A method for acquiring parameter values of a flexible press-pack IGBT and related device

Abstract

This invention discloses a method and related apparatus for obtaining parameter values ​​of a resiliently pressed IGBT. The method includes: constructing a total impedance model of the resiliently pressed IGBT based on its physical structure and physical characteristics in the off-state; obtaining the total impedance curve of the resiliently pressed IGBT under actual operating pressure conditions; using unknown parameters in the total impedance model as parameters of the resiliently pressed IGBT to be obtained; and using a genetic algorithm based on the total impedance model and the total impedance curve to obtain the values ​​of the resiliently pressed IGBT parameters. This invention comprehensively considers the physical structure and physical characteristics of the resiliently pressed IGBT in the off-state to construct a total impedance model. Based on the total impedance model and the measured total impedance curve, parameter values ​​that match the total impedance curve can be obtained, providing strong support for analyzing the performance of resiliently pressed IGBTs in complex circuits.

Description

Technical Field

[0001] This invention relates to a method and related apparatus for obtaining parameter values ​​of elastically press-fit IGBTs, belonging to the field of transistor technology. Background Technology

[0002] Flexible-press IGBTs (Insulated Gate Bipolar Transistors), as a type of press-fit IGBT, optimize pressure compensation and stress absorption to some extent by introducing a disc spring structure. This reduces the requirements for process precision and ensures uniform pressure on the chip surface, achieving a certain balance between cost control and performance optimization. However, to analyze the performance of flexible-press IGBTs in complex circuits, it is necessary to pre-determine their parameter values, but currently, no such method exists. Summary of the Invention

[0003] This invention provides a method and related apparatus for obtaining parameter values ​​of elastically press-fit IGBTs, which solves the problems disclosed in the background art.

[0004] According to one aspect of this application, a method for obtaining parameter values ​​of a resiliently pressed IGBT is provided, comprising:

[0005] Based on the physical structure and physical characteristics of the elastically press-fit IGBT in the off state, a total impedance model of the elastically press-fit IGBT is constructed.

[0006] Obtain the total impedance curve of the elastically press-fit IGBT under actual working pressure conditions;

[0007] The unknown parameters in the total impedance model are used as the parameters of the elastically pressed IGBT to be obtained. Based on the total impedance model and the total impedance curve, a genetic algorithm is used to obtain the values ​​of the elastically pressed IGBT parameters.

[0008] Furthermore, the total impedance model for the elastically press-fit IGBT is as follows:

[0009] ;

[0010] In the formula, Z(jω) is the total impedance of the flexible IGBT, ω is the angular frequency, R1 and L1 are the stray resistance and stray inductance of the lead wire, R2 is the equivalent resistance of the internal buffer layer and drift region of the flexible IGBT, C2 is the equivalent capacitance of the anti-parallel diode and parasitic capacitance, R3 and C3 are the equivalent resistance and equivalent capacitance of the PN junction dielectric loss, C1 is the equivalent capacitance formed by the PN junction in the off state of the flexible IGBT, R4 is the equivalent resistance of the dielectric loss of the anti-parallel diode, C1, R2 and R3 form the broadband model of the flexible IGBT, and C2 and R4 form the broadband equivalent model of the anti-parallel diode.

[0011] Furthermore, the unknown parameters in the total impedance model are used as the parameters of the elastically pressed IGBT to be obtained. Based on the total impedance model and the total impedance curve, a genetic algorithm is used to obtain the values ​​of the elastically pressed IGBT parameters, including:

[0012] 1) Initialize the population; the population includes multiple individuals, each individual is a set of elastic compression IGBT parameters whose values ​​are to be acquired. The initialization is to assign values ​​to each elastic compression IGBT parameter according to the individual number and the elastic compression IGBT parameter number in the individual.

[0013] 2) Based on the current iteration stage and the mutation factor of the previous generation, determine the mutation factor of the current generation. Perform mutation operation on each parameter of the individual according to the mutation factor to generate mutated individuals. In each iteration, when a new optimal individual is found, add the new optimal individual to the list of historical optimal individuals. The optimal individual is the individual with the lowest fitness.

[0014] 3) Based on the diversity index of the current population, determine the crossover probability of the current population, and perform crossover operation on each parameter of the mutated individuals according to the crossover probability to generate crossover individuals;

[0015] 4) Calculate the fitness of the crossover individuals and the fitness of the corresponding original individuals based on the total impedance model and total impedance curve; where the original individuals are those between those without mutation and crossover operations.

[0016] 5) Based on the fitness of the crossover individuals and the fitness of the corresponding original individuals, select individuals from the crossover individuals and the original individuals to be used as the next generation population. If the best individual in the next generation population meets the preset requirements or the number of iterations has reached the preset total number of iterations, then the value corresponding to the best individual in the next generation population is used as the value of the elastic compression IGBT parameter; otherwise, the next generation population is used as the current population, and go to 2).

[0017] Furthermore, the initialization formula is:

[0018] ;

[0019] In the formula, The parameter is the p-th elastic compression IGBT parameter for the q-th individual, r is a random number calculated based on a string hash value, and the string is a string constructed from the ID of the q-th individual and the ID of the p-th elastic compression IGBT parameter in the q-th individual. These are the upper and lower bounds of the parameters for the p-th elastically pressed IGBT, respectively.

[0020] Furthermore, in the genetic algorithm, the iterative process is divided into the first iteration stage, the second iteration stage, and the third iteration stage according to the time sequence of the iterations;

[0021] Based on the current iteration stage and the variation factors of the previous generation, determine the variation factors of the current population, including:

[0022] When the current iteration is in the first iteration phase, if the fitness of the best individual in the previous generation is O... f Average fitness A of individuals from the previous generation f If the difference is less than the first threshold, then the variation factor of the current population is equal to the sum of the variation factor of the previous generation and the first step length. If O f With A f If the difference is greater than the first threshold and less than the second threshold, then the variation factor of the current population is equal to the sum of the variation factor of the previous generation and the second step size; if O f With A f If the difference is greater than the second threshold, then the variation factor of the current generation population is equal to the variation factor of the previous generation population; where the first step length is greater than the second step length; the variation factor of the first generation population is a preset value.

[0023] When the current iteration is in the second iteration phase, if the standard deviation σ of the fitness in the list of historical best individuals corresponding to the previous generation population is... W If the variance is less than the third threshold, then the variance factor of the current population is equal to the difference between the variance factor of the previous generation and the third step length. If σ W If the variance factor is greater than the third threshold and less than the fourth threshold, then the variance factor of the current generation population is equal to that of the previous generation population. W If the variation factor is greater than the fourth threshold, then the variation factor of the current population is equal to the sum of the variation factor of the previous generation population and the fourth step length.

[0024] When the current iteration is in the third iteration stage, the mutation factor of the current population is equal to the mutation factor of the previous generation population minus 0.2×(g-2G / 3); where g is the current iteration number and G is the preset total iteration number.

