Flexible control method and system for transformer design parameter adjustment

Abstract

The invention provides a flexible control method and system for transformer design parameter adjustment, and relates to the field of transformer design, and the method comprises the steps: calculating an influence deviation value of each design parameter on a performance index; the design parameters are divided into key layer parameters, buffer layer parameters and adaptive layer parameters according to the ratio of the influence deviation value to the reference value; adopting a genetic algorithm to calculate a stable interval of key layer parameters, and setting a safety margin range and a scalable parameter group; if the parameter adjustment amount is within the safety margin range, the adjustment cost of each design parameter is calculated through a dynamic programming algorithm, and a target adjustment path with the lowest adjustment cost is determined; and calculating the final adjustment amount of the buffer layer parameters according to the target adjustment path and the telescopic parameter group. By implementing the method, the calculation resource consumption in the design parameter adjustment process can be reduced through hierarchical adjustment and optimized path selection, and the parameter adjustment efficiency in the transformer design process is improved.

Description

TECHNICAL FIELD

[0001] The present application relates to the field of transformer design, and in particular to a flexible control method and system for adjusting transformer design parameters. BACKGROUND

[0002] With the continuous expansion of the power system scale and the sustained growth of power demand, as a key device in the power system, the design level of the transformer directly affects the safe and stable operation of the power grid. In the design process of the transformer, multiple design parameters are involved, and there is a complex interrelation between these parameters. The adjustment of the parameters directly affects the performance indicators of the transformer. Therefore, how to reasonably adjust the design parameters of the transformer has become the focus of the industry.

[0003] In related technologies, the transformer design system generally uses simulation software to adjust parameters. Design personnel establish a transformer model through the simulation software, set the initial values of various parameters in the model, and then use the simulation analysis function to gradually adjust the parameters. Specifically, when a parameter needs to be adjusted, the simulation software calculates the change range of the related parameters according to the built-in calculation model, and the design personnel selects the appropriate parameter combination according to the simulation results.

[0004] However, in actual application, every time a parameter needs to be adjusted, the simulation software generally needs to perform complete performance verification calculation on all related parameters. This full-quantity calculation method not only consumes a large amount of computing resources, but also in the case of a large number of design parameters, the complete performance verification process often takes a long time, which greatly reduces the design efficiency of the transformer. SUMMARY

[0005] The present application provides a flexible control method and system for adjusting transformer design parameters, which is used to solve the problem of low design efficiency caused by large consumption of computing resources in the process of adjusting transformer design parameters.

[0006] In a first aspect, the present application provides a flexible control method for adjusting transformer design parameters, applied to a transformer design system, the method comprising: obtaining corresponding data of each design parameter of the transformer and the performance indicators, establishing a parameter-performance feature mapping database, and calculating the influence deviation value of each design parameter on the performance indicators based on the feature mapping database; determining the influence degree as the ratio of the influence deviation value to the reference value, and dividing the design parameters into key layer parameters, buffer layer parameters and adaptive layer parameters based on the influence degree; adopting a genetic algorithm to calculate a stable interval of the key layer parameter, and setting a safety margin range and a scalable parameter group based on the stable interval and the buffer layer parameter respectively, the scalable parameter group being used to adjust the buffer layer parameter and the adaptive layer parameter; When a parameter adjustment signal is detected, if a parameter adjustment amount corresponding to the parameter adjustment signal is within the safety margin range, an adjustment cost of each design parameter is calculated through a dynamic programming algorithm, and a target adjustment path with the lowest adjustment cost is determined; A final adjustment amount of the buffer layer parameter is calculated according to the target adjustment path and the scalable parameter group.

[0007] Through the above embodiment, the transformer design system realizes hierarchical management of transformer design parameters by establishing a parameter-performance feature mapping database and hierarchically classifying parameters based on the influence degree. Combined with the genetic algorithm to determine the stable interval of the key parameter and the dynamic programming algorithm to find the optimal adjustment path, the parameter adjustment process no longer needs to calculate all parameters. This method reduces the consumption of computing resources in the design parameter adjustment process through hierarchical adjustment and optimal path selection, while ensuring the rationality and reliability of parameter adjustment, and improving the efficiency of parameter adjustment in the transformer design process.

[0008] In some embodiments, the step of obtaining corresponding data of each design parameter of the transformer and a performance index, and establishing a parameter-performance feature mapping database, specifically includes: The corresponding data is divided into high-frequency mapping data and low-frequency mapping data according to the access frequency and stored in the memory cache area and the hard disk cache area respectively to generate a feature mapping database; Based on the historical adjustment records in the feature mapping database, a parameter association model for identifying the association relationship between the high-frequency mapping data and the low-frequency mapping data is established by using a data analysis method; When the access frequency of the mapping data in the memory cache area is lower than a preset frequency threshold, the mapping data is transferred to the hard disk cache area, and the mapping data in the hard disk cache area that has a strong association relationship with the current high-frequency access data is transferred to the memory cache area.

[0009] Through the above embodiment, the transformer design system realizes hierarchical cache management of data by classifying and storing mapping data according to access frequency and establishing a parameter association model. The system can store frequently accessed data in memory for fast access, and store low-frequency data in hard disk to save memory resources. At the same time, based on the parameter association model, the data is dynamically migrated to ensure fast access of frequently used data and reduce resource consumption during data reading, thereby improving the data processing efficiency in the parameter adjustment process.

[0010] In some embodiments, the step of dividing the design parameters into the key layer parameters, the buffer layer parameters and the adaptive layer parameters based on the influence degree specifically comprises: calculating a deviation between the performance optimization target and the actual performance indicator of the current design stage, and determining a dominant direction of performance optimization according to the deviation; performing weighted calculation on the influence degree based on the dominant direction to obtain a modified influence degree; counting the number of layer level conversions of each design parameter in the historical adjustment process, and calculating the stability of the parameter level; dynamically layering the design parameters according to the weighted values of the modified influence degree and the stability.

