Method and device for determining temperature field of winding of oil-immersed transformer

Abstract

The application provides a kind of oil-immersed transformer winding temperature field determination method and device, belong to transformer field.The method comprises: according to target working condition and pre-established discrete point temperature model, determine the multiple discrete point temperatures of oil-immersed transformer winding under target working condition;Wherein, discrete point temperature model is determined according to different kinds of sampling working condition and the corresponding multiple discrete point temperatures inside field domain under each sampling working condition;According to multiple discrete point temperatures and pre-established winding temperature field analysis model, determine the temperature field distribution of oil-immersed transformer winding;Wherein, winding temperature field analysis model is determined according to the temperature field distribution of oil-immersed transformer under the first preset quantity of different kinds of sampling working condition, and the second preset quantity of discrete point temperatures inside field domain under target working condition.The application can quickly and accurately obtain the temperature distribution of transformer winding.

Description

Technical Field

[0001] This invention relates to the field of transformer technology, and in particular to a method and apparatus for determining the temperature field of an oil-immersed transformer winding. Background Technology

[0002] Oil-immersed power transformers are core components in power systems, enabling power transmission and voltage conversion. When a transformer experiences a sudden short circuit or severe overload, the winding temperature may rise. This not only shortens the lifespan of the internal insulation materials but also reduces the mechanical strength of the conductors, making them more susceptible to damage or deformation under electromagnetic forces. Therefore, winding temperature and hot spot temperature are crucial indicators of the operational stability and safety of oil-immersed transformers. Rapidly obtaining these temperatures during operation is essential for real-time monitoring of the transformer's status and early fault warning.

[0003] Due to limitations imposed by transformer structure and field applications, winding temperature is difficult to measure directly. Multiphysics coupling modeling and simulation are the primary techniques for obtaining transformer winding temperature distribution. Generally, the boundary value problems corresponding to each physical field can be transformed into corresponding discrete equation systems for solution using different numerical methods. However, the discrete equation systems used in the temperature field calculation of large power transformers are characterized by high order and strong nonlinearity, making the solution of these high-order nonlinear equation systems extremely time-consuming and data-intensive, which is difficult to meet the needs of engineering applications.

[0004] Related techniques can solve this problem by introducing a reduction-order algorithm, which obtains a more accurate mathematical description by using a computational order much lower than the original numerical system, while limiting the computational error to a certain range. However, the surrogate model established during the solution process has a large error, resulting in a significant error in the final temperature field distribution. Summary of the Invention

[0005] This invention provides a method and apparatus for determining the temperature field of an oil-immersed transformer winding, thereby improving the accuracy of determining the temperature field of an oil-immersed transformer winding and quickly and accurately obtaining the temperature distribution of the transformer winding.

[0006] In a first aspect, embodiments of the present invention provide a method for determining the temperature field of an oil-immersed transformer winding, comprising:

[0007] Based on the target operating condition and the pre-established discrete point temperature model, the temperatures of multiple discrete points of the oil-immersed transformer winding under the target operating condition are determined; wherein, the discrete point temperature model is determined based on different types of sampling operating conditions and the multiple discrete point temperatures corresponding to the field under each sampling operating condition.

[0008] The temperature field distribution of the oil-immersed transformer winding is determined based on the multiple discrete point temperatures and the pre-established winding temperature field analysis model; wherein, the winding temperature field analysis model is determined based on the temperature field distribution of the oil-immersed transformer under a first preset number of different types of sampling conditions, and the temperature of a second preset number of discrete points within the field under the target condition.

[0009] In one possible implementation, before determining the temperature field distribution of the oil-immersed transformer winding based on the multiple discrete point temperatures and a pre-established winding temperature field analysis model, the method further includes:

[0010] Obtain the temperature field distribution under a first preset number of different types of sampling conditions;

[0011] Based on the temperature field distribution under all sampling conditions, a snapshot matrix is ​​established;

[0012] Based on the snapshot matrix, the reduced-order mode matrix and the mode coefficient matrix are determined;

[0013] Select a second preset number of discrete points;

[0014] Construct a discrete point temperature matrix based on the discrete point temperatures corresponding to the discrete points under the target operating conditions.

