Transformer multi-physics field direct coupling order reduction processing method and related device

Abstract

The embodiment of the invention discloses an order reduction processing method for direct coupling of multiple physical fields of a transformer and a related device, and the method comprises the steps: obtaining and utilizing geometric parameters, material attributes and operation conditions of the transformer, and building a simulation geometric model of the transformer; based on heating, fluid flowing and fluid-solid heat transfer processes during operation of the transformer, constructing an initial coupling model of direct coupling of a heat-fluid-solid multi-physics field of the transformer; and establishing a reduced-order model of a heat-fluid-solid multi-physics field by using a random forest algorithm and the initial coupling model, and performing multi-physics field operation simulation of the transformer according to the simulation geometric model and the reduced-order model to obtain an operation simulation result reflecting the operation condition of the transformer in the multi-physics field. By reducing the order of the initial coupling model through the random forest algorithm, the complexity of the model can be simplified, the degree of freedom in the simulation model can be reduced, the calculation efficiency of simulation can be improved, the requirement for calculation resources can be reduced, and meanwhile the reliability of a result is kept.

Description

Technical Field

[0001] The present invention relates to the technical field of transformer multi-physical field coupling calculation, and in particular to a transformer multi-physical field direct coupling reduction processing method and related devices. Background Art

[0002] In the operation and selection of transformers, multi-physics field simulation is involved for margin verification. During the operation of the transformer, heat is generated due to copper loss, iron loss, etc., and is transferred through the medium (transformer oil, insulation material, casing, air, etc.). The flow of transformer oil is affected by the internal structure of the transformer (such as windings, iron core, etc.). At the same time, the flow of oil will also affect the heat distribution of the entire system, thereby affecting the efficiency and safety of the transformer. In addition, the propagation of heat will cause thermal expansion of solid components, thereby affecting the mechanical structure and electrical performance of the transformer.

[0003] The above physical processes exist simultaneously and influence each other. When directly coupling and calculating multiple physical fields, simulation usually consumes huge computing resources, especially when it involves refined grids and complex boundary conditions. Due to the mutual coupling of processes such as heat conduction, fluid flow and solid deformation, traditional numerical simulation methods often have difficulty in achieving a balance between computing efficiency and accuracy when dealing with such problems.

[0004] Therefore, there is an urgent need for a method that can improve computational efficiency and accuracy when directly coupling and calculating multiple physical fields. Summary of the invention

[0005] The main purpose of the present invention is to provide a method for reducing the order of direct coupling of transformer multi-physical fields and related devices, which can solve the problem in the prior art of lacking means to improve calculation efficiency and accuracy when directly coupling and calculating multi-physical fields.

[0006] To achieve the above object, the present invention provides a first aspect of a transformer multi-physics field direct coupling reduction processing method, the method comprising:

[0007] Acquire and use geometric parameters, material properties and operating conditions of the transformer to establish a simulation geometric model of the transformer;

[0008] Based on the heat generation, fluid flow and fluid-solid heat transfer process of the transformer during operation, an initial coupling model of direct coupling of thermal-fluid-solid multi-physics fields of the transformer is constructed;

[0009] Using random forest and the initial coupling model, a reduced-order model of the thermal-fluid-solid multi-physics field is established;

[0010] A multi-physical field operation simulation of the transformer is performed according to the simulation geometric model and the reduced-order model to obtain an operation simulation result, and the operation simulation result is used to reflect the operating condition of the transformer in the multi-physical field.

[0011] In a feasible implementation, the use of random forest and the initial coupling model to establish the reduced-order model of the heat-fluid-solid multi-physics field includes:

[0012] Calculate the importance of all candidate data features by random forest, wherein the candidate data features are used to indicate data features that affect the running simulation results when the multi-physics field running simulation of the transformer is performed;

[0013] Calculate the Pearson correlation coefficient between each pair of the candidate data features;

[0014] Determine whether there is a strong correlation between the candidate data features based on the Pearson correlation coefficient;

[0015] If there is a strong correlation between the candidate data features, then the candidate data features are determined to be redundant features.

[0016] The target data features with low importance among the redundant features are removed from the initial coupling model to obtain a reduced-order model.

[0017] In a feasible implementation, the use of random forest and the initial coupling model to establish the reduced-order model of the heat-fluid-solid multi-physics field includes:

[0018] Calculate the importance of all candidate data features by random forest, wherein the candidate data features are used to indicate data features that affect the running simulation results when the multi-physics field running simulation of the transformer is performed;

[0019] Calculate the Pearson correlation coefficient between each pair of the candidate data features;

[0020] Determine whether there is a strong correlation between the candidate data features based on the Pearson correlation coefficient;

[0021] If there is a strong correlation between the candidate data features, then the candidate data features are determined to be redundant features.

[0022] The target data features with low importance among the redundant features are removed from the initial coupling model to obtain a reduced-order model.

