuals with the highest fitness in the target population are taken as the target temperature monitoring points of the oil-immersed transformer winding.

[0055] It should be noted that after the maximum number of iterations is reached, i.e., after the improved genetic algorithm ends, the selected target temperature monitoring points can still be verified. By combining the reduced-order model and monitoring point data, a rapid calculation of the entire temperature field distribution can be performed to invert the temperature monitored by the target temperature monitoring points. The calculation results can then be compared with the Fluent simulation results (i.e., the calculation results of the simulation model) to verify the effectiveness of the invention.

[0056] The specific verification process is as follows: First, select the principal modes: The process of selecting the principal modes is as follows: perform singular value decomposition on the snapshot matrix, truncate it according to the magnitude of the singular values, and approximate it with the first d-order modes to obtain the following formula:

[0057] Among them, the left orthogonal matrix can be regarded as a set of orthogonal bases of the snapshot matrix in the reduced-order subspace, which characterizes the main features of the temperature field distribution. The diagonal elements of the diagonal matrix are the singular values ​​of the snapshot matrix, arranged in descending order, and can be used to determine the value of d.

[0058] The formula for determining d is:

[0059] in, This represents the proportion of energy in the first d-order modes. Let i be the i-th singular value.

[0060] exist In the given information, the first d singular values ​​are much larger than the subsequent singular values; therefore, in general, When the value is greater than or equal to 99.9%, the original system can be approximated with high precision using a first-d order matrix, where d is much smaller than n. At this point, the temperature field snapshot at any given time in the snapshot matrix... The representation in the reduced-order subspace is as follows:

[0061] in, For the reduced-order mode matrix, Here is the modal coefficient matrix. .

[0062] In addition, in this invention, the full-order calculation results of the first 40 time steps are used to construct a snapshot matrix, and the full-order calculation results of the following 160 time steps are used for method verification. During the verification process, it was found that when the 5th order mode is taken, the cumulative energy ratio can be higher than 99.9% of the full-order model, which is sufficient to approximately capture all the characteristics of temperature field changes.

[0063] Second, construct the relationship between monitoring points and global temperature inversion: After obtaining a series of snapshot data, use the basic idea of ​​the GappyPOD method to solve the modal coefficients, establish the mathematical connection from discrete point temperature to global temperature field distribution, and quickly calculate the complete winding temperature field distribution based on the discrete point data and reduced-order modes of winding temperature.

[0064] Third, the inversion results of the winding's full-field temperature are compared with the Fluent simulation results to analyze the calculation accuracy and efficiency of the invention and verify its effectiveness.

[0065] Taking the transient temperature inversion problem of a 110kV oil-immersed transformer winding as an example, a reduced-order model is constructed using snapshot data from the first 40 time steps. The temperature field of subsequent time steps is then rapidly reconstructed and calculated, and the comparison process and results are shown below: Using the snapshot matrix as input, an improved genetic algorithm is employed to select the optimal monitoring point locations when there are 5 monitoring points. The final results are nodes 43532, 234465, 261549, 943717, and 1951263. The cumulative energy percentage for different numbers of modes is then calculated based on the snapshot matrix. Figure 7 As shown, the cumulative energy of the first five modes accounts for as much as 99.99%, which is sufficient to approximately capture all the characteristics of the temperature field change. Therefore, to improve computational efficiency, only the first five modes are used for temperature field calculation. Regarding computational accuracy, the results of this invention are further compared with simulation results, and the root mean square error and average relative error at each time step are calculated as follows: Figure 8 As shown. Simultaneously, the hotspot locations during the steady-state phase are identified, and their transient temperature rise curves and error curves are plotted as follows. Figure 9 As shown. See also Figure 8 and Figure 9 During the 0-2000s period, the root mean square error and hotspot error of the entire temperature field at each time step increased with time, but eventually stabilized. The maximum root mean square error was approximately 0.22K, the maximum average relative error was approximately 0.59%, and the maximum hotspot error was approximately 1.03K; all error indicators were within a reasonable range. To analyze the superiority of this invention over existing monitoring point selection methods in full-field temperature inversion, the calculation errors of this invention, the greedy algorithm, and the column-pivot QR decomposition method were compared, and the results are shown in Table 1. The above analysis shows that this invention has high accuracy in global reconstruction of transient temperature fields and calculation of hotspot temperatures.

[0066] Table 1. Statistics of various error indicators

[0067] As shown in Table 1, compared with existing monitoring point selection methods, the monitoring point layout scheme determined by the method of this invention can obtain more accurate full-field temperature inversion results. Furthermore, as the principle shows, this invention can optimize the selection of monitoring points within any range of candidate nodes, which helps to arrange monitoring points according to the actual conditions allowed by the transformer.