[0011] Through the above embodiments, the transformer design system calculates the deviation between the performance optimization target and the actual indicator, and combines the parameter influence degree and the historical adjustment record to realize dynamic optimization of parameter layering. This method considers the dominant direction of performance optimization and the stability of the parameter level, making the parameter layering more accurate and reasonable. This dynamic layering mechanism reduces the invalid calculation during parameter adjustment, concentrates the calculation resources on the optimization of key parameters, and improves the accuracy and efficiency of parameter adjustment.

[0012] In some embodiments, the step of calculating the adjustment cost of each design parameter by a dynamic programming algorithm and determining the target adjustment path with the lowest adjustment cost specifically comprises: dividing the parameter adjustment amount into multiple adjustment subintervals, and determining the initial adjustment direction of each adjustment subinterval by a simplified calculation method to generate an initial adjustment path, wherein the initial adjustment path comprises multiple adjustment nodes; calculating the influence degree of each adjustment node on the adjustment cost, and determining the adjustment nodes with the influence degree greater than a preset influence threshold as key nodes for fine calculation to generate an improved adjustment path; when detecting that the difference between the adjustment costs calculated in two adjacent times is less than a preset convergence threshold, determining the current improved adjustment path as the target adjustment path; if the difference between the adjustment costs is greater than the preset convergence threshold, the adjustment nodes are further calculated in the order of the size of the influence degree.

[0013] Through the above embodiments, the transformer design system divides the parameter adjustment amount into multiple subintervals and identifies key nodes to realize hierarchical optimization of the parameter adjustment path. The system only performs fine calculation on the key nodes with greater influence degree, and other nodes are calculated by a simplified calculation method, which greatly reduces the calculation amount. At the same time, the iterative optimization method ensures the convergence of the adjustment path, which significantly reduces the calculation resource consumption while ensuring the adjustment quality.

[0014] In some embodiments, the step of determining the adjustment nodes whose impact is greater than a preset impact threshold as key nodes, performing refined calculations, and generating an improved adjustment path specifically includes: Calculating a parameter sensitivity matrix of the key node based on the degree of influence, and determining an adjustment priority of each design parameter at the key node based on the parameter sensitivity matrix; Subdivide the adjustment subintervals of the key nodes according to the adjustment priorities to generate an encrypted adjustment point sequence; A local optimal adjustment path is calculated based on the encrypted adjustment point sequence, and the adjustment path segment at the original key node is replaced by the local optimal adjustment path to generate an improved adjustment path.

[0015] Through the above-described embodiment, the transformer design system achieves local refinement of the adjustment path by calculating a parameter sensitivity matrix and optimizing key nodes based on adjustment priorities. This method, by segmenting and locally optimizing key nodes, avoids performing detailed calculations on all nodes. This ensures adjustment accuracy at key locations while reducing computing resource consumption and improving the overall efficiency of parameter adjustment.

[0016] In some embodiments, after the step of calculating the stable interval of the key layer parameter using a genetic algorithm, the method further includes: Constructing a mutual influence matrix between the key layer parameters, and calculating the coupling degree of each key layer parameter based on the mutual influence matrix; Dividing the key layer parameters into multiple parameter combinations according to the coupling degree, and calculating the intersection of the stable intervals of each parameter combination; When it is detected that the intersection of the stable intervals is less than a preset interval threshold, the original stable intervals are modified based on a machine learning method until the intersection of the stable intervals is greater than or equal to the preset interval threshold.

[0017] Through the above-described embodiment, the transformer design system achieves dynamic optimization of parameter stability intervals by constructing a matrix of key layer parameters and calculating parameter coupling. This method divides parameter combinations, calculates the intersection of stable intervals, and incorporates machine learning methods to modify these stable intervals, ensuring the reliability of key parameter adjustments. This optimization approach, based on parameter coupling relationships, reduces unnecessary parameter verification calculations and improves computational efficiency.

[0018] In some embodiments, after the step of calculating the final adjustment amount of the buffer layer parameter according to the target adjustment path and the scalable parameter group, the method further includes: Collecting the adjustment status of the buffer layer parameters in real time, and calculating the adjustment deviation value corresponding to the adjustment status based on a preset evaluation model; update the stable interval based on the adjustment deviation value, and adjust the safety margin range and the value range of the scalable parameter group according to the updated stable interval, wherein the adjustment directions of the safety margin range and the scalable parameter group are inversely related to the variation trend of the adjustment deviation value.

[0019] Through the above embodiments, the transformer design system realizes adaptive optimization of the parameter adjustment range by real-time collection of the adjustment state and calculation of the adjustment deviation value. The system can dynamically update the stable interval and the safety margin range according to the adjustment deviation, ensuring the accuracy of parameter adjustment. This adaptive adjustment mechanism reduces the number of repeated calculations and verifications in the parameter adjustment process, improves the efficiency of parameter adjustment, and ensures the reliability of the adjustment result.

[0020] In a second aspect, the present application provides a transformer design system, comprising: one or more processors and a memory; The memory is coupled to the one or more processors, and the memory is configured to store computer program codes, the computer program codes comprising computer instructions, and the one or more processors invoke the computer instructions so that the transformer design system can implement the flexible control method for transformer design parameter adjustment provided in the above embodiments, which will not be described herein.

[0021] In a third aspect, the present application provides a computer-readable storage medium comprising instructions, which, when executed on a transformer design system, cause the transformer design system to implement the flexible control method for transformer design parameter adjustment provided in the above embodiments, which will not be described herein.

[0022] In a fourth aspect, the present application provides a computer program product, which, when executed on a transformer design system, causes the transformer design system to implement the flexible control method for transformer design parameter adjustment provided in the above embodiments, which will not be described herein.