[0025] Furthermore, based on the diversity indicators of the current population, the crossover probability of the current population is determined, including:

[0026] If the diversity index of the current population is greater than the high diversity threshold, the crossover probability of the current population is equal to the product of the initial crossover probability and the fifth step length.

[0027] If the diversity index of the current population is less than the low diversity threshold, the crossover probability of the current population is equal to the product of the initial crossover probability and the sixth step length; where the sixth step length is greater than the fifth step length.

[0028] If the diversity index of the current population is less than the high diversity threshold but greater than the low diversity threshold, the crossover probability of the current population is equal to the initial crossover probability.

[0029] Furthermore, based on the total impedance model and the total impedance curve, the fitness of the crossover individuals and the fitness of the corresponding original individuals are calculated, including:

[0030] Substitute the cross individuals into the total impedance model to obtain the theoretical total impedance. Calculate the fitness of the cross individuals based on the theoretical total impedance and the actual total impedance in the total impedance curve.

[0031] Substitute the original individual into the total impedance model to obtain the theoretical total impedance. Based on the theoretical total impedance and the actual total impedance in the total impedance curve, calculate the fitness of the original individual.

[0032] Furthermore, based on the fitness of the crossover individuals and the fitness of the corresponding original individuals, individuals are selected from the crossover individuals and the original individuals to serve as the next generation population, including:

[0033] Select a predetermined number of minimum fitness values ​​from the fitness values ​​of the crossover individuals and the fitness values ​​of the corresponding original individuals, and use the selected minimum fitness values ​​as the excellent fitness values.

[0034] The original individuals with excellent fitness will be used as individuals in the next generation of the population;

[0035] The selection strategy for individuals with non-superior fitness is as follows:

[0036] ;

[0037] In the formula, Let q be the individual after the crossover operation in the g-th iteration. for The corresponding original individual, They are respectively and fitness It represents the q-th individual in the (g+1)-th iteration.

[0038] According to another aspect of this application, a device for obtaining parameter values ​​of resiliently pressed IGBTs is provided, comprising:

[0039] The model building module constructs a total impedance model for the elastically press-fit IGBT based on its physical structure and physical characteristics under the cut-off state.

[0040] The curve acquisition module acquires the total impedance curve of the elastically pressed IGBT under actual working pressure conditions.

[0041] The parameter value acquisition module takes the unknown parameters in the total impedance model as the parameters of the elastically pressed IGBT to be acquired, and uses a genetic algorithm to obtain the values ​​of the elastically pressed IGBT parameters based on the total impedance model and the total impedance curve.

[0042] Furthermore, in the model building module, the total impedance model of the elastically press-fit IGBT is:

[0043] ;

[0044] In the formula, Z(jω) is the total impedance of the flexible IGBT, ω is the angular frequency, R1 and L1 are the stray resistance and stray inductance of the lead wire, R2 is the equivalent resistance of the internal buffer layer and drift region of the flexible IGBT, C2 is the equivalent capacitance of the anti-parallel diode and parasitic capacitance, R3 and C3 are the equivalent resistance and equivalent capacitance of the PN junction dielectric loss, C1 is the equivalent capacitance formed by the PN junction in the off state of the flexible IGBT, R4 is the equivalent resistance of the dielectric loss of the anti-parallel diode, C1, R2 and R3 form the broadband model of the flexible IGBT, and C2 and R4 form the broadband equivalent model of the anti-parallel diode.

[0045] Furthermore, the parameter value acquisition module includes:

[0046] The initialization module initializes the population; the population includes multiple individuals, each individual is a set of elastic compression IGBT parameters whose values ​​are to be acquired, and the initialization is to assign values ​​to each elastic compression IGBT parameter according to the individual number and the elastic compression IGBT parameter number in the individual.

[0047] The mutated individual module determines the mutation factor of the current population based on the current iteration stage and the mutation factor of the previous generation. It then performs mutation operations on each parameter of the individual based on the mutation factor to generate mutated individuals. In each iteration, when a new optimal individual is found, it is added to the list of historical optimal individuals. The optimal individual is the one with the lowest fitness.

[0048] The crossover individual module determines the crossover probability of the current population based on the diversity index of the current population, and performs crossover operations on each parameter of the mutated individuals according to the crossover probability to generate crossover individuals;

[0049] The fitness module calculates the fitness of crossover individuals and the fitness of the corresponding original individuals based on the total impedance model and the total impedance curve; where the original individuals are those between those without mutation and crossover operations.

[0050] The judgment module selects individuals from the crossover individuals and the corresponding original individuals as the next generation population based on the fitness of the crossover individuals and the fitness of the original individuals. If the best individual in the next generation population meets the preset requirements or the number of iterations has reached the preset total number of iterations, the value corresponding to the best individual in the next generation population is used as the value of the elastic compression IGBT parameter. Otherwise, the next generation population is used as the current population and transferred to the mutation individual module for processing.

[0051] Furthermore, in the initialization module, the initialization formula can be expressed as:

[0052] ;

[0053] In the formula, The parameter is the p-th elastic compression IGBT parameter for the q-th individual, r is a random number calculated based on a string hash value, and the string is a string constructed from the ID of the q-th individual and the ID of the p-th elastic compression IGBT parameter in the q-th individual. These are the upper and lower bounds of the parameters for the p-th elastically pressed IGBT, respectively.

[0054] Furthermore, in the genetic algorithm, the iteration process is divided into the first iteration stage, the second iteration stage, and the third iteration stage according to the time sequence of the iterations.

[0055] In the variant individual module, based on the current iteration stage and the variant factors of the previous generation, the variant factors of the current population are determined, including:

[0056] When the current iteration is in the first iteration phase, if the fitness of the best individual in the previous generation is O... f Average fitness A of individuals from the previous generation f If the difference is less than the first threshold, then the variation factor of the current population is equal to the sum of the variation factor of the previous generation and the first step length. If O f With A f If the difference is greater than the first threshold and less than the second threshold, then the variation factor of the current population is equal to the sum of the variation factor of the previous generation and the second step size; if O f With A f If the difference is greater than the second threshold, then the variation factor of the current generation population is equal to the variation factor of the previous generation population; where the first step length is greater than the second step length; the variation factor of the first generation population is a preset value.

[0057] When the current iteration is in the second iteration phase, if the standard deviation σ of the fitness in the list of historical best individuals corresponding to the previous generation population is... W If the variance is less than the third threshold, then the variance factor of the current population is equal to the difference between the variance factor of the previous generation and the third step length. If σ WIf the variance factor is greater than the third threshold and less than the fourth threshold, then the variance factor of the current generation population is equal to that of the previous generation population. W If the variation factor is greater than the fourth threshold, then the variation factor of the current population is equal to the sum of the variation factor of the previous generation population and the fourth step length.

[0058] When the current iteration is in the third iteration stage, the mutation factor of the current population is equal to the mutation factor of the previous generation population minus 0.2×(g-2G / 3); where g is the current iteration number and G is the preset total iteration number.