[0068] Using a pre-constructed simulation model of the oil-immersed transformer winding, the fluid-structure interaction heat transfer process of the oil-immersed transformer winding is simulated and calculated to obtain a snapshot matrix of the oil-immersed transformer winding. Based on the snapshot matrix, the fitness function in the improved genetic algorithm is determined. Furthermore, the improved genetic algorithm is used to screen multiple target temperature monitoring points of the oil-immersed transformer winding from multiple target nodes. The target nodes are nodes on the oil-immersed transformer winding where temperature monitoring facilities can be installed. This setting ensures that all target temperature monitoring points can be arranged. In addition, by combining the improved genetic algorithm with the snapshot matrix to optimize the target nodes, the gene with the highest fitness obtained in each iteration of the improved genetic algorithm is retained. Genes with non-highest fitness are cross-crossed and mutated. This can avoid local optima while maintaining good genes, ensuring that the finally selected target temperature monitoring points can meet the temperature monitoring requirements of the oil-immersed transformer winding to the greatest extent.

[0069] It should be understood that the sequence number of each step in the above embodiments does not imply the order of execution. The execution order of each process should be determined by its function and internal logic, and should not constitute any limitation on the implementation process of the embodiments of the present invention.

[0070] The following are device embodiments of the present invention. For details not described in detail, please refer to the corresponding method embodiments described above.

[0071] Figure 10 A schematic diagram of the structure of the oil-immersed transformer winding temperature monitoring point selection device provided in an embodiment of the present invention is shown. For ease of explanation, only the parts related to the embodiment of the present invention are shown, and are described in detail below: like Figure 10 As shown, the oil-immersed transformer winding temperature monitoring point selection device 10 includes: Simulation module 101 is used to simulate and calculate the fluid-structure interaction heat transfer process of oil-immersed transformer windings based on a pre-built simulation model of oil-immersed transformer windings, and obtain a snapshot matrix of oil-immersed transformer windings. The screening module 102 is used to determine the fitness function in the improved genetic algorithm based on the snapshot matrix, and to screen out multiple target temperature monitoring points of the oil-immersed transformer winding from multiple target nodes based on the improved genetic algorithm; wherein, the target node is a node on the oil-immersed transformer winding that can be equipped with temperature monitoring facilities.

[0072] In one possible implementation, the screening module 102 specifically includes: performing a reduced-order calculation on the fluid-structure interaction heat transfer process of the oil-immersed transformer winding based on the simulation model, obtaining the reduced-order calculation result of each node of the oil-immersed transformer winding at each time step; calculating the root mean square error of the reduced-order calculation results and the full-order calculation results of all nodes at each time step; and determining the mean of the root mean square error of all time steps as the fitness function in the improved genetic algorithm.

[0073] In one possible implementation, the screening module 102 further includes: constructing an initial population in the improved genetic algorithm based on the target nodes, a predetermined population size, and individual gene length; each individual in the initial population is a candidate monitoring point selection scheme; each monitoring point selection scheme includes N target nodes at different locations; where N is a preset value; iterating the initial population according to a predetermined maximum number of iterations and a fitness function to obtain a target population, and determining the multiple target nodes including the individual with the highest fitness in the target population as multiple target temperature monitoring points of the oil-immersed transformer winding.

[0074] In one possible implementation, the screening module 102 further includes: performing the following steps in each iteration: calculating the fitness of each individual in the population obtained in the previous iteration according to the fitness function; updating the population obtained in the previous iteration according to the fitness of each individual to obtain the population for the current iteration; determining whether the iteration number corresponding to the current iteration is the maximum iteration number; if it is not the maximum iteration number, continuing the iteration; if it is the maximum iteration number, determining the population obtained in the current iteration as the target population of the oil-immersed transformer winding.

[0075] In one possible implementation, the screening module 102 further includes: using a binary tournament selection operator to determine the parent individuals in the current iteration based on the fitness function of each individual and the individuals with non-highest fitness in the population obtained in the previous iteration; performing crossover on all parent individuals based on the fitness of each individual to obtain multiple child individuals; calculating the mutation probability of each child individual based on the fitness of each child individual, and performing single-point mutation on the corresponding child individuals according to the mutation probability; and combining all child individuals after single-point mutation with the individual with the highest fitness in the population obtained in the previous iteration to obtain the population in the current iteration.

[0076] In one possible implementation, the screening module 102 further includes: determining the crossover probability of each parent individual combination based on the fitness of each parent individual, and determining the parent individual combinations that need to be crossed based on the crossover probability of each parent individual combination; wherein, the parent individual combination contains any two parent individuals; determining the two parent individuals in the parent individual combination that do not need to be crossed as the two child individuals of the parent individual combination; and crossing the two parent individuals in the parent individual combination to obtain two child individuals.