[0023] The one or more technical solutions provided in the embodiments of the present application have at least the following technical effects or advantages: 1. By establishing a parameter-performance mapping database, the parameters are divided into key layer, buffer layer and adaptive layer according to the influence degree, and the parameter level is dynamically adjusted in combination with the performance optimization target. This mechanism determines the stable interval of the key parameters by using genetic algorithm, and uses dynamic programming algorithm to find the optimal adjustment path, realizing hierarchical management and optimization of parameter adjustment. This hierarchical adjustment strategy avoids full-quantity parameter calculation, reduces the consumption of computing resources, and ensures the reliability of the adjustment result.

[0024] 2、Adopt hierarchical storage scheme based on access frequency, store high-frequency access data in memory, and low-frequency data in hard disk, and realize intelligent migration of data through parameter correlation model. The system can automatically identify data access mode, predict possible data loading to memory in advance, and realize efficient utilization of computing resources. This intelligent data management method greatly improves the data processing efficiency in parameter adjustment process.

[0025] 3、By identifying the key nodes in the parameter adjustment process, combined with parameter sensitivity matrix and coupling degree analysis, the local fine calculation of adjustment path is realized. The system can real-time monitor the adjustment state, dynamically update the parameter stable interval and safety margin range, and adjust the optimization strategy according to the adjustment deviation. This adaptive optimization method not only ensures the adjustment accuracy of key positions, but also avoids a large amount of redundant calculation, and improves the efficiency of parameter adjustment. BRIEF DESCRIPTION OF DRAWINGS

[0026] Figure 1 is a flowchart of a flexible control method for transformer design parameter adjustment in an embodiment of the present application; Figure 2 is another flowchart of a flexible control method for transformer design parameter adjustment in an embodiment of the present application; Figure 3 is a schematic diagram of an entity device structure of a transformer design system in an embodiment of the present application. DETAILED DESCRIPTION

[0027] The terms used in the following embodiments of the present application are only for the purpose of describing the specific embodiments, and are not intended to be limiting on the present application. As used in the specification and the appended claims of the present application, the singular forms "a," "an," and "the" are intended to include both singular and plural forms, unless the context clearly indicates otherwise. It should also be understood that the term "and / or" used in the present application means any or all possible combinations of one or more listed items.

[0028] Hereinafter, the terms "first" and "second" are only for the purpose of description, and cannot be understood as implying or suggesting relative importance or implicitly indicating the number of indicated technical features. Therefore, the features defined with "first" and "second" can explicitly or implicitly include one or more of the features, and in the description of the embodiments of the present application, the meaning of "multiple" is two or more, unless otherwise specified.

[0029] For the sake of understanding, the method provided by the present embodiment is described in the flow. Please refer to Figure 1 is a flowchart of a flexible control method for transformer design parameter adjustment in an embodiment of the present application.

[0030] S101, acquire corresponding data of each design parameter and performance index of the transformer, establish a feature mapping database between the parameters and the performance, and calculate an influence deviation value of each design parameter on the performance index based on the feature mapping database.

[0031] Wherein, the transformer design parameters represent various technical parameters that can be adjusted when designing the transformer, such as winding turns, core material, coil wire diameter, etc.; the performance index refers to various parameters for measuring the performance of the transformer, such as voltage ratio accuracy, short-circuit impedance, no-load loss, load loss, etc.; the feature mapping database refers to a data set for storing the corresponding relationship between each design parameter and the performance index of the transformer; and the influence deviation value is used to represent the deviation value caused by the change of each design parameter on the performance index.

[0032] Specifically, the transformer design system first acquires corresponding data of each design parameter and performance index of the transformer from various data sources, which can include previous design project documents, experimental test data records, etc. Then the acquired data is sorted and stored to construct a feature mapping database between the parameters and the performance. After the construction is completed, the system analyzes the influence of the change of each design parameter on various performance indexes based on the database through a specific calculation method, thereby calculating the influence deviation value of each design parameter on the performance index. For example, when changing the winding turns, the system will analyze the change of the performance indexes such as voltage ratio accuracy, no-load loss, etc. according to the existing data in the database, and then obtain the influence deviation value of the winding turns on each performance index.

[0033] Optionally, the system can be connected to the design data server of the enterprise through the network interface to acquire the design data stored on the server. The data is stored in the database using a database management system such as MySQL, and the corresponding table structure is established to store the corresponding relationship between the design parameters and the performance indexes, forming a feature mapping database. Then SQL query statements are written to calculate the average value, standard deviation and other statistical quantities of the performance index under different design parameter values, analyze the influence of the design parameters on the performance index, and then obtain the influence deviation value.

[0034] It can be understood that other ways can also be used to acquire data, establish a database and calculate an influence deviation value, such as directly importing data from professional power design software for processing, etc., which is not limited here.

[0035] In addition, the system can also obtain corresponding data of each design parameter and performance index of the transformer, divide the data into high-frequency and low-frequency mapping data through statistical access frequency, and store the data into the memory cache area and the hard disk cache area respectively to construct a feature mapping database. Then, the system extracts historical adjustment records from the database, uses data analysis methods such as correlation analysis and regression analysis or machine learning algorithms to mine the association between the high-frequency and low-frequency mapping data, and establishes a parameter association model. After that, the system continuously monitors the data access frequency of the memory cache area, and when the data access frequency is lower than the preset frequency threshold, the data is stored in the hard disk cache area, and the data in the hard disk cache area that is strongly associated with the current high-frequency access data is stored in the memory cache area according to the parameter association model, so as to optimize data storage and improve data access efficiency.

[0036] S102, determine the ratio of the influence deviation value and the reference value as the influence degree, and divide the design parameters into key layer parameters, buffer layer parameters and adaptive layer parameters based on the influence degree.