[0023] In a feasible implementation, the random forest includes multiple decision trees, and the importance of all candidate data features is calculated by the random forest, including:

[0024] Determine the target contribution of each candidate data feature in each decision tree;

[0025] A comprehensive calculation is performed based on the respective target contributions to determine the total contribution of each of the candidate data features, wherein the importance includes the total contribution.

[0026] In a feasible implementation, determining the contribution of each candidate data feature in each decision tree includes:

[0027] Calculate the first Gini index of the candidate data feature at the m-th node, the second Gini index of the l-th node, and the third Gini index of the r-th node of the k-th decision tree, where the l-th node and the r-th node are branch nodes corresponding to the splitting of the m-th node;

[0028] Determine the contribution of the candidate data feature to the mth node of the kth decision tree by using the first Gini index, the second Gini index, and the third Gini index;

[0029] A comprehensive calculation is performed based on the contribution of the candidate data feature at each node of the i-th decision tree to determine the target contribution of the candidate data feature at the k-th decision tree.

[0030] In a feasible implementation, the calculating of the Pearson correlation coefficients between the candidate data features each pair includes:

[0031] Determining the standard deviation of the candidate data features and the covariance between each pair of the candidate data features;

[0032] The Pearson correlation coefficient between each pair of the candidate data features is obtained using a preset Pearson correlation coefficient algorithm, standard deviation and covariance.

[0033] In a feasible implementation, the total contribution is determined using the following mathematical expression:

[0034]

[0035] In the formula, VIM(X i ) is the candidate data feature X i Total contribution of VIM k (X i ) is the candidate data feature X i The target contribution of the kth decision tree, K is the total number of decision trees, where

[0036]

[0037] Where VIM k,m (X i ) is the candidate data feature X i The contribution of node m in the kth decision tree, N is Xi The number of splits on decision tree k, where

[0038] VIM k,m (X i )=GI k,m (X i )-[GI k,l (X i )+GI k,r (X i )];

[0039] In the formula, GI k,m (X i ) is the first Gini index, GI k,l (X i ) is the second Gini index, GI k,r (X i ) is the third Gini index.

[0040] In one feasible implementation, the Pearson correlation coefficient algorithm includes the following mathematical expression:

[0041]

[0042] Where r(X,Y) is the Pearson correlation coefficient between candidate data feature X and candidate data feature Y, Cov(X,Y) represents the covariance between candidate data feature X and candidate data feature Y, Var[X] and Var[Y] represent the standard deviations of candidate data feature X and candidate data feature Y, respectively.

[0043] To achieve the above object, the second aspect of the present invention provides a transformer multi-physical field direct coupling reduction order processing device, the device comprising:

[0044] The first modeling module is used to obtain and use the geometric parameters, material properties and operating conditions of the transformer to establish a simulation geometric model of the transformer;

[0045] The second modeling module is used to construct an initial coupling model of direct coupling of thermal-fluid-solid multi-physics fields of the transformer based on the heat generation, fluid flow and fluid-solid heat transfer process during the operation of the transformer;

[0046] A reduced-order processing module: used to establish a reduced-order model of the thermal-fluid-solid multi-physics field by using random forest and the initial coupling model;

[0047] Operation simulation module: used to perform multi-physical field operation simulation of the transformer according to the simulation geometric model and the reduced-order model to obtain operation simulation results, and the operation simulation results are used to reflect the operating conditions of the transformer in the multi-physical field.

[0048] To achieve the above-mentioned purpose, the third aspect of the present invention provides a computer-readable storage medium storing a computer program. When the computer program is executed by a processor, the processor executes the steps shown in the first aspect and any feasible implementation method.

[0049] To achieve the above-mentioned purpose, the fourth aspect of the present invention provides a computer device, including a memory and a processor, wherein the memory stores a computer program, and when the computer program is executed by the processor, the processor executes the following steps as shown in the first aspect and any feasible implementation method.

[0050] The embodiments of the present invention have the following beneficial effects:

[0051] The present invention provides a method for reducing the order of direct coupling of transformer multi-physics fields, the method comprising: obtaining and utilizing geometric parameters, material properties and operating conditions of the transformer to establish a simulation geometric model of the transformer; constructing an initial coupling model of direct coupling of heat-fluid-solid multi-physics fields of the transformer based on the heating, fluid flow and fluid-solid heat transfer process during transformer operation; utilizing a random forest algorithm and an initial coupling model to establish a reduced-order model of the heat-fluid-solid multi-physics field, performing multi-physics field operation simulation of the transformer according to the simulation geometric model and the reduced-order model, and obtaining an operation simulation result, which is used to reflect the operating conditions of the transformer in the multi-physics field. By utilizing a random forest algorithm and an initial coupling model to establish a reduced-order model of the heat-fluid-solid multi-physics field, the reduction of the order of the initial coupling model is achieved, and the reduction can simplify the complexity of the model, thereby reducing the degree of freedom in the simulation model. When simulating the operation of the multi-physics field, the calculation efficiency of the simulation can be greatly improved, the demand for calculation resources can be reduced, and the reliability of the operation simulation results can be maintained. BRIEF DESCRIPTION OF THE DRAWINGS

[0052] In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the drawings required for use in the embodiments or the description of the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying creative work.