[0077] In one possible implementation, the screening module 102 further includes: determining the fitness of the parent individuals with higher fitness in each parent individual combination as the fitness of the corresponding parent individual combination; inputting the fitness of each parent individual combination into the crossover probability calculation expression to obtain the crossover probability of each parent individual combination; the crossover probability calculation expression is:

[0078] in, For crossover probability, This represents the maximum crossover probability. This represents the minimum crossover probability. For the fitness of the parent generation's individual combinations, This represents the average fitness of all individuals in the population obtained from the previous iteration. This represents the current iteration number. This represents the maximum number of iterations.

[0079] In one possible implementation, the screening module 102 further includes: inputting the fitness of each offspring individual into the mutation probability calculation formula to calculate the mutation probability of each offspring individual; the mutation probability calculation formula is:

[0080] in, The mutation probability, This represents the maximum mutation probability. This represents the minimum probability of mutation. For offspring individuals The degree of adaptability.

[0081] Figure 11 This is a schematic diagram of an electronic device provided in an embodiment of the present invention. For example... Figure 11 As shown, the electronic device 1110 of this embodiment includes a processor 1111 and a memory 1112. The memory 1112 stores a computer program 1113. When the processor 1111 executes the computer program 1113, it implements the steps in the various method embodiments described above. Alternatively, when the processor 1111 executes the computer program 1113, it implements the functions of each module / unit in the various device embodiments described above.

[0082] For example, computer program 1113 may be divided into one or more modules / units, which are stored in memory 1112 and executed by processor 1111 to complete the present invention. The one or more modules / units may be a series of computer program instruction segments capable of performing a specific function, which describe the execution process of computer program 1113 in electronic device 1110.

[0083] Electronic device 1110 may include, but is not limited to, processor 1111 and memory 1112. Those skilled in the art will understand that... Figure 11 This is merely an example of electronic device 1110 and does not constitute a limitation on electronic device 1110. It may include more or fewer components than shown, or combine certain components, or different components. For example, electronic device 1110 may also include input / output devices, network access devices, buses, etc.

[0084] For the sake of simplicity and clarity, only the above-described functional modules / units are used as examples. In practical applications, the functions described above can be assigned to different functional modules / units as needed. These modules / units can be implemented in hardware, software, or a combination of both.

[0085] In the above embodiments, the descriptions of each embodiment have their own emphasis. Parts not detailed or described in a particular embodiment can be referred to in the relevant descriptions of other embodiments. Unless otherwise specified or in conflict with logic, the terminology and / or descriptions between different embodiments are consistent and can be referenced interchangeably. Technical features in different embodiments can be combined to form new embodiments based on their inherent logical relationships.

[0086] The above-described embodiments are only used to illustrate the technical solutions of the present invention, and are not intended to limit it. Although the present invention has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that modifications can still be made to the technical solutions described in the foregoing embodiments, or equivalent substitutions can be made to some of the technical features. Such modifications or substitutions do not cause the essence of the corresponding technical solutions to deviate from the spirit and scope of the technical solutions of the embodiments of the present invention, and should all be included within the protection scope of the present invention.

Claims

1. A method for selecting temperature monitoring points for oil-immersed transformer windings, characterized in that, include: Based on a pre-built simulation model of the oil-immersed transformer winding, the fluid-structure interaction heat transfer process of the oil-immersed transformer winding is simulated and calculated to obtain a snapshot matrix of the oil-immersed transformer winding. Based on the snapshot matrix, the fitness function in the improved genetic algorithm is determined, and based on the improved genetic algorithm, multiple target temperature monitoring points of the oil-immersed transformer winding are selected from multiple target nodes; wherein, the target node is a node on the oil-immersed transformer winding that can be equipped with temperature monitoring facilities.

2. The method for selecting temperature monitoring points of an oil-immersed transformer winding according to claim 1, characterized in that, The oil-immersed transformer winding includes multiple nodes; the snapshot matrix contains the full-order calculation results of each node at each time step; determining the fitness function in the improved genetic algorithm based on the snapshot matrix includes: Based on the simulation model, the fluid-structure interaction heat transfer process of the oil-immersed transformer winding is calculated in a reduced order, and the reduced order calculation results of each node of the oil-immersed transformer winding at each time step are obtained. Calculate the root mean square error of the reduced-order calculation results and the full-order calculation results for all nodes at each time step; The mean of the root mean square error of all time steps is determined as the fitness function in the improved genetic algorithm.