[0037] Specifically, the transformer design system divides the design parameters into different arrays according to the size of the influence degree value and the preset threshold value, respectively representing the key layer parameters, the buffer layer parameters and the adaptive layer parameters. For example, set the influence degree greater than 0.8 as the key layer parameter, between 0.3 and 0.8 as the buffer layer parameter, and less than 0.3 as the adaptive layer parameter.

[0038] S103, calculate the stable interval of the key layer parameter by using the genetic algorithm, and set the safety margin range and the scalable parameter group based on the stable interval and the buffer layer parameter respectively.

[0039] Specifically, the transformer design system uses the genetic algorithm, takes the key layer parameter as the variable of the genetic algorithm, and takes the satisfaction of the transformer performance index with the design requirements as the objective function. Through the iterative calculation of the genetic algorithm, the stable interval of the key layer parameter is searched in the parameter space. After obtaining the stable interval, the system sets the safety margin range according to the preset rules, for example, increases and decreases a certain proportion of values based on the upper and lower limits of the stable interval to obtain the safety margin range. At the same time, according to the characteristics and design requirements of the buffer layer parameters, the scalable parameter group is set, and the parameter values in the parameter group can be adjusted according to the actual situation, which is used for subsequent adjustment of the buffer layer parameters and the adaptive layer parameters.

[0040] Optionally, the system can use the genetic algorithm toolbox of MATLAB. In MATLAB, define the variable range and fitness function of the key layer parameters, configure the parameters of the genetic algorithm such as selection function, cross function, mutation function, etc. Run the genetic algorithm to solve and get the stable interval of the key layer parameters. Then, determine the safety margin range by calculating 5% of the upper and lower limits of the stable interval. For the scalable parameter group, set a group of initial values according to the type of the buffer layer parameters and design experience, and formulate the rules for dynamic adjustment, such as dynamically changing the value of the scalable parameter group according to the adjustment effect of the buffer layer parameters.

[0041] It can be understood that other ways can also be used to implement the calculation of the stable interval by the genetic algorithm, the setting of the safety margin range and the scalable parameter group, such as writing genetic algorithm code by oneself, etc., which are not limited here.

[0042] It should be noted that the genetic algorithm is a random search algorithm simulating the natural evolution process, which finds the optimal solution or approximate optimal solution in the parameter space by simulating genetic operations such as selection, cross and mutation; the stable interval of the key layer parameters refers to the value interval of the key layer parameters that can ensure that the transformer performance indicators meet the design requirements when the key layer parameters vary within a certain range; the safety margin range refers to a range of allowed parameter fluctuations reserved to prevent unexpected situations during parameter adjustment based on the stable interval of the key layer parameters; the scalable parameter group is a set of parameters used to adjust the buffer layer parameters and the adaptive layer parameters, and the value range of the scalable parameter group will change according to the actual situation.

[0043] S104, when the parameter adjustment signal is detected, if the parameter adjustment amount corresponding to the parameter adjustment signal is within the safety margin range, the adjustment cost of each design parameter is calculated by a dynamic programming algorithm, and the target adjustment path with the lowest adjustment cost is determined.

[0044] When the transformer design system receives an externally input parameter adjustment demand (such as a design personnel initiates an adjustment according to new design requirements, or the system monitors that the running state changes need to adjust parameters), and the system checks the parameter adjustment amount in the adjustment demand and judges that it is within the safety margin range, this step is executed.

[0045] Specifically, when the transformer design system detects a parameter adjustment signal, it first determines whether the parameter adjustment amount corresponding to the signal is within a pre-set safety margin. If it is within the range, the system calls a dynamic programming algorithm to calculate the adjustment cost of each design parameter. Based on the actual situation of the transformer design, the system determines the factors that affect the adjustment cost, such as the degree of impact of different parameter adjustments on performance indicators and the computing resource consumption required during the adjustment process. The system then divides the design parameter adjustment process into multiple stages, each corresponding to a different parameter adjustment situation, and uses a dynamic programming algorithm to find the adjustment path with the lowest total cost in these stages. During the calculation process, the system gradually constructs the adjustment path and compares the adjustment costs under different paths, ultimately determining the target adjustment path with the lowest adjustment cost.

[0046] Optionally, the system can first convert the design parameters and safety margins into a MATLAB matrix within the MATLAB environment. Using MATLAB loops and conditional statements, and following the principles of dynamic programming, the cost of each parameter adjustment is calculated step by step. The cost of different paths is compared, and the path with the lowest cost is recorded. During the calculation process, MATLAB's plotting capabilities can be used to visualize the adjusted paths and cost changes for better analysis and verification of the results.

[0047] S105: Calculate the final adjustment amount of the buffer layer parameters according to the target adjustment path and the scalable parameter group.

[0048] Specifically, the transformer design system obtains the target adjustment path obtained in step S104 and the pre-set scalable parameter group. Based on the adjustment information regarding the buffer layer parameters in the target adjustment path and the parameter values ​​in the scalable parameter group, the system determines the final adjustment amount for the buffer layer parameters using specific calculation rules. For example, if the target adjustment path specifies that buffer layer parameter A must be adjusted within a certain range, and a parameter in the scalable parameter group has a scaling effect on the adjustment range of buffer layer parameter A, the system comprehensively considers these factors to calculate the final adjustment amount for buffer layer parameter A.

[0049] Alternatively, the system can be implemented using formulas and macros in Excel. Enter the target adjustment path, scalable parameter group, and buffer layer parameter data into an Excel spreadsheet. Formulas can be set in the spreadsheet to perform calculations based on the specified rules. For example, a multiplication formula can be used to multiply the adjustment value in the target adjustment path by the corresponding value in the scalable parameter group to obtain the adjustment value for the buffer layer parameter. If the calculation rules are complex, Excel's macro functionality can be used to write VBA code to implement more flexible calculation logic.