[0053] in:

[0054] Figure 1 It is a flow chart of a method for reducing the order of direct coupling of transformer multi-physical fields in an embodiment of the present invention;

[0055] Figure 2 Another flow chart of a method for reducing the order of direct coupling of transformer multi-physics fields in an embodiment of the present invention;

[0056] Figure 3It is a structural block diagram of a transformer multi-physical field direct coupling reduction order processing device in an embodiment of the present invention;

[0057] Figure 4 4 is a structural block diagram of a computer device in an embodiment of the present invention. DETAILED DESCRIPTION

[0058] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0059] In order to solve the problem that traditional numerical simulation methods often have difficulty in striking a balance between computational efficiency and accuracy when dealing with such problems due to the mutual coupling of processes such as heat conduction, fluid flow, and solid deformation, a reduced-order calculation method can be used. The reduced-order method reduces the number of variables required for calculation by simplifying the complexity of the model, while trying to maintain the key features of the original physical process. Random forest is a classic ensemble learning method that achieves classification and regression functions by training multiple decision trees, and can calculate the importance of each feature in the data set during the training process. The main variables that affect the behavior of the system are extracted through certain rules, thereby reducing the degrees of freedom in the simulation model, greatly improving the computational efficiency of the simulation, reducing the demand for computing resources, and maintaining the reliability of the results.

[0060] By effectively applying the reduced-order calculation method of direct coupling of multiple physical fields, the performance of the transformer under actual working conditions can be predicted more accurately and efficiently. This has important scientific and engineering application value for transformer design optimization, fault analysis, life prediction, and digital twin model construction. For details, please refer to the following content.

[0061] See also Figure 1 , Figure 1 The flowchart of a method for reducing the order of direct coupling of transformer multi-physics fields in an embodiment of the present invention is shown in FIG. The method can be applied to both a terminal and a server. The terminal can be a desktop terminal or a mobile terminal. The mobile terminal can be at least one of a mobile phone, a tablet computer, and a laptop computer. The server can be implemented by an independent server or a server cluster composed of multiple servers. This embodiment is described by applying to a terminal as an example. Figure 1 The method shown comprises the following steps:

[0062] 101. Acquire and use geometric parameters, material properties and operating conditions of the transformer to establish a simulation geometric model of the transformer;

[0063] 102. Based on the heat generation, fluid flow, and fluid-solid heat transfer process during the operation of the transformer, an initial coupling model of direct coupling of thermal-fluid-solid multi-physics fields of the transformer is constructed;

[0064] First of all, it is necessary to accurately capture the physical characteristics and operating environment of the transformer in order to accurately restore the real environment and perform modeling, including but not limited to the geometric parameters, material properties and operating conditions of the transformer, etc. For example, the geometric dimensions of the transformer components, material properties (such as density, thermal conductivity, specific heat capacity, etc.) and operating conditions (such as load current, ambient temperature and other environmental factors) should be recorded in detail. These predetermined actual transformer data will be used as the basis for establishing a multi-physics field model to simulate the thermal behavior, fluid dynamics and structural response of the transformer in actual operation.

[0065] First, the geometric parameters, material properties and operating conditions of the transformer are obtained and used to establish a simulation geometric model of the transformer. Secondly, the heating, fluid flow, fluid-solid heat transfer and other processes are considered. Based on the heating, fluid flow and fluid-solid heat transfer processes during the operation of the transformer, an initial coupling model of the direct coupling of the thermal-fluid-solid multi-physics fields of the transformer is constructed. This initial coupling model is used as the thermal-fluid-solid multi-physics field coupling model of the transformer.

[0066] An exemplary thermal-fluid-solid multiphysics coupling model includes:

[0067] Taking into account the viscosity and pressure of the fluid, the Navier-Stokes equations are used to describe the fluid motion (1):

[0068]

[0069] Where ρ is the density, v is the velocity field, p is the pressure, μ is the dynamic viscosity, and f is the body force.

[0070] Considering conduction, convection and radiation, the expression describing the transfer of energy in fluids and solids (2) is:

[0071]

[0072] Where T is the temperature, c p is the specific heat capacity, k is the thermal conductivity, and Q is the heat source term. The heat source term Q comes from the internal heat source generated by chemical reactions or interactions with radiation, and the radiation heat transfer is described by the Stefan-Boltzmann law, as shown in formula (3):

[0073] q Stefan-Boltzmann =σεT 4 (3)

[0074] where σ is the Stefan-Boltzmann constant.