[0059] Furthermore, in the crossover individual module, the crossover probability of the current population is determined based on the diversity index of the current population, including: if the diversity index of the current population is greater than the high diversity threshold, the crossover probability of the current population is equal to the product of the initial crossover probability and the fifth step length; if the diversity index of the current population is less than the low diversity threshold, the crossover probability of the current population is equal to the product of the initial crossover probability and the sixth step length; wherein, the sixth step length is greater than the fifth step length; if the diversity index of the current population is less than the high diversity threshold and greater than the low diversity threshold, the crossover probability of the current population is equal to the initial crossover probability.

[0060] Furthermore, in the fitness module, the fitness of the cross individual and the fitness of the corresponding original individual are calculated based on the total impedance model and the total impedance curve. This includes: substituting the cross individual into the total impedance model to obtain the theoretical total impedance, and calculating the fitness of the cross individual based on the theoretical total impedance and the actual total impedance in the total impedance curve; and substituting the original individual into the total impedance model to obtain the theoretical total impedance, and calculating the fitness of the original individual based on the theoretical total impedance and the actual total impedance in the total impedance curve.

[0061] Furthermore, in the decision module, individuals are selected from the crossover individuals and the corresponding original individuals to serve as the next generation population, based on the fitness of the crossover individuals and the fitness of the corresponding original individuals. These individuals include:

[0062] Select a predetermined number of minimum fitness values ​​from the fitness values ​​of the crossover individuals and the fitness values ​​of the corresponding original individuals, and use the selected minimum fitness values ​​as the excellent fitness values.

[0063] The original individuals with excellent fitness will be used as individuals in the next generation of the population;

[0064] The selection strategy for individuals with non-superior fitness is as follows:

[0065] ;

[0066] In the formula, Let q be the individual after the crossover operation in the g-th iteration. for The corresponding original individual, They are respectively and fitness It represents the q-th individual in the (g+1)-th iteration.

[0067] According to another aspect of this application, a computer-readable storage medium is provided that stores one or more programs, the one or more programs including instructions that, when executed by a computing device, cause the computing device to perform a method for obtaining parameter values ​​of a resiliently pressed IGBT.

[0068] According to another aspect of this application, a computer device is provided, including one or more processors and one or more memories, wherein one or more programs are stored in the one or more memories and configured to be executed by the one or more processors, and the one or more programs include instructions for performing a method for obtaining parameter values ​​of resiliently pressed IGBTs.

[0069] The beneficial effects achieved by this invention are as follows: This invention comprehensively considers the physical structure and physical characteristics of the flexible press-fit IGBT in the off state, and constructs a total impedance model of the flexible press-fit IGBT. Compared with the traditional model, it can more accurately simulate the electrical characteristics of the flexible press-fit IGBT at different frequencies. Based on the total impedance model and the measured total impedance curve, parameter values ​​that match the total impedance curve can be obtained, which can provide strong support for analyzing the performance of the flexible press-fit IGBT in complex circuits. Attached Figure Description

[0070] Figure 1 A flowchart illustrating the method for obtaining parameter values ​​of elastically press-fit IGBTs;

[0071] Figure 2 The equivalent circuit diagram for a flexible compression IGBT;

[0072] Figure 3 This is a schematic diagram of impedance measurement for a flexible compression-bonded IGBT.

[0073] Figure 4 This is a flowchart of a genetic algorithm;

[0074] Figure 5 A comparison chart of the measured amplitude-frequency results of the total impedance of the elastically pressurized IGBT and the results obtained by the method of the present invention;

[0075] Figure 6 A comparison chart of the measured results of the total impedance phase frequency of the elastically pressurized IGBT and the results of the method of the present invention;

[0076] Figure 7 This is a block diagram of a device for obtaining parameter values ​​of elastically press-fitted IGBTs. Detailed Implementation

[0077] The technical solutions of the embodiments of this application will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of this application, and not all embodiments. The following description of at least one exemplary embodiment is merely illustrative and is in no way intended to limit this application or its application or use. All other embodiments obtained by those skilled in the art based on the embodiments of this application without creative effort are within the scope of protection of this application.

[0078] Unless otherwise specifically stated, the relative arrangement, numerical expressions, and values ​​of the components and steps described in these embodiments do not limit the scope of this application.

[0079] At the same time, it should be understood that, for ease of description, the dimensions of the various parts shown in the accompanying drawings are not drawn according to actual scale.

[0080] Techniques, methods, and equipment known to those skilled in the art may not be discussed in detail, but where appropriate, such techniques, methods, and equipment should be considered part of the specification.

[0081] In all examples shown and discussed herein, any specific values ​​should be interpreted as merely exemplary and not as limitations. Therefore, other examples of exemplary embodiments may have different values.

[0082] It should be noted that similar symbols and letters in the following figures represent similar items; therefore, once an item is defined in one figure, it does not need to be discussed further in subsequent figures.

[0083] Furthermore, in the description of the embodiments of this application, the terms "first," "second," etc., are used only for distinguishing descriptions and should not be construed as indicating or implying relative importance. Therefore, features defined with "first" or "second" may explicitly or implicitly include one or more features.

[0084] To address the current problem of being unable to obtain parameter values ​​for elastically clamped IGBTs, this application proposes a method for obtaining these parameters. This method can be executed by a parameter value acquisition device, which can be a terminal device or a server. The terminal device can include, but is not limited to, mobile phones, computers, etc., as described in this application. The server can be a standalone physical server, a server cluster or distributed system composed of multiple physical servers, or a cloud server providing basic cloud computing services such as cloud services, cloud databases, cloud computing, big data, and artificial intelligence platforms, etc., as described in this application. Optionally, this method can also be executed collaboratively by multiple electronic devices with computing power. For ease of explanation, subsequent embodiments will be described as being executed by a parameter value acquisition device.

[0085] See Figure 1 , Figure 1 This is a flowchart of a method for obtaining parameter values ​​of a resiliently pressed IGBT according to an embodiment of this application. The method can be executed by a parameter value acquisition device and may include at least the following steps:

[0086] Step 1: Based on the physical structure and physical characteristics of the elastically crimped IGBT in the off state, construct the total impedance model of the elastically crimped IGBT.

[0087] It should be noted that the physical structure of a flexible compression IGBT can be leads, buffer layers, drift regions, PN junctions, etc., and its physical characteristics in the off state can be capacitance effects, dielectric losses, etc. An equivalent circuit can be constructed through circuit theory analysis, and a total impedance model can be derived.

[0088] like Figure 2 The diagram shows the equivalent circuit of a flexible IGBT. In the figure, R1 and L1 are the stray resistance and stray inductance of the leads, respectively. R2 is the equivalent resistance of the internal buffer layer and drift region of the flexible IGBT. C2 is the equivalent capacitance of the anti-parallel diode and parasitic capacitance. R3 and C3 are the equivalent resistance and equivalent capacitance of the PN junction dielectric loss, respectively. C1 is the equivalent capacitance formed by the PN junction in the off state of the flexible IGBT. R4 is the equivalent resistance of the dielectric loss of the anti-parallel diode. C1, R2, and R3 form the broadband model of the flexible IGBT, and C2 and R4 form the broadband equivalent model of the anti-parallel diode.