[0050] In addition, after calculating the final adjustment amount of the buffer layer parameters and making the adjustment, the transformer design system can also collect the adjustment state of the buffer layer parameters in real time, including the actual value of the parameters, the adjustment time and amplitude, and the like, and input these data into a preset evaluation model, calculate the adjustment deviation value by comparing the actual performance index after adjustment with the performance index required by the design, and then update the stable interval according to the adjustment deviation value. If the deviation value is large, the stable interval is reduced to improve the adjustment accuracy, and if the deviation value is small, the stable interval is appropriately expanded to increase the adjustment flexibility. Finally, according to the updated stable interval, the value range of the safety margin range and the scalable parameter group is adjusted according to the principle of being inversely related to the change trend of the adjustment deviation value, that is, the value range of the safety margin range and the scalable parameter group is reduced when the deviation value increases, and the value range of the safety margin range and the scalable parameter group is expanded when the deviation value decreases, so as to optimize the subsequent parameter adjustment strategy.

[0051] In the above embodiment, the transformer design system realizes the hierarchical management of the transformer design parameters by establishing a parameter-performance feature mapping database and hierarchically classifying the parameters based on the influence degree. The stable interval of the key parameters is determined in combination with the genetic algorithm, and the optimal adjustment path is found by the dynamic programming algorithm, so that the parameter adjustment process no longer needs to calculate all parameters. This method reduces the consumption of computing resources in the design parameter adjustment process through hierarchical adjustment and optimization of the path selection, while ensuring the rationality and reliability of the parameter adjustment, and improving the efficiency of the parameter adjustment in the transformer design process.

[0052] The method provided by the embodiment will be further described in a more specific flow. Please refer to Figure 2 for another flowchart of the flexible control method of the transformer design parameter adjustment in the embodiment of the present application.

[0053] S201, calculate the deviation amount between the performance optimization target and the actual performance index in the current design stage, and determine the dominant direction of performance optimization according to the deviation amount.

[0054] In the transformer design process, whenever a new design stage is entered or the current design is adjusted and optimized, the direction of performance optimization needs to be determined, and this step is executed at this time. The performance optimization target represents the ideal value of each performance index expected to be achieved in the current transformer design stage, which is set according to the design requirements and related standards.

[0055] Specifically, the transformer design system first acquires the performance optimization targets preset in the current design stage and the actual performance indicators obtained through actual measurement, calculation or extraction from existing design data. Then, each indicator in the performance optimization targets is subtracted from the corresponding actual performance indicator to obtain the deviation of each performance indicator. For example, if the voltage ratio accuracy requirement in the performance optimization targets is 99.5%, and the actual measured voltage ratio accuracy is 98%, the deviation between the two is 1.5%. Next, the system analyzes these deviation amounts and determines the performance indicators with larger deviation amounts. The direction of improving or improving these indicators is determined as the dominant direction of performance optimization.

[0056] S202, based on the dominant direction, the influence degree is weighted and calculated, the influence degree is calculated, and the number of level conversion times of each design parameter in the historical adjustment process is counted to calculate the stability of the parameter level.

[0057] Specifically, the transformer design system first acquires the performance optimization main direction determined in step S201 and the influence degree of each design parameter on the performance indicators calculated in step S102. According to the dominant direction, the system determines the weight of different design parameters. For example, if the current performance optimization dominant direction is to improve the voltage ratio accuracy, the weight of the design parameter (such as the number of turns of the winding) that has a greater influence on the voltage ratio accuracy will be relatively high. Then, the system multiplies the influence degree of each design parameter by the corresponding weight to perform weighted calculation and obtain the corrected influence degree.

[0058] At the same time, the system extracts the level conversion information of each design parameter in the past adjustment process from the design history record, and counts the level conversion times of each design parameter. For example, a certain design parameter has undergone level conversion for 3 times in the past 5 adjustments, so its level conversion times is 3. Finally, according to the level conversion times, the stability of the parameter level is calculated through a specific formula. Assuming that the level conversion times is n and the total adjustment times is N, the stability can be calculated by formula 1-n / N, wherein N can be determined according to the total adjustment operation times in the historical record.

[0059] S203, according to the weighted value of the corrected influence degree and the stability, the design parameters are dynamically layered.

[0060] Specifically, the transformer design system obtains the correction influence degree and stability of each design parameter in step S202. The system determines the weight of the correction influence degree and the stability according to design experience or pre-set rules. For example, the weight of the correction influence degree is set to 0.6, and the weight of the stability is set to 0.4. Then, the system multiplies the correction influence degree of each design parameter by the corresponding weight, and adds the stability multiplied by the corresponding weight to obtain a comprehensive weighted value. For example, the correction influence degree of a certain design parameter is 0.7, and the stability is 0.8. According to the above weight calculation, the comprehensive weighted value is 0.74. Finally, according to the size of the comprehensive weighted value, the design parameters are divided into different levels, including key layer parameters, buffer layer parameters, and adaptive layer parameters. Optionally, it can be set that the comprehensive weighted value greater than 0.8 is a key layer parameter, the comprehensive weighted value between 0.5 and 0.8 is a buffer layer parameter, and the comprehensive weighted value less than 0.5 is an adaptive layer parameter, which is not limited here.

[0061] In the above embodiment, the transformer design system realizes dynamic optimization of parameter layering by calculating the deviation amount of the performance optimization target and the actual index, and combining the parameter influence degree and the historical adjustment record. This method considers the dominant direction of performance optimization and the stability of parameter level, so that the parameter layering is more accurate and reasonable. This dynamic layering mechanism reduces the invalid calculation during parameter adjustment, so that the calculation resources are concentrated on the optimization of key parameters, and the accuracy and efficiency of parameter adjustment are improved.

[0062] S204, construct a mutual influence matrix between the key layer parameters, and calculate the coupling degree of each key layer parameter based on the mutual influence matrix.