[0075] Furthermore, the heat generation of the winding is described using Joule's law, as shown in equation (4):

[0076] q Joule = J·E (4)

[0077] At the same time, the spread of heat will cause the thermal expansion of solid parts, thus affecting the mechanical structure and electrical performance of the transformer. Considering the influence of fluid-solid heat transfer on solids, taking into account the thermal stress and thermal expansion effects, the thermal-solid coupling equation is constructed, as shown in Equation (5).

[0078]

[0079] Where a is the thermal expansion coefficient, T0 is the reference temperature, and e T is thermal strain.

[0080] Considering the normal stress (generated by fluid pressure) and shear stress (represented by stress tensor) exerted by the fluid on the solid surface, the dot product of this combined tensor and the normal vector of the solid surface is used to obtain the total force acting on the solid surface, thereby establishing the fluid-solid coupling equation, as shown in Equation (6):

[0081] F A =[-pI+K]·n (6)

[0082] Among them, F A is the force of the fluid on the solid surface, p is the fluid pressure, K is the stress tensor representing the shear stress and normal stress inside the fluid, and n is the unit normal vector of the solid surface.

[0083] 103. Using random forest and the initial coupling model, a reduced-order model of the heat-fluid-solid multi-physics field is established;

[0084] Furthermore, the reduced-order model is used to perform reduced-order calculations in the process of directly coupling and calculating multi-physics fields, and random forests can be used for order reduction processing. For example, the importance of all candidate data features can be calculated by random forests, and the order can be reduced by the importance of the candidate data features, such as retaining data features with high importance. Among them, the candidate data features are data features obtained from the measurement data of the transformer collected in advance, and the candidate data features are used to indicate the data features that affect the operation simulation results when the multi-physics field operation simulation of the transformer is performed, and the random forest is trained using the above-mentioned measurement data as training sample data.

[0085] 104. Perform multi-physical field operation simulation of the transformer according to the simulation geometric model and the reduced-order model to obtain an operation simulation result, wherein the operation simulation result is used to reflect the operating condition of the transformer in the multi-physical field.

[0086] Furthermore, simulation can be performed based on the reduced-order model and the simulation geometry model, and multi-physical field operation simulation of the transformer can be performed based on the simulation geometry model and the reduced-order model to quickly obtain operation simulation results that reflect the operating conditions of the transformer in multi-physical fields.

[0087] The present invention provides a method for reducing the order of direct coupling of transformer multi-physics fields, the method comprising: obtaining and utilizing geometric parameters, material properties and operating conditions of the transformer to establish a simulation geometric model of the transformer; constructing an initial coupling model of direct coupling of heat-fluid-solid multi-physics fields of the transformer based on the heating, fluid flow and fluid-solid heat transfer process during transformer operation; utilizing a random forest algorithm and an initial coupling model to establish a reduced-order model of the heat-fluid-solid multi-physics field, performing multi-physics field operation simulation of the transformer according to the simulation geometric model and the reduced-order model, and obtaining an operation simulation result, which is used to reflect the operating conditions of the transformer in the multi-physics field. By utilizing a random forest algorithm and an initial coupling model to establish a reduced-order model of the heat-fluid-solid multi-physics field, the reduction of the order of the initial coupling model is achieved, and the reduction can simplify the complexity of the model, thereby reducing the degree of freedom in the simulation model. When simulating the operation of the multi-physics field, the calculation efficiency of the simulation can be greatly improved, the demand for calculation resources can be reduced, and the reliability of the operation simulation results can be maintained.

[0088] See also Figure 2 , Figure 2 FIG. 1 is another flow chart of a method for reducing the order of direct coupling of multi-physical fields of a transformer according to an embodiment of the present invention. Figure 2 The method shown comprises the following steps:

[0089] 201. Acquire and use geometric parameters, material properties and operating conditions of the transformer to establish a simulation geometric model of the transformer;

[0090] 202. Based on the heat generation, fluid flow, and fluid-solid heat transfer process during operation of the transformer, construct an initial coupling model of direct coupling of thermal-fluid-solid multi-physics fields of the transformer;

[0091] It should be noted that steps 201 and 202 are Figure 1 The contents of steps 101 and 102 are similar, and are not described here to avoid repetition. For details, please refer to Figure 1 The contents of steps 101 and 102 are shown

[0092] In a feasible implementation method, in order to improve computing efficiency while ensuring data accuracy and accurate simulation, a reduced-order model of the heat-fluid-solid multi-physics field is established using random forests and the initial coupling model. It can also be combined with two-stage feature reduction to combine the importance and correlation of sample data features to determine whether different features are strongly correlated. If the correlation is strong, it is determined to be a redundant feature, and the features with lower importance are removed. The reduced-order model of the heat-fluid-solid multi-physics coupling field is constructed by combining the obtained two-stage feature reduction. This can ensure the accuracy of the simulation while reducing the feature dimension. The importance of the data features can be determined by random forests, and the correlation can be determined by the Pearson correlation coefficient. Please refer to the following for details.