[0089] Therefore, in some embodiments, the total impedance model of the elastically press-fit IGBT can be expressed as:

[0090] ;

[0091] In the formula, Z(jω) is the total impedance of the elastically press-fit IGBT, and ω is the angular frequency.

[0092] The above model not only considers the physical structure of the elastically pressed IGBT, but also its physical characteristics in the off state. Compared with the traditional model, it can more accurately simulate the electrical characteristics of the elastically pressed IGBT at different frequencies, and can provide a reliable basis for in-depth analysis of the behavior of the elastically pressed IGBT in complex circuits.

[0093] Step 2: Obtain the total impedance curve of the elastically crimped IGBT under the actual working pressure environment.

[0094] It should be noted that the total impedance curve can be obtained experimentally; see [link to relevant documentation]. Figure 3 Specifically, specialized pressure-applying equipment such as clamps can be used to apply appropriate pressure to the elastically crimped IGBT, simulating the actual working pressure environment. An impedance analyzer is then used to perform a frequency sweep test on the elastically crimped IGBT by superimposing a small AC signal, obtaining the amplitude-frequency characteristic curve and phase-frequency characteristic curve at both ends of the elastically crimped IGBT. The total impedance curve consists of the amplitude-frequency characteristic curve (actual impedance amplitude changes with frequency) and the phase-frequency characteristic curve (actual impedance phase angle changes with frequency), obtained through frequency sweep testing. The frequency range of the impedance analyzer can be set to 100Hz-30MHz, the DC bias voltage to 10V, and the amplitude of the AC signal to 200mV, thereby ensuring that the applied AC signal meets the requirements of the small-signal test.

[0095] Step 3: Using the unknown parameters in the total impedance model as the parameters of the elastically pressed IGBT to be obtained, and using a genetic algorithm based on the total impedance model and the total impedance curve, the values ​​of the elastically pressed IGBT parameters are obtained.

[0096] It should be noted that, because the mutation factor and crossover probability of traditional genetic algorithms are usually fixed or simply random, they are prone to getting trapped in local optima and have slow convergence speeds. Therefore, in some embodiments, see... Figure 4 A genetic algorithm is used to obtain the values ​​of the elastically press-fit IGBT parameters, including:

[0097] 1) Initialize the population; the population includes multiple individuals, each individual is a set of elastic compression IGBT parameters whose values ​​are to be obtained. The initialization is to assign values ​​to each elastic compression IGBT parameter according to the individual number and the elastic compression IGBT parameter number in the individual.

[0098] It should be noted that, as can be seen from the above total impedance model, the parameters of the elastically press-fit IGBT include R1~R4, L1, C1~C3, a total of 8 parameters, and each individual contains all 8 parameters. The number of individuals, i.e., the population size N, is... pThe population size cannot be too large or too small. Too small a population will reduce the accuracy of parameter value acquisition, while too large a population will consume excessive computing resources and time. Therefore, the population size needs to be determined based on the complexity of the problem and the available computing resources. In this embodiment, considering multiple factors such as data volume and processing accuracy, it can be set to 400.

[0099] Before assigning values ​​to parameters, a reasonable range of values ​​needs to be determined for each parameter. The assigned value should not exceed the corresponding range. Specifically, a value can be randomly selected from this range for parameter assignment. This can be expressed by the formula:

[0100] ;

[0101] In the formula, Let r be the parameter of the p-th elastic compression IGBT of the q-th individual during initialization, where r is a random number calculated based on a string hash value, with a value range of [0,1]. These are the upper and lower bounds of the parameters for the p-th elastically pressed IGBT, respectively.

[0102] In some embodiments, the assignment is associated with the individual's ID and the ID of the elastic IGBT parameter within the individual. Specifically, the individual's ID and the ID of the elastic IGBT parameter within the individual can be combined into a string "ind_q_param_p". This string is then input into the hash function SHA-256 (Secure Hash Algorithm 256-bit) to obtain a hash value. This hash value is used to calculate r, which can be represented as follows:

[0103] ;

[0104] In the formula, H is the hash value of the string. The string is a string composed of the number of the q-th individual and the number of the p-th elastic compression IGBT parameter in the q-th individual. The hash function maps the individual and parameter numbers to random numbers, which ensures the repeatability of the initialization (facilitating debugging and verification) while maintaining randomness and avoiding initial population bias.

[0105] Therefore, the above parameter assignment formula can be expressed as:

[0106] .

[0107] 2) Determine the mutation factor of the current population based on the current iteration stage and the mutation factor of the previous generation. Specifically, this can be determined based on the current iteration stage, the mutation factor of the previous generation, the fitness of the best individual (i.e., the individual with the best fitness) of the previous generation, and the average fitness of individuals in the previous generation; or based on the current iteration stage, the mutation factor of the previous generation, and the standard deviation of fitness in the historical best individual list corresponding to the previous generation; or based on the current iteration stage and the mutation factor of the previous generation. Mutate each parameter in the individual according to the mutation factor to generate mutated individuals. In each iteration, when a new best individual is found, add the new best individual to the historical best individual list; the best individual is the one with the lowest fitness.

[0108] Traditional mutation factors are often fixed values ​​(such as 0.5) or randomly generated, which may lead to low search efficiency (too large a value results in slow convergence, while too small a value makes it easy to get trapped in local optima). The dynamic adjustment strategy here can adapt to the evolutionary stage and improve efficiency.

[0109] In some embodiments, the genetic algorithm divides the iteration process into a first iteration stage, a second iteration stage, and a third iteration stage according to the iteration time sequence, corresponding to the early iteration stage, the middle iteration stage, and the late iteration stage, respectively. The fitness of the best individual in the previous generation is defined as O. f The average fitness of individuals in the previous generation population is represented by A. f The standard deviation of fitness in the list of historical best individuals from the previous generation is denoted as σ. W The process of determining the variation factors of the contemporary population can be described as follows:

[0110] A. When the current iteration is in the early iteration stage, if the fitness of the best individual in the previous generation is O f Average fitness A of individuals from the previous generation f If the difference is less than the first threshold, then the variation factor of the current population is equal to the sum of the variation factor of the previous generation and the first step length. If O f With A f If the difference is greater than the first threshold and less than the second threshold, then the variation factor of the current population is equal to the sum of the variation factor of the previous generation and the second step size; if O f With A f If the difference is greater than the second threshold, then the variation factor of the current generation population is equal to the variation factor of the previous generation population; where the first step length is greater than the first step length; the variation factor of the first generation population is a preset value, such as 0.7.

[0111] Assuming the preset total number of iterations is G, the early iteration phase can be defined as the first iteration to G / 3 iterations (if G is not divisible by 3, then round down). The first threshold can be set to 0.2, the first step length can be set to 0.2, the second threshold can be set to 0.5, and the second step length can be set to 0.1.