[0063] Specifically, the transformer design system analyzes the mutual influence relationship between each two key layer parameters for the determined key layer parameters. For example, if the key layer parameters include winding turns, core material, etc., the system analyzes the influence of winding turns on the performance of core material, and the influence of core material on the performance related to winding turns. According to these influence relationships, a mutual influence matrix is constructed. The elements in the matrix can be determined by experimental data, theoretical analysis or historical design experience. After the mutual influence matrix is constructed, the system calculates the coupling degree between each key layer parameter based on the matrix by using a specific algorithm (such as correlation analysis algorithm). For example, by calculating the correlation coefficient and the like, the coupling degree between the key layer parameters is quantified to obtain the coupling degree value of each key layer parameter with other key layer parameters.

[0064] Optionally, the system can input the key layer parameters into the MATLAB workspace under the MATLAB environment, and generate an empty matrix matching the number of key layer parameters using the matrix creation function of MATLAB. By writing a loop and conditional statement, the mutual influence matrix is assigned values according to the influence relationship between the key layer parameters. Then, the correlation coefficient between each key layer parameter is calculated using the corrcoef function of MATLAB, which is used as a quantitative indicator of coupling degree. It can be understood that other methods can also be used, such as using the parameter correlation analysis function provided in professional circuit design analysis software, which is not limited here.

[0065] S205, according to the coupling degree, the key layer parameters are divided into a plurality of parameter combinations, and the stable interval intersection of each parameter combination is calculated.

[0066] The transformer design system first sets a coupling degree threshold according to the coupling degree of each key layer parameter calculated in step S204. The key layer parameters with a coupling degree greater than the threshold are divided into a group to form a plurality of parameter combinations. For example, if the coupling degree threshold is set to 0.7, when the coupling degree between parameters A and B is 0.8 and the coupling degree between parameters C and D is 0.75, A and B are divided into a group, and C and D are divided into a group. After dividing the parameter combinations, the system calculates the stable interval for each parameter combination. The determination method of the stable interval is the same as the principle of calculating the stable interval of the key layer parameters using the genetic algorithm (such as step S103), which will not be repeated here. Finally, the intersection of the stable intervals of different parameter combinations is calculated to obtain a common value range, and the stable interval intersection is obtained.

[0067] S206, when it is detected that the stable interval intersection is less than the preset interval threshold, the original stable interval is modified based on the machine learning method until the stable interval intersection is greater than or equal to the preset interval threshold.

[0068] Specifically, the transformer design system compares the stable interval intersection calculated in step S205 with the preset interval threshold. If the stable interval intersection is less than the preset interval threshold, the system starts the machine learning algorithm. Optionally, a neural network algorithm is selected, and the key layer parameters and their related performance indicator data are input into the neural network model for training. During the training process, the model learns the complex relationship between the key layer parameters and the performance indicators. Based on the trained model, the original stable interval is adjusted and modified. After each modification, the stable interval intersection of the parameter combination is recalculated and compared with the preset interval threshold again. This is repeated until the stable interval intersection is greater than or equal to the preset interval threshold, and the stable interval obtained at this time is the reasonable stable interval after modification.

[0069] In the above embodiment, the transformer design system realizes dynamic optimization of the parameter stable interval by constructing the mutual influence matrix of the key layer parameters and calculating the parameter coupling degree. This method divides the parameter combination and calculates the intersection of the stable interval, and combines the machine learning method to correct the stable interval, ensuring the reliability of the key parameter adjustment. This optimization method based on the parameter coupling relationship reduces unnecessary parameter verification calculation and improves the calculation efficiency.

[0070] S207, if the parameter adjustment amount corresponding to the parameter adjustment signal is within the safety margin range, the parameter adjustment amount is divided into multiple adjustment subintervals, and the initial adjustment direction of each adjustment subinterval is determined by using a simplified calculation method to generate an initial adjustment path.

[0071] Specifically, when the system confirms that the parameter adjustment amount is within the safety margin range, the parameter adjustment amount is divided into multiple adjustment subintervals according to the preset division rule. The division rule can be determined according to the characteristics of the parameters, previous design experience, or related algorithms, such as dividing the adjustment amount into several equidistant subintervals, or dividing it unequally according to the sensitivity of the parameter to the performance index. For each adjustment subinterval, the system uses a simplified calculation method to determine its initial adjustment direction. This simplified calculation method can be an empirical formula calculation based on historical data, or it can be a judgment based on the general relationship between the parameters and the performance index. For example, if it is known that when a certain parameter increases, a certain performance index will change in the direction of optimization, then in this adjustment subinterval, the direction of parameter increase is preliminarily determined as the initial adjustment direction. Finally, the system connects the initial adjustment directions of each adjustment subinterval in turn to form an initial adjustment path containing multiple adjustment nodes, providing a basic framework for subsequent optimization calculations.

[0072] S208, calculate the influence degree of each adjustment node on the adjustment cost, calculate the parameter sensitivity matrix of the key node based on the influence degree, and determine the adjustment priority of each design parameter at the key node based on the parameter sensitivity matrix.

[0073] Specifically, the transformer design system calculates the influence degree of each adjustment node on the initial adjustment path on the adjustment cost according to the pre-set adjustment cost calculation rule. These rules may take into account the influence of parameter adjustment on performance indicators, the computing resources required during adjustment, the risks that may be brought about after adjustment, and other factors. For example, if the parameter adjustment of a certain adjustment node will cause the performance indicators to deviate significantly from the design requirements, while a large amount of computing resources are required to determine the new parameter value, then the influence degree of the adjustment node on the adjustment cost is greater. The system compares the influence degree of each adjustment node with the pre-set influence threshold, and determines the adjustment nodes with an influence degree greater than the pre-set influence threshold as key nodes. For these key nodes, the system calculates the parameter sensitivity matrix based on the influence degree. In the calculation, the influence of a slight change in each design parameter at the key node on each performance indicator is analyzed, and these influence degrees are quantified and filled into the matrix. Based on the obtained parameter sensitivity matrix, the system determines the adjustment priority of each design parameter at the key node. For example, at a certain key node, if the influence degree of changing the number of turns on the voltage ratio accuracy is greater than the influence degree of changing the core material on the voltage ratio accuracy, then the adjustment priority of the number of turns is higher than that of the core material.