[0093] 203. Calculate the importance of all candidate data features by random forest, where the candidate data features are used to indicate data features that affect the running simulation results when the multi-physics field running simulation of the transformer is performed;

[0094] It is understandable that random forest is a classic ensemble learning method. It achieves classification and regression functions by training multiple decision trees, and can calculate the importance of each feature in the data set during the training process. The data set can be obtained from measurement data, and the measurement data is used as sample data to obtain the subsequent training set. Therefore, the importance of all candidate data features is calculated through random forest, and can be sorted from high to low. The implementation steps are as follows:

[0095] 1) Each time, N training samples are extracted from the training set with replacement to form a new training set;

[0096] 2) Draw M times in total and use the new training set to train M decision trees. Randomly select d features without duplication at each generated node and use these d features to divide the sample set;

[0097] 3) When performing a classification task, each decision tree will output a classification result, and the final classification result is determined by voting of all decision trees.

[0098] The contribution of each data feature is evaluated using the Gini Index. The Gini Index represents the probability that a randomly selected sample in the sample set is misclassified. i The Gini index GI of the kth decision tree at node m k,m (X i )If (7):

[0099]

[0100] Where: C is the number of categories, p m,cis the proportion of class c at node m, that is, the probability of correct classification. The smaller the Gini index, the smaller the probability that the selected sample in the set is misclassified, that is, the higher the purity of the set, and vice versa, the more impure the set.

[0101] Data feature X i The importance at (k,m) is expressed as the change in the Gini index before and after the node m branch VIM k,m (X i ), and the calculation method is as follows:

[0102] VIM k,m (X i )=GI k,m (X i )-[GI k,l (X i )+GI k,r (X i )] (8)

[0103] Where: GI k,l and GI k,r Represents the Gini index of the two new nodes l and r after branching.

[0104] Data feature X i The decision tree k is split N times, and the data feature X i Sum all the nodes that appear in the tree, then X i Contribution to this tree VIM k As shown in formula (9):

[0105]

[0106] If there are K trees in the random forest, then X i The total contribution VIM in the random forest is as follows:

[0107]

[0108] Each feature in the sample data is processed, the contribution of all features is calculated, and finally normalized, so that the contribution score VIM (X) of each feature is obtained. i ).

[0109] In a feasible implementation, the random forest includes multiple decision trees, and step 203 includes steps A01 and A02:

[0110] A01. Determine the target contribution of each candidate data feature in each decision tree;

[0111] In a feasible implementation, step A01 may include the following steps B01, B02 and B03:

[0112] B01. Calculate the first Gini index of the candidate data feature at the m-th node, the second Gini index of the l-th node, and the third Gini index of the r-th node of the k-th decision tree, where the l-th node and the r-th node are the branch nodes corresponding to the m-th node after splitting;

[0113] B02. Determine the contribution of the candidate data feature to the mth node of the kth decision tree by using the first Gini index, the second Gini index, and the third Gini index;

[0114] B03. Perform a comprehensive calculation based on the contribution of the candidate data feature at each node of the i-th decision tree to determine the target contribution of the candidate data feature at the k-th decision tree.

[0115] A02. Perform comprehensive calculation based on each target contribution to determine the total contribution of each candidate data feature, where the importance includes the total contribution.

[0116] In a feasible implementation, the total contribution is determined using the following mathematical expression:

[0117]

[0118] In the formula, VIM(X i ) is the candidate data feature X i Total contribution of VIM k (X i ) is the candidate data feature X i The target contribution of the kth decision tree, K is the total number of decision trees, where

[0119]

[0120] Where VIM k,m (X i ) is the candidate data feature X i The contribution of node m in the kth decision tree, N is X i The number of splits on decision tree k, where

[0121] VIM k,m (X i )=GI k,m (X i )-[GI k,l (X i )+GI k,r (X i )];

[0122] In the formula, GI k,m (X i) is the first Gini index, GI k,l (X i ) is the second Gini index, GI k,r (X i ) is the third Gini index.

[0123] 204. Calculate the Pearson correlation coefficient between each pair of the candidate data features;

[0124] Furthermore, the Pearson correlation coefficient between each measured data is calculated to form a Pearson correlation coefficient matrix, so as to obtain the Pearson correlation coefficient between each candidate data feature. The Pearson correlation coefficient reflects the correlation between variables X and Y. Its value is between -1 and 1. The closer it is to 1, the stronger the positive correlation between the two variables, and the closer it is to -1, the stronger the negative correlation. Usually, its absolute value is taken to measure whether there is a correlation between two variables. Assuming that X and Y are two different nodes in the distribution network, their flow data are collected, and their Pearson correlation coefficient r is obtained as formula (11):

[0125]

[0126] Where Cov(X,Y) represents the covariance of X and Y, Var[X] and Var[Y] represent the standard deviations of variables X and Y respectively. The covariance is divided by the standard deviation to normalize the result and measure the correlation between each pair of variables.