[0112] When the current iteration is in the early iteration stage, if O f -A f If 0.2 < 0, then F = F0 + 0.2 to help escape local optima and expand the search range. f -A f If <0.5, then F=F0+0.1, encouraging further exploration. If O f -A f If F > 0.5, then F = F0; where F is the variation factor of the current population and F0 is the variation factor of the previous generation population.

[0113] B. When the current iteration is in the middle iteration stage, if the standard deviation σ of the fitness in the list of historical best individuals corresponding to the previous generation population is... W If the variance is less than the third threshold, then the variance factor of the current population is equal to the difference between the variance factor of the previous generation and the third step length. If σ W If the variance factor is greater than the third threshold and less than the fourth threshold, then the variance factor of the current generation population is equal to that of the previous generation population. W If the variation factor is greater than the fourth threshold, then the variation factor of the current population is equal to the sum of the variation factor of the previous generation population and the fourth step length.

[0114] Similarly, the intermediate iteration phase can be artificially defined as G / 3 iterations to 2G / 3 iterations, the third threshold can be set to 0.15, the third step size can be set to 0.1, the fourth threshold can be set to 0.3, and the fourth step size can be set to 0.05.

[0115] When the current iteration is in the middle iteration stage, if σ W If σ < 0.15, then F = F0 - 0.1 for a more refined search; 0.15 < σ W If σ < 0.3, then F = F0, if σ W If the value is greater than 0.3, then F = F0 + 0.05 to expand the search range.

[0116] C. When the current iteration is in the late iteration stage, the mutation factor of the current population is equal to the mutation factor of the previous generation population minus 0.2×(g-2G / 3); where g is the current iteration number and G is the preset total iteration number.

[0117] Similarly, the last iteration of the later iteration stage can be artificially defined as 2G / 3 times. In order to ensure convergence, the mutation factor can be gradually reduced by using the formula F=F0-0.2×(g-2G / 3) to gradually reduce the mutation amplitude and promote the convergence of the algorithm.

[0118] The mutation factor is dynamically adjusted based on the iteration stage (early, middle, and late stages) and the population status (poor fitness, standard deviation). In the early stage, exploration is encouraged; in the middle stage, exploration and utilization are balanced; and in the late stage, convergence is promoted, thereby improving the global search capability and convergence accuracy.

[0119] Given a known variable factor, individual variation can be calculated based on the following formula:

[0120] ;

[0121] In the formula, Let q be the individual after the mutation operation in the g-th iteration. Let be the a-th, b-th, and c-th individuals randomly selected from the current population, where a≠b≠c≠q. In the formula, F controls the degree of influence of the difference on the parameters of the new individuals, and 0.7 can be set as the initial mutation factor.

[0122] 3) Based on the diversity index of the current population, determine the crossover probability of the current population, and perform crossover operation on each parameter of the mutated individuals according to the crossover probability to generate crossover individuals;

[0123] It should be noted that the diversity index of a population is used to quantitatively characterize the degree of difference among individuals within a population, reflecting the exploration range of the algorithm in the parameter space. In some embodiments, the diversity index of the current population can be measured by the mean Euclidean distance between individuals, which can be expressed by the formula:

[0124] ;

[0125] In the formula, For the q1th individual T q1 With the q2th individual T q2 The Euclidean distance between them, D is the number of elastically pressed IGBT parameters in an individual, and T p,q1 For the q1th individual T q1 The parameter of the p-th elastic compression IGBT, T p,q2 For the q2th individual T q2 The parameters of the p-th elastic compression IGBT.

[0126] The diversity index of contemporary populations can be expressed as:

[0127] ;

[0128] Where S is the diversity index of the current population, and N p is the total number of individuals in the population (i.e., the population size).

[0129] It should be noted that, in order to improve the accuracy of obtaining parameters, in some embodiments, the crossover probability will be dynamically adjusted according to the evolutionary state of the population, specifically as follows:

[0130] If the diversity index of the current population is greater than the high diversity threshold, the crossover probability of the current population is equal to the product of the initial crossover probability and the fifth step size; if the diversity index of the current population is less than the low diversity threshold, the crossover probability of the current population is equal to the product of the initial crossover probability and the sixth step size; where the sixth step size is greater than the fifth step size; if the diversity index of the current population is less than the high diversity threshold and greater than the low diversity threshold, the crossover probability of the current population is equal to the initial crossover probability.

[0131] It should be noted that the initial crossover probability CR0 can be set to 0.4, which can balance introducing new information and retaining the original information, improving the accuracy of obtaining parameters. The high diversity threshold can be set to 0.8, the fifth step size can be set to 0.9, the low diversity threshold can be set to 0.2, and the sixth step size can be set to 1.1.

[0132] When S > 0.8, it indicates that the population has rich diversity. To make full use of the current diversity and accelerate convergence, the crossover probability is appropriately reduced, CR = CR0 × 0.9; when S < 0.2, it indicates that the population has insufficient diversity and may fall into local optimum. At this time, the crossover probability needs to be increased, CR = CR0 × 1.1; when 0.2 < S < 0.8, CR = CR0; where CR is the crossover probability of the current population.

[0133] Based on the diversity index to adjust the crossover probability, when the diversity is high, reduce the crossover probability to retain good individuals, and when the diversity is low, increase the crossover probability to introduce new genes, so as to prevent premature convergence and improve the search efficiency.

[0134] In the crossover operation, it is determined whether to replace it with the corresponding value of the mutation vector or retain the parameters in the original individual through the crossover probability. The formula for the crossover operation can be expressed as:

[0135] ;

[0136] In the formula, is the q-th individual after the crossover operation in the g-th iteration, rand(0,1) represents taking a random number between 0 and 1, and p rand is a random integer in [1, D].[[]END]

[0137] 4) Calculate the fitness of the crossover individual and the fitness of the corresponding original individual based on the total impedance model and total impedance curve; where the original individual is the individual between the mutation operation and the crossover operation.

[0138] The specific process for calculating fitness is as follows: Substitute the cross individual into the total impedance model to obtain the theoretical total impedance, and calculate the fitness of the cross individual based on the theoretical total impedance and the actual total impedance in the total impedance curve; substitute the original individual into the total impedance model to obtain the theoretical total impedance, and calculate the fitness of the original individual based on the theoretical total impedance and the actual total impedance in the total impedance curve.

[0139] It should be noted that the root mean square error (RMSE) can be used as the fitness metric here, and can be expressed by the formula:

[0140] ;

[0141] In the formula, These are the actual total impedance (covering the 100Hz–30MHz frequency band) and the theoretical total impedance, respectively, where i represents the i-th frequency point in the frequency scan test, and ω... i Let be the angular frequency corresponding to the i-th frequency point.

[0142] 5) Based on the fitness of the crossover individuals and the fitness of the corresponding original individuals, select individuals from the crossover individuals and the original individuals to be used as the next generation population. If the best individual (i.e. the one with the smallest fitness value) in the next generation population meets the preset requirements or the number of iterations has reached the preset total number of iterations, then the value corresponding to the best individual in the next generation population is used as the value of the elastic compression IGBT parameter; otherwise, the next generation population is used as the current population, and go to 2).