[0074] S209, according to the adjustment priority, the adjustment sub-interval of the key node is subdivided, the encrypted adjustment point sequence is generated, and the local optimal adjustment path is calculated based on the encrypted adjustment point sequence, the local optimal adjustment path is replaced with the adjustment path segment at the original key node, and the improved adjustment path is generated.

[0075] The transformer design system subdivides the adjustment sub-interval of the key node according to the adjustment priority determined in step S208. For example, if a certain key node involves multiple design parameters, and according to the adjustment priority, it is determined that a certain parameter is adjusted first, then the adjustment sub-interval corresponding to the parameter is subdivided more carefully. By subdividing the adjustment sub-interval, the encrypted adjustment point sequence is generated, and increasing the density of adjustment points can more accurately explore the parameter space. Based on this encrypted adjustment point sequence, the system calculates the local optimal adjustment path using an optimization algorithm (such as a dynamic programming algorithm or a greedy algorithm). During the calculation, factors such as adjustment cost, performance indicator changes, and other factors are considered to find the path with the lowest adjustment cost in the local area. Finally, the local optimal adjustment path calculated is replaced with the path segment at the key node in the original initial adjustment path, thereby generating an improved adjustment path, making the overall parameter adjustment path more reasonable and optimized.

[0076] S210, when it is detected that the difference between the adjustment costs calculated adjacent to each other is less than the pre-set convergence threshold, the current improved adjustment path is determined as the target adjustment path; if the difference between the adjustment costs is greater than the pre-set convergence threshold, the adjustment nodes are further calculated according to the size order of the influence degree.

[0077] Specifically, the transformer design system calculates the adjustment cost of the current improved adjustment path after each time the calculation of the improved adjustment path is completed, and the difference between the adjustment cost calculated this time and the adjustment cost calculated last time. The system compares the difference with the preset convergence threshold. If the difference is less than the preset convergence threshold, it is determined that the adjustment cost has tended to be stable, and the effect of continuous optimization is not obvious. At this time, the system considers that a better adjustment path has been found, and determines the current improved adjustment path as the target adjustment path. If the difference is greater than the preset convergence threshold, it is determined that the adjustment cost has not reached a stable state, and the adjustment path still has optimization space. The system continues to calculate the adjustment nodes that have not been fully optimized in the order of the size of the adjustment node influence degree. For example, the adjustment nodes with greater influence degree but not yet accurately calculated are calculated and optimized again, and the iteration is continuously performed until the difference between the adjustment costs calculated in the adjacent two times is less than the preset convergence threshold, and the final target adjustment path is determined.

[0078] In the above embodiment, the transformer design system realizes the hierarchical optimization of the parameter adjustment path by dividing the parameter adjustment amount into multiple subintervals and identifying the key nodes. The system only performs fine calculation on the key nodes with greater influence degree, and uses a simplified calculation method for other nodes, thereby greatly reducing the calculation amount. At the same time, the iterative optimization method ensures the convergence of the adjustment path, thereby significantly reducing the consumption of calculation resources while ensuring the adjustment quality.

[0079] The transformer design system of the embodiment of the present application is applied to an electronic device, Figure 3 An architectural schematic diagram of an electronic device suitable for implementing the embodiment of the present application is shown.

[0080] It should be noted that, Figure 3 The electronic device shown is only an example, and should not impose any limitation on the functions and use range of the embodiment of the present application.

[0081] Those skilled in the art can understand that all or part of the steps in the various methods of the above embodiments can be completed by instructions (computer programs) or controlled by related hardware through instructions (computer programs). The instructions can be stored in a computer readable storage medium and loaded and executed by a processor. The electronic device of the embodiment of the present application includes a storage medium and a processor, wherein the storage medium stores a plurality of instructions, and the instructions can be loaded by the processor to execute any step of the method provided by the embodiment of the present application.

[0082] Specifically, the storage medium and the processor are directly or indirectly electrically connected to realize the transmission or interaction of data. For example, the elements can be electrically connected to each other through one or more signal lines. The storage medium stores computer execution instructions for realizing the data access control method, including at least one software function module stored in the storage medium in the form of software or firmware. The processor executes various function applications and data processing by running the software program and the module stored in the storage medium. The storage medium can be, but is not limited to, a random access memory (RAM), a read-only memory (ROM), a programmable read-only memory (PROM), an erasable programmable read-only memory (EPROM), an electrically erasable programmable read-only memory (EEPROM), etc. Among them, the storage medium is used to store programs, and the processor executes the programs after receiving execution instructions.

[0083] Further, the software program and the module in the storage medium can also include an operating system, which can include various software components and / or drivers for managing system tasks (such as memory management, storage device control, power management, etc.), and can communicate with various hardware or software components to provide a running environment for other software components. The processor can be an integrated circuit chip with signal processing capability. The processor described above can be a general-purpose processor, including a central processing unit (CPU), a network processor (NP), etc., which can realize or execute the methods, steps and logic flow diagrams disclosed in the embodiments. The general-purpose processor can be a microprocessor or the processor can also be any conventional processor.

[0084] Due to the instructions stored in the storage medium, the steps in any method provided by the embodiments of the present application can be executed, and thus the beneficial effects of any method provided by the embodiments of the present application can be achieved. Details are described in the foregoing embodiments, which will not be repeated here.

[0085] The above merely describes preferred specific embodiments of the present application, but the protection scope of the present application is not limited thereto, any changes or replacements within the technical scope disclosed by the present application, which can be easily thought by those skilled in the art, should be covered within the protection scope of the present application. Therefore, the protection scope of the present application should be subject to the protection scope of the claims.