[0127] That is, the Pearson correlation coefficient between each pair of the candidate data features is calculated, including determining the standard deviation of the candidate data features and the covariance between each pair of the candidate data features; using the preset Pearson correlation coefficient algorithm, standard deviation and covariance, the Pearson correlation coefficient between each pair of the candidate data features is obtained.

[0128] It can be understood that the Pearson correlation coefficient algorithm includes the following mathematical expression:

[0129]

[0130] Where r(X,Y) is the Pearson correlation coefficient between candidate data feature X and candidate data feature Y, Cov(X,Y) represents the covariance between candidate data feature X and candidate data feature Y, Var[X] and Var[Y] represent the standard deviations of candidate data feature X and candidate data feature Y, respectively.

[0131] 205. Determine whether there is a strong correlation between any two of the candidate data features based on the Pearson correlation coefficient;

[0132] 206. If there is a strong correlation between the candidate data features, then determining that the candidate data features are redundant features;

[0133] 207. Eliminate the target data features with low importance among the redundant features from the initial coupling model to obtain a reduced-order model;

[0134] It should be noted that when judging whether different measurements are strongly correlated, if the correlation is strong, it is judged as redundant features and the features with lower importance are removed. Specifically, the correlation coefficient is first used to judge whether the features are strongly correlated. If there is a strong correlation, it means that there is redundancy, and the strongly correlated features can be removed. At this time, the importance of the features is further evaluated to remove the features with lower importance (redundant features) with strong correlation, thus ensuring data accuracy and achieving order reduction.

[0135] Through the above steps, the main features can be effectively extracted from the multivariate data, the dimension of the data set can be reduced, and the information of the original data set can be retained as much as possible, thereby obtaining the reduced-order calculation equation under the thermal-fluid-solid coupling of the transformer.

[0136] After the above steps, the transformer geometric model is established, the material parameters are imported, and the thermal-fluid-solid multi-physics field coupling equation is constructed. Then, through standardization, solving eigenvalues, eigenvectors and other calculations, the reduced-order equation is constructed to perform rapid calculations of the thermal-fluid-solid multi-physics field.

[0137] 208. Perform multi-physical field operation simulation of the transformer according to the simulation geometric model and the reduced-order model to obtain an operation simulation result, wherein the operation simulation result is used to reflect the operating condition of the transformer in the multi-physical field.

[0138] It should be noted that step 208 and Figure 1 The contents of step 104 are similar and will not be described here to avoid repetition. Figure 1 The content of step 104 is shown.

[0139] The present invention provides a method for reducing the order of direct coupling of transformer multi-physics fields, which can be regarded as a method for reducing the order of direct coupling of transformer heat-fluid-solid multi-physics fields. Obtain the physical characteristics and actual working conditions of the transformer, and establish a corresponding simulation geometric model in the simulation software in combination with the geometric dimensions and material parameters of the transformer; consider the heating, fluid flow, and fluid-solid heat transfer processes to establish a heat-fluid-solid coupling multi-physics field model; combine the importance and correlation of sample data features in combination with two-stage feature reduction, remove the features with insufficient contribution, and construct a reduced-order model of the heat-fluid-solid multi-physics coupling field in combination with the obtained two-stage feature reduction. Combine the geometric model and the physical model to perform direct coupling reduction calculation of heat-fluid-solid multi-physics fields. The present invention comprehensively considers physical processes such as heating, fluid flow, and fluid-solid heat transfer to establish a heat-fluid-solid multi-physics field direct coupling model of the transformer. The present invention establishes a heat-fluid-solid multi-physics field reduction model, uses a two-stage feature reduction method to reduce the calculation dimension, and can realize rapid calculation of multi-physics fields under heat-fluid-solid direct coupling.

[0140] See also Figure 3 , Figure 3 : is a structural block diagram of a transformer multi-physical field direct coupling reduction order processing device according to an embodiment of the present invention, such as Figure 3 The device shown comprises:

[0141] The first modeling module 301 is used to obtain and use the geometric parameters, material properties and operating conditions of the transformer to establish a simulation geometric model of the transformer;

[0142] The second modeling module 302 is used to construct an initial coupling model of direct coupling of thermal-fluid-solid multi-physics fields of the transformer based on the heat generation, fluid flow and fluid-solid heat transfer process during the operation of the transformer;

[0143] The order reduction processing module 303 is used to establish the order reduction model of the heat-fluid-solid multi-physics field by using the random forest and the initial coupling model;

[0144] Operation simulation module 304: used to perform multi-physical field operation simulation of the transformer according to the simulation geometric model and the reduced-order model to obtain operation simulation results, and the operation simulation results are used to reflect the operating conditions of the transformer in the multi-physical fields.