[0015] The inversion matrix is ​​determined based on the discrete point temperature matrix, the reduced-order mode matrix, and the mode coefficient matrix.

[0016] A winding temperature field analysis model is established based on the discrete point temperature matrix, the inversion matrix, and the reduced-order mode matrix.

[0017] In one possible implementation, the reduced-order mode matrix and the mode coefficient matrix are determined based on the snapshot matrix, including:

[0018] Singular value decomposition is performed on the snapshot matrix to obtain the singular value representation of the snapshot matrix;

[0019] Based on the singular value representation of the snapshot matrix, an eigenorthogonal decomposition is performed to obtain the reduced-order modal representation of the snapshot matrix;

[0020] Based on the reduced-order modal representation of the snapshot matrix, the reduced-order modal matrix and modal coefficient matrix are determined.

[0021] In one possible implementation, determining the inversion matrix based on the discrete-point temperature matrix, the reduced-order mode matrix, and the mode coefficient matrix includes:

[0022] Based on the discrete point temperature matrix, determine the node number of each discrete point;

[0023] Based on the discrete point temperature matrix, the node number of each discrete point, and the modal coefficient matrix, a row vector with the same node number as the discrete point is selected from the reduced-order modal matrix to obtain the inversion matrix.

[0024] In one possible implementation, before determining the multiple discrete-point temperatures of the oil-immersed transformer windings under the target operating condition based on the target operating condition and a pre-established discrete-point temperature model, the method further includes:

[0025] Obtain various sampling conditions and the corresponding discrete point temperatures within the field under each sampling condition;

[0026] Based on multiple sampling conditions, the discrete point temperature corresponding to each sampling condition, and the response surface model, a discrete point temperature model between the sampling conditions and the discrete point temperature of the oil-immersed transformer winding is established.

[0027] In one possible implementation, before determining the temperature field distribution of the oil-immersed transformer winding based on the multiple discrete point temperatures and a pre-established winding temperature field analysis model, the method further includes:

[0028] Under the target operating condition, the measured temperature of multiple discrete points of the oil-immersed transformer winding is obtained by using multiple sensors installed at the winding of the oil-immersed transformer.

[0029] Determine the temperature difference between the measured temperature and the discrete point temperature for each discrete point;

[0030] Determine whether the temperature difference is greater than a preset threshold;

[0031] If the temperature difference is greater than the preset threshold, the discrete point temperature model is adjusted according to the measured temperature.

[0032] In one possible implementation, based on the discrete point temperature matrix, the node number of each discrete point, and the modal coefficient matrix, a row vector with the same node number as the discrete point is selected from the reduced-order modal matrix to obtain the inversion matrix, including:

[0033] According to the expression: Determine the inversion matrix;

[0034] In the formula, T D p represents the discrete-point temperature matrix. j α represents the row vector corresponding to the j-th discrete point in the inversion matrix. j u' represents the modal coefficient corresponding to the j-th discrete point in the modal coefficient matrix. kThis represents the row vector containing the k-th node number corresponding to the discrete points in the reduced-order mode matrix, where j represents the j-th discrete point, d represents the total number of discrete points, and k represents the k-th node number. Dj Represents the row vector corresponding to the j-th discrete point in the discrete point temperature matrix, index(T) Dj ) represents the node number corresponding to the j-th discrete point, k = index(T) Dj The ) indicates that the node number of the k-th node in the reduced-order mode matrix corresponds to the node number of the j-th discrete point.

[0035] In one possible implementation, the expression for the winding temperature field analysis model is:

[0036] T = U'P -1 T D ;

[0037] In the formula, T represents the snapshot matrix, U' represents the reduced mode matrix, P represents the inversion matrix, and T D This represents the temperature matrix at discrete points.