[0143] It should be noted that the selection of individuals from the crossover individuals and the original individuals to serve as the next generation population can be done as follows:

[0144] Select a predetermined number of minimum fitness values ​​from the fitness values ​​of the crossover individuals and the fitness values ​​of the corresponding original individuals. The selected minimum fitness values ​​are taken as the excellent fitness values. The original individuals corresponding to the excellent fitness values ​​(called "elite individuals") are taken as the individuals of the next generation of the population.

[0145] The selection strategy for individuals with non-superior fitness is as follows:

[0146] ;

[0147] In the formula, for The corresponding original individual, They are respectively and fitness, i.e. and The corresponding root mean square error, It represents the q-th individual in the (g+1)-th iteration.

[0148] It should be noted that the preset requirement can be set to the fitness (i.e. root mean square error) change rate of the optimal individual being less than 1% for 10 consecutive generations. The condition of 10 consecutive generations can avoid random factors and ensure the accuracy and stability of parameter value acquisition. The parameter value corresponding to the optimal individual is the value of the elastic compression IGBT parameter.

[0149] The aforementioned genetic algorithm, by dynamically adjusting the mutation factor (based on fitness, average fitness, and the standard deviation of the best historical individual) and the crossover probability (based on diversity indicators), can balance global and local search, improve convergence speed and accuracy, avoid premature convergence, and enhance the accuracy and robustness of parameter acquisition. The aforementioned genetic algorithm can quickly locate the optimal solution in a complex parameter space, significantly improving the accuracy and efficiency of parameter value acquisition.

[0150] To verify the above method, the results obtained were substituted into the total impedance model, and the theoretical results were compared with the measured results. (See [reference needed]). Figure 5 and Figure 6 As can be seen from the figure, within the range of 100Hz to 300Hz, the fitting theoretical results are very close to the measured results, indicating that the above method is effective and reliable.

[0151] The above method comprehensively considers the physical structure and physical characteristics of the flexible-pressed IGBT in the off state, and constructs a total impedance model of the flexible-pressed IGBT. Compared with the traditional model, it can more accurately simulate the electrical characteristics of the flexible-pressed IGBT at different frequencies. Based on the total impedance model and the measured total impedance curve, parameter values ​​that fit the total impedance curve can be obtained. These parameter values ​​are very close to the actual situation and can provide strong support for analyzing the performance of the flexible-pressed IGBT in complex circuits.

[0152] See Figure 7 , Figure 7 This is a block diagram of a device for obtaining parameter values ​​of a resiliently pressed IGBT according to an embodiment of this application. This device is a virtual device that can be loaded and executed by a computer device, which may include the aforementioned parameter value acquisition device. Figure 7 The apparatus may include a model building module, a curve acquisition module, and a parameter value acquisition module, which, when executing the above-mentioned parameter value acquisition method, can:

[0153] The model building module constructs a total impedance model for the elastically press-fit IGBT based on its physical structure and physical characteristics in the off state.

[0154] In the model building module, the total impedance model of the elastically pressed IGBT can be expressed as:

[0155] ;

[0156] In the formula, Z(jω) is the total impedance of the flexible IGBT, ω is the angular frequency, R1 and L1 are the stray resistance and stray inductance of the lead wire, R2 is the equivalent resistance of the internal buffer layer and drift region of the flexible IGBT, C2 is the equivalent capacitance of the anti-parallel diode and parasitic capacitance, R3 and C3 are the equivalent resistance and equivalent capacitance of the PN junction dielectric loss, C1 is the equivalent capacitance formed by the PN junction in the off state of the flexible IGBT, R4 is the equivalent resistance of the dielectric loss of the anti-parallel diode, C1, R2 and R3 form the broadband model of the flexible IGBT, and C2 and R4 form the broadband equivalent model of the anti-parallel diode.

[0157] The curve acquisition module acquires the total impedance curve of the elastically pressurized IGBT under actual working pressure conditions.

[0158] The parameter value acquisition module takes the unknown parameters in the total impedance model as the parameters of the elastically pressed IGBT to be acquired, and uses a genetic algorithm to obtain the values ​​of the elastically pressed IGBT parameters based on the total impedance model and the total impedance curve.

[0159] The parameter value acquisition module may include:

[0160] The initialization module initializes the population; the population includes multiple individuals, each individual is a set of elastic compression IGBT parameters whose values ​​are to be acquired. The initialization is performed by assigning values ​​to each elastic compression IGBT parameter according to the individual's number and the number of the elastic compression IGBT parameter in the individual.

[0161] In the initialization module, the initialization formula can be expressed as:

[0162] ;

[0163] In the formula, The parameter is the p-th elastic compression IGBT parameter for the q-th individual, r is a random number calculated based on a string hash value, and the string is a string constructed from the ID of the q-th individual and the ID of the p-th elastic compression IGBT parameter in the q-th individual. These are the upper and lower bounds of the parameters for the p-th elastically pressed IGBT, respectively.

[0164] The mutated individual module determines the mutation factor of the current population based on the current iteration stage, the mutation factor of the previous generation, the fitness of the best individual in the previous generation, and the average fitness of individuals in the previous generation; or based on the current iteration stage, the mutation factor of the previous generation, and the standard deviation of fitness in the historical best individual list corresponding to the previous generation; or based on the current iteration stage and the mutation factor of the previous generation. It then performs mutation operations on each parameter of the individual based on the mutation factor to generate mutated individuals. In each iteration, when a new best individual is found, it is added to the historical best individual list; the best individual is the one with the lowest fitness.

[0165] It should be noted that in genetic algorithms, the iteration process is divided into the first iteration stage, the second iteration stage, and the third iteration stage according to the time sequence of the iterations.

[0166] In the mutated individual module, the mutation factor of the current population is determined based on the current iteration stage, the mutation factor of the previous generation, the fitness of the best individual in the previous generation, and the average fitness of individuals in the previous generation; or based on the current iteration stage, the mutation factor of the previous generation, and the standard deviation of fitness in the historical best individual list corresponding to the previous generation; or based on the current iteration stage and the mutation factor of the previous generation. This can include:

[0167] When the current iteration is in the first iteration phase, if the fitness of the best individual in the previous generation is O... f Average fitness A of individuals from the previous generation f If the difference is less than the first threshold, then the variation factor of the current population is equal to the sum of the variation factor of the previous generation and the first step length. If O f With A f If the difference is greater than the first threshold and less than the second threshold, then the variation factor of the current population is equal to the sum of the variation factor of the previous generation and the second step size; if O f With A f If the difference is greater than the second threshold, then the variation factor of the current generation population is equal to the variation factor of the previous generation population; where the first step length is greater than the second step length; the variation factor of the first generation population is a preset value.