[0145] It should be noted that if Figure 3 The device shown is Figure 1 The contents of each step in the method shown are similar, so they are not described here to avoid repetition. For details, please refer to Figure 1 The content of each step in the method shown.

[0146] The present invention provides a transformer multi-physics field direct coupling reduction processing device, the device includes: a first modeling module: used to obtain and use the geometric parameters, material properties and operating conditions of the transformer to establish a simulation geometric model of the transformer; a second modeling module: used to construct an initial coupling model of the direct coupling of the heat-fluid-solid multi-physics field of the transformer based on the heat generation, fluid flow and fluid-solid heat transfer process during the operation of the transformer; a reduction processing module: used to use random forest and initial coupling model to establish a reduced-order model of the heat-fluid-solid multi-physics field; an operation simulation module: used to perform multi-physics field operation simulation of the transformer according to the simulation geometric model and the reduced-order model to obtain an operation simulation result, and the operation simulation result is used to reflect the operating conditions of the transformer in the multi-physics field. By using the random forest algorithm and the initial coupling model to establish a reduced-order model of the heat-fluid-solid multi-physics field, the reduction of the initial coupling model is achieved, and the reduction can simplify the complexity of the model, thereby reducing the degree of freedom in the simulation model. When simulating the operation of the multi-physics field, the calculation efficiency of the simulation can be greatly improved, the demand for computing resources can be reduced, and the reliability of the operation simulation results can be maintained.

[0147] Figure 4 FIG. 1 shows an internal structure diagram of a computer device in an embodiment. The computer device may be a terminal or a server. Figure 4 As shown, the computer device includes a processor, a memory and a network interface connected via a system bus. The memory includes a non-volatile storage medium and an internal memory. The non-volatile storage medium of the computer device stores an operating system and may also store a computer program. When the computer program is executed by the processor, the processor can implement the above method. The internal memory may also store a computer program. When the computer program is executed by the processor, the processor can implement the above method. Those skilled in the art can understand that Figure 4 The structure shown in the figure is only a block diagram of a part of the structure related to the solution of the present application, and does not constitute a limitation on the computer device to which the solution of the present application is applied. The specific computer device may include more or fewer components than those shown in the figure, or combine certain components, or have a different arrangement of components.

[0148] In one embodiment, a computer device is provided, comprising a memory and a processor, wherein the memory stores a computer program, and when the computer program is executed by the processor, the processor performs the following steps: Figure 1 or Figure 2 Steps shown.

[0149] In one embodiment, a computer-readable storage medium is provided, which stores a computer program. When the computer program is executed by a processor, the processor performs the following steps: Figure 1 or Figure 2 Steps shown.

[0150] Those skilled in the art can understand that all or part of the processes in the above-mentioned embodiment methods can be completed by instructing the relevant hardware through a computer program, and the program can be stored in a non-volatile computer-readable storage medium. When the program is executed, it can include the processes of the embodiments of the above-mentioned methods. Among them, any reference to memory, storage, database or other media used in the embodiments provided in this application can include non-volatile and / or volatile memory. Non-volatile memory may include read-only memory (ROM), programmable ROM (PROM), electrically programmable ROM (EPROM), electrically erasable programmable ROM (EEPROM) or flash memory. Volatile memory may include random access memory (RAM) or external cache memory. As an illustration and not limitation, RAM is available in many forms, such as static RAM (SRAM), dynamic RAM (DRAM), synchronous DRAM (SDRAM), double data rate SDRAM (DDRSDRAM), enhanced SDRAM (ESDRAM), synchronous link (Synchlink) DRAM (SLDRAM), memory bus (Rambus) direct RAM (RDRAM), direct memory bus dynamic RAM (DRDRAM), and memory bus dynamic RAM (RDRAM).

[0151] The technical features of the above embodiments may be arbitrarily combined. To make the description concise, not all possible combinations of the technical features in the above embodiments are described. However, as long as there is no contradiction in the combination of these technical features, they should be considered to be within the scope of this specification.

[0152] The above-mentioned embodiments only express several implementation methods of the present application, and the descriptions thereof are relatively specific and detailed, but they cannot be understood as limiting the scope of the present application. It should be pointed out that, for a person of ordinary skill in the art, several variations and improvements can be made without departing from the concept of the present application, and these all belong to the protection scope of the present application. Therefore, the protection scope of the present application shall be subject to the attached claims.