[0038] In one possible implementation, the expression for the discrete-point temperature model is:

[0039] T Dx =Xβ;

[0040] In the formula, T Dx Let represent the temperature at the x-th discrete point, X represent the parameter matrix corresponding to the sampling conditions of the winding, and β represent the polynomial coefficient matrix.

[0041] Secondly, embodiments of the present invention provide a device for determining the temperature field of an oil-immersed transformer winding, comprising:

[0042] The first determining module is used to determine multiple discrete point temperatures of the oil-immersed transformer winding under the target operating condition based on the target operating condition and a pre-established discrete point temperature model; wherein, the discrete point temperature model is determined based on different types of sampling operating conditions and multiple discrete point temperatures corresponding to the field under each sampling operating condition.

[0043] The second determining module is used to determine the temperature field distribution of the oil-immersed transformer winding based on the multiple discrete point temperatures and a pre-established winding temperature field analysis model; wherein, the winding temperature field analysis model is determined based on the temperature field distribution of the oil-immersed transformer under a first preset number of different types of sampling conditions, and the temperature of a second preset number of discrete points within the field under the target condition.

[0044] The beneficial effects of the embodiments of the present invention compared with the prior art are as follows:

[0045] This invention establishes a discrete point temperature model by sampling different operating conditions of an oil-immersed transformer and multiple discrete point temperatures within the field under each sampling condition. This model establishes the relationship between the sampling operating conditions and discrete point temperatures of the oil-immersed transformer. Furthermore, by using the temperature field distribution under different sampling operating conditions of the oil-immersed transformer and the discrete point temperatures within the field under the target operating condition, a winding temperature field analysis model is established. This takes into account the significant correlation between discrete point temperatures and external operating conditions, as well as the significant correlation between discrete point temperatures and temperature field distribution. Finally, by using multiple discrete point temperatures of the oil-immersed transformer windings under the target operating condition and the pre-established winding temperature field analysis model, the temperature field distribution of the oil-immersed transformer windings is determined. This allows for rapid and accurate derivation of the entire field temperature from discrete point temperatures, improving the accuracy of the obtained oil-immersed transformer winding temperature field distribution. Attached Figure Description

[0046] To more clearly illustrate the technical solutions in the embodiments of the present invention, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0047] Figure 1 This is a schematic diagram of the structure of an oil-immersed transformer winding provided in an embodiment of the present invention;

[0048] Figure 2 This is a flowchart illustrating the implementation of the method for determining the winding temperature field of an oil-immersed transformer provided in this embodiment of the invention.

[0049] Figure 3 This is a schematic diagram of the intrinsic orthogonal decomposition process provided in an embodiment of the present invention;

[0050] Figure 4 This is a schematic diagram illustrating the composition of the inversion matrix provided in an embodiment of the present invention;

[0051] Figure 5 This is a schematic diagram of the singular value distribution under different modal orders provided in the embodiments of the present invention;

[0052] Figure 6 This is a schematic diagram of the energy ratio under different modal orders provided in the embodiments of the present invention;

[0053] Figure 7 This is a schematic diagram of the temperature field distribution results provided in an embodiment of the present invention. Figure 7 In the diagram, A represents the temperature field distribution results for mode order 1. Figure 7 In the diagram, B represents the temperature field distribution results for mode order 5. Figure 7C in the diagram represents the temperature field distribution results for mode order 10. Figure 7 D in the diagram is a schematic representation of the temperature field distribution for mode order 15.

[0054] Figure 8 This is a schematic diagram illustrating the discrete point temperature calculation error provided in an embodiment of the present invention;

[0055] Figure 9 This is a schematic diagram of the error distribution under different sampling conditions provided in the embodiments of the present invention;

[0056] Figure 10 This is a comparative schematic diagram of the temperature field distribution provided in the embodiments of the present invention;

[0057] Figure 11 This is a comparative schematic diagram of the hotspot temperature calculation results provided in the embodiments of the present invention;

[0058] Figure 12 This is a schematic diagram of the hotspot temperature error provided in an embodiment of the present invention;

[0059] Figure 13 This is a schematic diagram of the structure of the oil-immersed transformer winding temperature field determination device provided in an embodiment of the present invention. Detailed Implementation

[0060] In the following description, specific details such as particular system architectures and techniques are set forth for illustrative purposes and not for limitation, in order to provide a thorough understanding of the embodiments of the invention. However, those skilled in the art will understand that the invention can be implemented in other embodiments without these specific details. In other instances, detailed descriptions of well-known systems, apparatuses, circuits, and methods are omitted so as not to obscure the description of the invention with unnecessary detail.