3. The method for selecting temperature monitoring points of an oil-immersed transformer winding according to claim 1, characterized in that, The improved genetic algorithm is used to select multiple target temperature monitoring points for the oil-immersed transformer winding from multiple target nodes, including: Based on the target nodes, the predetermined population size, and the individual gene length, an initial population is constructed in the improved genetic algorithm; each individual in the initial population is a candidate monitoring point selection scheme; each monitoring point selection scheme includes N target nodes at different locations; where N is a preset value. Based on the predetermined maximum number of iterations and the fitness function, the initial population is iterated to obtain the target population, and the multiple target nodes including the individuals with the highest fitness in the target population are determined as multiple target temperature monitoring points of the oil-immersed transformer winding.

4. The method for selecting temperature monitoring points of oil-immersed transformer windings according to claim 3, characterized in that, The step of iterating the initial population to obtain the target population based on a predetermined maximum number of iterations and the fitness function includes: In each iteration, the following steps are performed: Calculate the fitness of each individual in the population obtained in the previous iteration based on the fitness function. Based on the fitness of each individual, the population obtained in the previous iteration is updated to obtain the population for the current iteration. Determine whether the number of iterations corresponding to this iteration process is the maximum number of iterations; If the maximum number of iterations is not reached, continue iterating. If it is the maximum number of iterations, then the population obtained in this iteration process is determined as the target population of the oil-immersed transformer winding.

5. The method for selecting winding temperature monitoring points of an oil-immersed transformer according to claim 4, characterized in that, The process of updating the population obtained in the previous iteration based on the fitness of each individual to obtain the population for the current iteration includes: Using the binary tournament selection operator, the parent individuals in the current iteration are determined based on the fitness function of each individual and the individuals in the population whose fitness is not the highest obtained in the previous iteration. Based on the fitness of each individual, crossover is performed on all parent individuals to obtain multiple offspring individuals; Based on the fitness of each offspring individual, the mutation probability of each offspring individual is calculated, and single-point mutation is performed on the corresponding offspring individual according to the mutation probability. The population for this iteration is obtained by combining all offspring individuals after a single mutation with the individual with the highest fitness from the population obtained in the previous iteration.

6. The method for selecting temperature monitoring points of an oil-immersed transformer winding according to claim 5, characterized in that, Based on the fitness of each individual, all parent individuals are cross-crossed to obtain multiple offspring individuals, including: Based on the fitness of each parent individual, the crossover probability of each parent individual combination is determined, and based on the crossover probability of each parent individual combination, the parent individual combinations that need to be crossovered are determined; wherein, the parent individual combination contains any two parent individuals. Two parent individuals in a combination of parent individuals that do not require crossover are identified as the two child individuals of that parent individual combination. Crossing the two parent individuals in the parent individual combination yields two child individuals.

7. The method for selecting temperature monitoring points of an oil-immersed transformer winding according to claim 6, characterized in that, The determination of the crossover probability for each combination of parent individuals based on the fitness of each parent individual includes: The fitness of the parent individuals with higher fitness in each parent combination is determined as the fitness of the corresponding parent combination. The fitness of each parent combination is input into the crossover probability calculation expression to obtain the crossover probability of each parent combination. The expression for calculating the crossover probability is: in, For crossover probability, This represents the maximum crossover probability. This represents the minimum crossover probability. For the fitness of the parent generation's individual combinations, This represents the average fitness of all individuals in the population obtained from the previous iteration. This represents the current iteration number. This represents the maximum number of iterations.

8. The method for selecting temperature monitoring points of an oil-immersed transformer winding according to claim 5, characterized in that, The calculation of the mutation probability of each offspring individual based on the fitness of each offspring individual includes: The fitness of each offspring individual is input into the mutation probability calculation formula to calculate the mutation probability of each offspring individual; The formula for calculating the mutation probability is: in, The mutation probability, This represents the maximum mutation probability. This represents the minimum probability of mutation. For offspring individuals The degree of adaptability.

9. A device for selecting temperature monitoring points of an oil-immersed transformer winding, characterized in that, include: The simulation module is used to simulate and calculate the fluid-structure interaction heat transfer process of the oil-immersed transformer winding based on a pre-built simulation model of the oil-immersed transformer winding, and obtain a snapshot matrix of the oil-immersed transformer winding. The screening module is used to determine the fitness function in the improved genetic algorithm based on the snapshot matrix, and to screen out multiple target temperature monitoring points of the oil-immersed transformer winding from multiple target nodes based on the improved genetic algorithm; wherein, the target node is a node on the oil-immersed transformer winding that can be equipped with temperature monitoring facilities.

10. An electronic device, characterized in that, It includes a memory and a processor, the memory storing a computer program, and the processor executing the computer program to implement the method as described in any one of claims 1 to 8.

Images