Claims

1. A flexible control method for adjusting transformer design parameters, applied to a transformer design system, characterized in that: The method comprises: Obtain corresponding data of each design parameter and performance index of the transformer, establish a parameter-performance feature mapping database, and calculate the deviation value of the influence of each design parameter on the performance index based on the feature mapping database; Determining the ratio of the impact deviation value to the reference value as the impact degree, and dividing the design parameters into key layer parameters, buffer layer parameters, and adaptation layer parameters based on the impact degree; Calculating a stable interval of the key layer parameters using a genetic algorithm, and setting a safety margin range and a scalable parameter group based on the stable interval and the buffer layer parameters, respectively, wherein the scalable parameter group is used to adjust the buffer layer parameters and the adaptation layer parameters; When a parameter adjustment signal is detected, if the parameter adjustment amount corresponding to the parameter adjustment signal is within the safety margin range, the adjustment cost of each design parameter is calculated using a dynamic programming algorithm, and a target adjustment path with the lowest adjustment cost is determined; A final adjustment amount of the buffer layer parameter is calculated according to the target adjustment path and the scalable parameter group.

2. The method according to claim 1, characterized in that The step of obtaining corresponding data of each transformer design parameter and performance index and establishing a parameter-performance feature mapping database specifically includes: The corresponding data is divided into high-frequency mapping data and low-frequency mapping data according to the access frequency and stored in the memory cache area and the hard disk cache area respectively to generate a feature mapping database; Based on the historical adjustment records in the feature mapping database, a parameter association model for identifying the association relationship between the high-frequency mapping data and the low-frequency mapping data is established using a data analysis method; When the access frequency of the mapping data in the memory cache area is lower than the preset frequency threshold, the mapping data is transferred to the hard disk cache area, and the mapping data in the hard disk cache area that has a strong correlation with the current high-frequency access data is transferred to the memory cache area.

3. The method according to claim 1, characterized in that The step of dividing the design parameters into key layer parameters, buffer layer parameters, and adaptation layer parameters based on the influence degree specifically includes: Calculating the deviation between the performance optimization target of the current design stage and the actual performance index, and determining the dominant direction of performance optimization based on the deviation; Performing weighted calculation on the influence degree based on the dominant direction to obtain a modified influence degree; Count the number of level conversions of each design parameter during the historical adjustment process and calculate the stability of the parameter level; The design parameters are dynamically layered according to the weighted values ​​of the correction influence degree and the stability.

4. The method according to claim 1, wherein The step of calculating the adjustment cost of each design parameter by a dynamic programming algorithm and determining the target adjustment path with the lowest adjustment cost specifically includes: Dividing the parameter adjustment amount into a plurality of adjustment subintervals, and determining an initial adjustment direction of each adjustment subinterval using a simplified calculation method to generate an initial adjustment path, wherein the initial adjustment path includes a plurality of adjustment nodes; Calculating the degree of influence of each adjustment node on the adjustment cost, and determining the adjustment nodes whose influence is greater than a preset influence threshold as key nodes for refined calculation to generate an improved adjustment path; When it is detected that the difference in adjustment costs between two adjacent calculations is less than a preset convergence threshold, the current improved adjustment path is determined as the target adjustment path; if the adjustment cost difference is greater than the preset convergence threshold, the adjustment nodes are further refined in order of their impact.

5. The method according to claim 4, characterized in that The step of determining the adjustment nodes whose impact is greater than a preset impact threshold as key nodes and performing refined calculations to generate an improved adjustment path specifically includes: Calculating a parameter sensitivity matrix of the key node based on the degree of influence, and determining an adjustment priority of each design parameter at the key node based on the parameter sensitivity matrix; Subdivide the adjustment subintervals of the key nodes according to the adjustment priorities to generate an encrypted adjustment point sequence; A local optimal adjustment path is calculated based on the encrypted adjustment point sequence, and the adjustment path segment at the original key node is replaced by the local optimal adjustment path to generate an improved adjustment path.

6. The method according to claim 1, characterized in that After the step of calculating the stable interval of the key layer parameters using a genetic algorithm, the method further includes: Constructing a mutual influence matrix between the key layer parameters, and calculating the coupling degree of each key layer parameter based on the mutual influence matrix; Dividing the key layer parameters into multiple parameter combinations according to the coupling degree, and calculating the intersection of the stable intervals of each parameter combination; When it is detected that the intersection of the stable intervals is less than a preset interval threshold, the original stable intervals are modified based on a machine learning method until the intersection of the stable intervals is greater than or equal to the preset interval threshold.

7. The method according to claim 1, characterized in that After the step of calculating the final adjustment amount of the buffer layer parameter according to the target adjustment path and the scalable parameter group, the method further includes: Collecting the adjustment status of the buffer layer parameters in real time, and calculating the adjustment deviation value corresponding to the adjustment status based on a preset evaluation model; The stable interval is updated based on the adjusted deviation value, and the safety margin range and the value range of the scalable parameter group are adjusted according to the updated stable interval, where the adjustment direction of the safety margin range and the scalable parameter group is inversely correlated with the changing trend of the adjusted deviation value.

8. A transformer design system, characterized in that: The transformer design system includes: one or more processors and memory; The memory is coupled to the one or more processors, and is configured to store computer program codes, where the computer program codes include computer instructions. The one or more processors call the computer instructions to enable the transformer design system to execute the method according to any one of claims 1 to 7.

9. A computer-readable storage medium comprising instructions, characterized in that: When the instruction is executed on a transformer design system, the transformer design system is caused to execute the method according to any one of claims 1 to 7.

10. A computer program product, characterized in that When the computer program product is run on a transformer design system, the transformer design system is caused to perform the method according to any one of claims 1 to 7.

Images