[0168] When the current iteration is in the second iteration phase, if the standard deviation σ of the fitness in the list of historical best individuals corresponding to the previous generation population is... W If the variance is less than the third threshold, then the variance factor of the current population is equal to the difference between the variance factor of the previous generation and the third step length. If σ W If the variance factor is greater than the third threshold and less than the fourth threshold, then the variance factor of the current generation population is equal to that of the previous generation population. WIf the variation factor is greater than the fourth threshold, then the variation factor of the current population is equal to the sum of the variation factor of the previous generation population and the fourth step length.

[0169] When the current iteration is in the third iteration stage, the mutation factor of the current population is equal to the mutation factor of the previous generation population minus 0.2×(g-2G / 3); where g is the current iteration number and G is the preset total iteration number.

[0170] The crossover individual module determines the crossover probability of the current population based on the diversity index of the current population, and performs crossover operations on each parameter of the mutated individuals according to the crossover probability to generate crossover individuals.

[0171] In the crossover individual module, the crossover probability of the current population is determined based on the diversity index of the current population. This can include: if the diversity index of the current population is greater than the high diversity threshold, the crossover probability of the current population is equal to the product of the initial crossover probability and the fifth step length; if the diversity index of the current population is less than the low diversity threshold, the crossover probability of the current population is equal to the product of the initial crossover probability and the sixth step length; where the sixth step length is greater than the fifth step length; if the diversity index of the current population is less than the high diversity threshold but greater than the low diversity threshold, the crossover probability of the current population is equal to the initial crossover probability.

[0172] The fitness module calculates the fitness of crossover individuals and the fitness of the corresponding original individuals based on the total impedance model and the total impedance curve; where the original individuals are those between those without mutation and crossover operations.

[0173] In the fitness module, the fitness of the cross individual and the fitness of the corresponding original individual are calculated based on the total impedance model and the total impedance curve. This can include: substituting the cross individual into the total impedance model to obtain the theoretical total impedance, and calculating the fitness of the cross individual based on the theoretical total impedance and the actual total impedance in the total impedance curve; or substituting the original individual into the total impedance model to obtain the theoretical total impedance, and calculating the fitness of the original individual based on the theoretical total impedance and the actual total impedance in the total impedance curve.

[0174] The judgment module selects individuals from the crossover individuals and the corresponding original individuals as the next generation population based on the fitness of the crossover individuals and the fitness of the original individuals. If the best individual in the next generation population meets the preset requirements or the number of iterations has reached the preset total number of iterations, the value corresponding to the best individual in the next generation population is used as the value of the elastic compression IGBT parameter. Otherwise, the next generation population is used as the current population and transferred to the mutation individual module for processing.

[0175] In the decision module, individuals are selected from the crossover individuals and the corresponding original individuals to serve as the next generation population, based on the fitness of the crossover individuals and the fitness of the original individuals. This may include:

[0176] Select a predetermined number of minimum fitness values ​​from the fitness values ​​of the crossover individuals and the fitness values ​​of the corresponding original individuals, and use the selected minimum fitness values ​​as the excellent fitness values.

[0177] The original individuals with excellent fitness will be used as individuals in the next generation of the population;

[0178] The selection strategy for individuals with non-superior fitness is as follows:

[0179] ;

[0180] In the formula, Let q be the individual after the crossover operation in the g-th iteration. for The corresponding original individual, They are respectively and fitness It represents the q-th individual in the (g+1)-th iteration.

[0181] The aforementioned device comprehensively considers the physical structure and physical characteristics of the flexible-pressed IGBT in the off state, and constructs a total impedance model for the flexible-pressed IGBT. Compared with traditional models, it can more accurately simulate the electrical characteristics of the flexible-pressed IGBT at different frequencies. Based on the total impedance model and the measured total impedance curve, parameter values ​​that fit the total impedance curve can be obtained. These parameter values ​​are very close to the actual situation and can provide strong support for analyzing the performance of the flexible-pressed IGBT in complex circuits.

[0182] This application also relates to a computer-readable storage medium that stores one or more programs, the one or more programs including instructions that, when executed by a computing device, cause the computing device to perform a method for obtaining the parameter values ​​of a resiliently pressed IGBT.

[0183] This application also relates to a computer device including one or more processors and one or more memories, wherein one or more programs are stored in the one or more memories and configured to be executed by the one or more processors, and the one or more programs include instructions for performing a method for obtaining parameter values ​​of elastically pressed IGBTs.

[0184] Those skilled in the art will understand that embodiments of the present invention can be provided as methods, systems, or computer program products. Therefore, the present invention can take the form of a completely hardware embodiment, a completely software embodiment, or an embodiment combining software and hardware aspects. Furthermore, the present invention can take the form of a computer program product embodied on one or more computer-usable storage media (including, but not limited to, disk storage, CD-ROM, optical storage, etc.) containing computer-usable program code.

[0185] This invention is described with reference to flowchart illustrations and / or block diagrams of methods, apparatus (systems), and computer program products according to embodiments of the invention. It will be understood that each block of the flowchart illustrations and / or block diagrams, and combinations of blocks in the flowchart illustrations and / or block diagrams, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, special-purpose computer, embedded processor, or other programmable data processing apparatus to produce a machine, such that the instructions, which execute via the processor of the computer or other programmable data processing apparatus, generate instructions for implementing the flowchart illustrations and / or block diagrams. Figure 1 One or more processes and / or boxes Figure 1 A device that provides the functions specified in one or more boxes.

[0186] These computer program instructions may also be stored in a computer-readable storage medium that can direct a computer or other programmable data processing device to function in a particular manner, such that the instructions stored in the computer-readable storage medium produce an article of manufacture including instruction means, which are implemented in a process Figure 1 One or more processes and / or boxes Figure 1 The function specified in one or more boxes.

[0187] These computer program instructions may also be loaded onto a computer or other programmable data processing equipment to cause a series of operational steps to be performed on the computer or other programmable equipment to produce a computer-implemented process, thereby providing instructions that execute on the computer or other programmable equipment for implementing the process. Figure 1 One or more processes and / or boxes Figure 1 The steps of the function specified in one or more boxes.

[0188] The above are merely embodiments of the present invention and are not intended to limit the present invention. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of the present invention are included within the scope of the claims of the present invention pending approval.

Claims

1. A method of acquiring a parameter value of a flexible press-pack IGBT, characterized by, include: Based on the physical structure and physical characteristics of the elastically press-fit IGBT in the off state, a total impedance model of the elastically press-fit IGBT is constructed. Obtain the total impedance curve of the elastically press-fit IGBT under actual working pressure conditions; The unknown parameters in the total impedance model are used as the parameters of the elastically pressed IGBT to be obtained. Based on the total impedance model and the total impedance curve, a genetic algorithm is used to obtain the values ​​of the elastically pressed IGBT parameters.