Claims

1. A transformer multi-physics field direct coupling reduction processing method, characterized in that: The method comprises: Acquire and use geometric parameters, material properties and operating conditions of the transformer to establish a simulation geometric model of the transformer; Based on the heat generation, fluid flow and fluid-solid heat transfer process of the transformer during operation, an initial coupling model of direct coupling of thermal-fluid-solid multi-physics fields of the transformer is constructed; Using random forest and the initial coupling model, a reduced-order model of the thermal-fluid-solid multi-physics field is established; A multi-physical field operation simulation of the transformer is performed according to the simulation geometric model and the reduced-order model to obtain an operation simulation result, and the operation simulation result is used to reflect the operating condition of the transformer in the multi-physical field.

2. The method according to claim 1, characterized in that: The use of random forest and the initial coupling model to establish the reduced-order model of the heat-fluid-solid multi-physics field includes: Calculate the importance of all candidate data features by random forest, wherein the candidate data features are used to indicate data features that affect the running simulation results when the multi-physics field running simulation of the transformer is performed; Calculate the Pearson correlation coefficient between each pair of the candidate data features; Determine whether there is a strong correlation between the candidate data features based on the Pearson correlation coefficient; If there is a strong correlation between the candidate data features, then the candidate data features are determined to be redundant features. The target data features with low importance among the redundant features are removed from the initial coupling model to obtain a reduced-order model.

3. The method according to claim 2, characterized in that: The random forest includes multiple decision trees, and the importance of all candidate data features is calculated by the random forest, including: Determine the target contribution of each candidate data feature in each decision tree; A comprehensive calculation is performed based on the respective target contributions to determine the total contribution of each of the candidate data features, wherein the importance includes the total contribution.

4. The method according to claim 3, characterized in that: Determining the target contribution of each candidate data feature in each decision tree includes: Calculate the first Gini index of the candidate data feature at the m-th node, the second Gini index of the l-th node, and the third Gini index of the r-th node of the k-th decision tree, where the l-th node and the r-th node are branch nodes corresponding to the splitting of the m-th node; Determine the contribution of the candidate data feature to the mth node of the kth decision tree by using the first Gini index, the second Gini index, and the third Gini index; A comprehensive calculation is performed based on the contribution of the candidate data feature at each node of the i-th decision tree to determine the target contribution of the candidate data feature at the k-th decision tree.

5. The method according to claim 2, characterized in that: The calculating of the Pearson correlation coefficients between the candidate data features each pairwise includes: Determining the standard deviation of the candidate data features and the covariance between each pair of the candidate data features; The Pearson correlation coefficient between each pair of the candidate data features is obtained using a preset Pearson correlation coefficient algorithm, standard deviation and covariance.

6. The method according to claim 4, characterized in that: The total contribution is determined using the following mathematical expression: In the formula, VIM(X i ) is the candidate data feature X i Total contribution of VIM k (X i ) is the candidate data feature X i The target contribution of the kth decision tree, K is the total number of decision trees, where Where VIM k,m (X i ) is the candidate data feature X i The contribution of node m in the kth decision tree, N is X i The number of splits on decision tree k, where VIM k,m (X i )=GI k,m (X i )-[GI k,l (X i )+GI k,r (X i )]; In the formula, GI k,m (X i ) is the first Gini index, GI k,l (X i ) is the second Gini index, GI k,r (X i ) is the third Gini index.

7. The method according to claim 5, characterized in that: The Pearson correlation coefficient algorithm includes the following mathematical expression: Where r(X,Y) is the Pearson correlation coefficient between candidate data feature X and candidate data feature Y, Cov(X,Y) represents the covariance between candidate data feature X and candidate data feature Y, Var[X] and Var[Y] represent the standard deviations of candidate data feature X and candidate data feature Y, respectively.

8. A transformer multi-physics field direct coupling reduction order processing device, characterized in that: The device comprises: The first modeling module is used to obtain and use the geometric parameters, material properties and operating conditions of the transformer to establish a simulation geometric model of the transformer; The second modeling module is used to construct an initial coupling model of direct coupling of thermal-fluid-solid multi-physics fields of the transformer based on the heat generation, fluid flow and fluid-solid heat transfer process during the operation of the transformer; A reduced-order processing module: used to establish a reduced-order model of the thermal-fluid-solid multi-physics field by using random forest and the initial coupling model; Operation simulation module: used to perform multi-physical field operation simulation of the transformer according to the simulation geometric model and the reduced-order model to obtain operation simulation results, and the operation simulation results are used to reflect the operating conditions of the transformer in the multi-physical field.

9. A computer-readable storage medium storing a computer program, characterized in that: When the computer program is executed by a processor, the processor is caused to perform the steps of the method according to any one of claims 1 to 7.

10. A computer device comprising a memory and a processor, characterized in that: The memory stores a computer program, and when the computer program is executed by the processor, the processor is caused to perform the steps of the method according to any one of claims 1 to 7.

Images