[0061] To make the objectives, technical solutions, and advantages of the present invention clearer, specific embodiments will be described below in conjunction with the accompanying drawings.

[0062] The inventors have discovered that analyzing and determining the temperature field of oil-immersed transformer windings typically requires solving high-order nonlinear equations, incurring significant time and data storage costs. Two common approaches address this issue: one is to employ parallel computing techniques to improve hardware computational capabilities for high-performance solutions to large-scale equation systems; the other is to introduce order reduction algorithms, which use a computational order far less than the original numerical system to obtain a more accurate mathematical description. However, the former places high demands on hardware, resulting in high costs for large-scale engineering-level simulations. Furthermore, current parallel technologies are quite mature, making it difficult to achieve additional efficiency gains under the same hardware conditions. The latter, when performing order reduction calculations for nonlinear problems, still requires the nonlinear terms to be formed in the original space, making the computational efficiency improvement of the Proper Orthogonal Decomposition (POD) method in the face of large-scale strongly nonlinear problems insignificant.

[0063] In the rapid solution of parameterized partial differential equations with complex nonlinear terms, there exists the concept of non-intrusive POD (Programmable Optimization) mode coefficients. This involves fully utilizing order reduction algorithms to approximate the characteristics of the original full-order system using low-dimensional data. Using the reduced-order model as a carrier, a proxy relationship is established between external monitoring data and the full-order system, thereby bypassing complex nonlinear calculation processes and achieving efficient computation. Although external factors affect the temperature field distribution and thus change the POD mode coefficients, there is generally no clear correlation between the POD mode coefficients and the external influencing factors. The proxy model between the two has a large error, resulting in significant errors in the temperature field distribution obtained by combining the reduced-order modes, even under relatively ideal computational conditions. Therefore, it is necessary to consider a technique for determining the temperature field distribution of oil-immersed transformer windings.

[0064] To improve the accuracy of determining the temperature field of oil-immersed transformer windings, this invention establishes a relationship between the operating conditions of the oil-immersed transformer and discrete point temperatures through a discrete point temperature model, and a relationship between discrete point temperatures and temperature field distribution through a winding temperature field analysis model. Significant correlations exist between the transformer's operating conditions and discrete point temperatures, as well as between discrete point temperatures and temperature field distribution. By progressively deriving from transformer operating conditions to discrete point temperatures and then to temperature field distribution, the entire field temperature can be quickly and accurately derived, thus improving the accuracy of the obtained oil-immersed transformer winding temperature field distribution.

[0065] Figure 1 This is a schematic diagram of the structure of an oil-immersed transformer winding provided in an embodiment of the present invention. Figure 1 The image shows the windings of a 110kV oil-immersed transformer, from left to right: low-voltage, medium-voltage, and high-voltage windings. The transformer as a whole uses natural oil circulation for heat dissipation.

[0066] Figure 2 The implementation flowchart of the method for determining the winding temperature field of an oil-immersed transformer provided in the embodiments of the present invention is described in detail below:

[0067] Step S201: Based on the target operating condition and the pre-established discrete point temperature model, determine the multiple discrete point temperatures of the oil-immersed transformer winding under the target operating condition; wherein, the discrete point temperature model is determined based on different types of sampling operating conditions and the multiple discrete point temperatures corresponding to the field under each sampling operating condition.

[0068] In this embodiment, to obtain the temperature of discrete points within the field, and considering the significant correlation between the operating conditions of the oil-immersed transformer windings and the discrete point temperatures, a proxy model can be constructed between the sampled operating conditions of the oil-immersed transformer windings and their temperatures. This proxy model allows for the accurate determination of the discrete point temperatures under the target operating conditions.