2. The method according to claim 1, characterized in that, The total impedance model for elastically press-fitted IGBTs is: ； In the formula, Z(jω) is the total impedance of the flexible IGBT, ω is the angular frequency, R1 and L1 are the stray resistance and stray inductance of the lead wire, R2 is the equivalent resistance of the internal buffer layer and drift region of the flexible IGBT, C2 is the equivalent capacitance of the anti-parallel diode and parasitic capacitance, R3 and C3 are the equivalent resistance and equivalent capacitance of the PN junction dielectric loss, C1 is the equivalent capacitance formed by the PN junction in the off state of the flexible IGBT, R4 is the equivalent resistance of the dielectric loss of the anti-parallel diode, C1, R2 and R3 form the broadband model of the flexible IGBT, and C2 and R4 form the broadband equivalent model of the anti-parallel diode.

3. The method according to claim 1, characterized in that, Using the unknown parameters in the total impedance model as the parameters of the elastically pressed IGBT to be obtained, and employing a genetic algorithm based on the total impedance model and total impedance curve, the values ​​of the elastically pressed IGBT parameters are obtained, including: 1) Initialize the population; the population includes multiple individuals, each individual is a set of elastic compression IGBT parameters whose values ​​are to be acquired. The initialization is to assign values ​​to each elastic compression IGBT parameter according to the individual number and the elastic compression IGBT parameter number in the individual. 2) Based on the current iteration stage and the mutation factor of the previous generation, determine the mutation factor of the current generation. Perform mutation operation on each parameter of the individual according to the mutation factor to generate mutated individuals. In each iteration, when a new optimal individual is found, add the new optimal individual to the list of historical optimal individuals. The optimal individual is the individual with the lowest fitness. 3) Based on the diversity index of the current population, determine the crossover probability of the current population, and perform crossover operation on each parameter of the mutated individuals according to the crossover probability to generate crossover individuals; 4) Calculate the fitness of the crossover individuals and the fitness of the corresponding original individuals based on the total impedance model and total impedance curve; where the original individuals are those between those without mutation and crossover operations. 5) Based on the fitness of the crossover individuals and the fitness of the corresponding original individuals, select individuals from the crossover individuals and the original individuals to be used as the next generation population. If the best individual in the next generation population meets the preset requirements or the number of iterations has reached the preset total number of iterations, then the value corresponding to the best individual in the next generation population is used as the value of the elastic compression IGBT parameter; otherwise, the next generation population is used as the current population, and go to 2).

4. The method according to claim 3, characterized in that, The initialization formula is: ； In the formula, The parameter is the p-th elastic compression IGBT parameter for the q-th individual, r is a random number calculated based on a string hash value, and the string is a string constructed from the ID of the q-th individual and the ID of the p-th elastic compression IGBT parameter in the q-th individual. These are the upper and lower bounds of the parameters for the p-th elastically pressed IGBT, respectively.

5. The method according to claim 3, characterized in that, In genetic algorithms, the iterative process is divided into three stages according to the time sequence of the iterations: the first iteration stage, the second iteration stage, and the third iteration stage. Based on the current iteration stage and the variation factors of the previous generation, determine the variation factors of the current population, including: When the current iteration is in the first iteration stage, if the difference between the fitness O f of the optimal individual of the last generation population and the average fitness A f of the individuals of the last generation population is less than a first threshold value, the mutation factor of the current generation population is equal to the sum of the mutation factor of the last generation population and a first step length; if the difference between O f and A f is greater than the first threshold value and less than a second threshold value, the mutation factor of the current generation population is equal to the sum of the mutation factor of the last generation population and a second step length; if the difference between O f and A f is greater than the second threshold value, the mutation factor of the current generation population is equal to the mutation factor of the last generation population; wherein the first step length is greater than the second step length; and the mutation factor of the first generation population is a preset value. When the current iteration is in the second iteration phase, if the standard deviation σ of the fitness in the list of historical best individuals corresponding to the previous generation population is... W If the variance is less than the third threshold, then the variance factor of the current population is equal to the difference between the variance factor of the previous generation and the third step length. If σ W If the variance factor is greater than the third threshold and less than the fourth threshold, then the variance factor of the current generation population is equal to that of the previous generation population. W If the variation factor is greater than the fourth threshold, then the variation factor of the current population is equal to the sum of the variation factor of the previous generation population and the fourth step length. When the current iteration is in the third iteration stage, the mutation factor of the current population is equal to the mutation factor of the previous generation population minus 0.2×(g-2G / 3); where g is the current iteration number and G is the preset total iteration number.

6. The method according to claim 3, characterized in that, Based on the diversity indicators of the current population, determine the crossover probability of the current population, including: If the diversity index of the current population is greater than the high diversity threshold, the crossover probability of the current population is equal to the product of the initial crossover probability and the fifth step length. If the diversity index of the current population is less than the low diversity threshold, the crossover probability of the current population is equal to the product of the initial crossover probability and the sixth step length; where the sixth step length is greater than the fifth step length. If the diversity index of the current population is less than the high diversity threshold but greater than the low diversity threshold, the crossover probability of the current population is equal to the initial crossover probability.

7. The method according to claim 3, characterized in that, Based on the total impedance model and total impedance curve, calculate the fitness of the crossover individuals and the fitness of the corresponding original individuals, including: Substitute the cross individuals into the total impedance model to obtain the theoretical total impedance. Calculate the fitness of the cross individuals based on the theoretical total impedance and the actual total impedance in the total impedance curve. Substitute the original individual into the total impedance model to obtain the theoretical total impedance. Based on the theoretical total impedance and the actual total impedance in the total impedance curve, calculate the fitness of the original individual.

8. The method according to claim 3, characterized in that, Based on the fitness of the crossover individuals and the fitness of the corresponding original individuals, individuals are selected from the crossover individuals and the original individuals to serve as the next generation population, including: Select a predetermined number of minimum fitness values ​​from the fitness values ​​of the crossover individuals and the fitness values ​​of the corresponding original individuals, and use the selected minimum fitness values ​​as the excellent fitness values. The original individuals with excellent fitness will be used as individuals in the next generation of the population; The selection strategy for individuals with non-superior fitness is as follows: ； In the formula, Let q be the individual after the crossover operation in the g-th iteration. for The corresponding original individual, They are respectively and Adaptability, It represents the q-th individual in the (g+1)-th iteration.

9. A device for obtaining parameter values ​​of elastically pressed IGBTs, characterized in that, include: The model building module constructs a total impedance model for the elastically press-fit IGBT based on its physical structure and physical characteristics under the cut-off state. The curve acquisition module acquires the total impedance curve of the elastically pressed IGBT under actual working pressure conditions. The parameter value acquisition module takes the unknown parameters in the total impedance model as the parameters of the elastically pressed IGBT to be acquired, and uses a genetic algorithm to obtain the values ​​of the elastically pressed IGBT parameters based on the total impedance model and the total impedance curve.

10. A computer-readable storage medium, characterized in that, The computer-readable storage medium stores one or more programs, the one or more programs including instructions that, when executed by a computing device, cause the computing device to perform the method of any one of claims 1 to 8.

11. A computer device, characterized in that, include: One or more processors and one or more memories, one or more programs stored in one or more memories and configured to be executed by one or more processors, the one or more programs including instructions for performing the method of any one of claims 1 to 8.

Images