[0069] Step S202: Determine the temperature field distribution of the oil-immersed transformer winding based on multiple discrete point temperatures and a pre-established winding temperature field analysis model; wherein, the winding temperature field analysis model is determined based on the temperature field distribution of the oil-immersed transformer under a first preset number of different types of sampling conditions, and the second preset number of discrete point temperatures within the field under the target condition.

[0070] In this embodiment, the discrete point temperature is a part of the field of the winding temperature field. By the correlation between multiple discrete point temperatures and the temperature field distribution, the global temperature field distribution can be derived from multiple discrete point temperatures. That is, by using the winding temperature field analysis model to analyze multiple discrete point temperatures, the temperature field distribution of the oil-immersed transformer winding can be obtained.

[0071] This invention establishes a discrete point temperature model by sampling different operating conditions of an oil-immersed transformer and multiple discrete point temperatures within the field under each sampling condition. This model establishes the relationship between the sampling operating conditions and discrete point temperatures of the oil-immersed transformer. Furthermore, by using the temperature field distribution under different sampling operating conditions of the oil-immersed transformer and the discrete point temperatures within the field under the target operating condition, a winding temperature field analysis model is established. This takes into account the significant correlation between discrete point temperatures and external operating conditions, as well as the significant correlation between discrete point temperatures and temperature field distribution. Finally, by using multiple discrete point temperatures of the oil-immersed transformer windings under the target operating condition and the pre-established winding temperature field analysis model, the temperature field distribution of the oil-immersed transformer windings is determined. This allows for rapid and accurate derivation of the entire field temperature from discrete point temperatures, improving the accuracy of the obtained oil-immersed transformer winding temperature field distribution.

[0072] In some embodiments, before determining the temperature field distribution of the oil-immersed transformer winding based on multiple discrete point temperatures and a pre-established winding temperature field analysis model, a first preset number of temperature field distributions under different sampling conditions can be obtained; a snapshot matrix can be established based on the temperature field distributions under all sampling conditions; and a reduced-order mode matrix and mode coefficient matrix can be determined based on the snapshot matrix; then a second preset number of discrete points can be selected; a discrete point temperature matrix can be constructed based on the discrete point temperatures corresponding to the discrete points under the target operating condition; and an inversion matrix can be determined based on the discrete point temperature matrix, the reduced-order mode matrix, and the mode coefficient matrix; finally, a winding temperature field analysis model can be established based on the discrete point temperature matrix, the inversion matrix, and the reduced-order mode matrix.

[0073] In this embodiment, multiple different types of sampling conditions can be selected to obtain the corresponding temperature field distribution results. The temperature field distributions of all sampling conditions are arranged to form a matrix, which is the snapshot matrix.

[0074] For example, if s different types of sampling conditions are selected, the temperature field distribution corresponding to the i-th condition is T. i Then the snapshot matrix can be represented as T = {T1, T2, ..., T} i ,……,T s}

[0075] By performing eigenorthogonal decomposition on the snapshot matrix, we can obtain the reduced-order orthogonal basis, i.e., the reduced-order modes, and the orthogonal basis coefficients, i.e., the mode coefficients, in the reduced-order mode matrix and the mode coefficient matrix, thus achieving the reduction of the order of the snapshot matrix.

[0076] The winding temperature field analysis model established in this embodiment is the relationship between multiple discrete point temperatures and the temperature field distribution. It is necessary to infer the temperature field distribution from the discrete point temperatures. Therefore, the relationship between the discrete point temperature matrix, the reduced-order mode matrix and the mode coefficient matrix is ​​established first to obtain the inversion matrix. Then, the winding temperature field analysis model is established using the discrete point temperature matrix, the reduced-order mode matrix and the inversion matrix to realize the inversion from the discrete point temperatures to the temperature field distribution.

[0077] Here, the temperature field distribution under a first preset number of different sampling conditions can be obtained through calculation using the finite element method. Specifically:

[0078] The calculation of winding temperature in an oil-immersed transformer is a multi-field calculation process involving flow field and temperature field, etc. According to the law of conservation of energy, the expression of its governing equation can be determined as follows:

[0079]

[0080] In the formula, This represents the thermal convection term, and ρ represents the density of the medium, in kg / m³. 3 C p λ represents the specific heat capacity at constant pressure, with units of J / (kg·K), λ represents the thermal conductivity, with units of W / (m·K), U represents the fluid velocity, and T represents the temperature. S represents the heat conduction term, and the right-hand side represents the heat source term. T This represents the density of the heat source, expressed in W / m³. 3 .

[0081] By combining the boundary conditions of the equations and using the Galerkin finite element method for discretization, the discrete governing equations can be obtained:

[0082] KT = B;

[0083] In the formula, K represents the finite element stiffness matrix, K∈R n×n T represents the solution vector of the temperature field, T∈R n×1 B represents the right-hand side term of the discrete control equation, B∈R n×1 .

[0084] By combining the above formula with appropriate numerical methods, the temperature of transformer windings can be calculated, and the results of temperature field distribution under multiple different operating conditions can be obtained, thereby establishing a winding temperature field analysis model of the non-intrusive POD algorithm.

[0085] Furthermore, as can be seen from the above equations, this system of equations is complex in form and computationally intensive. Directly applying it to the calculation of winding temperature in large power transformers would require significant time and storage costs. Therefore, this embodiment introduces the POD method to reduce the computational scale of the equation system, thereby improving computational efficiency.

[0086] Optionally, in this embodiment, the reduced-order mode matrix and mode coefficient matrix are determined based on the snapshot matrix. This can be achieved by first performing singular value decomposition on the snapshot matrix to obtain its singular value representation; then performing eigenorthogonal decomposition based on the singular value representation of the snapshot matrix to obtain its reduced-order mode representation; and finally, determining the reduced-order mode matrix and mode coefficient matrix based on the reduced-order mode representation of the snapshot matrix.

[0087] In this embodiment, see Figure 3 The diagram shown illustrates the eigenorthogonal decomposition process. Singular value decomposition of the snapshot matrix can be performed using left singular matrix, right singular matrix, and singular value matrix. The snapshot matrix can be expressed as: T = U∑V T In the formula, T represents the snapshot matrix, U represents the left singular matrix, ∑ represents the singular value matrix, V represents the right singular matrix, and T represents the transpose of the matrix.

[0088] Both the left and right singular matrices are orthogonal matrices. The singular value matrix is ​​a diagonal matrix, and its diagonal elements are the singular values ​​of the snapshot matrix T arranged in descending order.

[0089] In general, the first d singular values ​​in the singular value matrix ∑ are much larger than the subsequent singular values. Therefore, the snapshot matrix T can be approximated by the first d singular values ​​in the singular value matrix ∑ and their corresponding left and right singular value matrices.

[0090] That is, T = U∑V T ≈U′∑′V′ T Where U′ denotes an approximate left singular matrix, U′∈R n×d ∑′ denotes an approximate singular value matrix, ∑′∈R d×d V′ denotes an approximate right singular matrix, V′∈R n×d .

[0091] Here, since the first d singular values ​​are much larger than the subsequent singular values ​​(i.e., d << n), the approximate left singular matrix U′ can be regarded as an orthogonal basis of the snapshot matrix T in the reduced-order subspace.

[0092]

[0093] In the formula, U′ represents the reduced-order mode matrix, or the reduced-order orthogonal basis matrix, which is the approximate left singular matrix mentioned above; α represents the mode coefficient matrix, that is, the orthogonal basis coefficients, α∈R d×1 u' i Let α represent the row vector corresponding to the i-th node number in the reduced-order mode matrix. i This represents the vector corresponding to the i-th node number in the reduced-order mode matrix. The reduced-order mode matrix and the mode coefficient matrix together constitute the expression of the solution vector in the reduced-order subspace, thus reducing the order of the snapshot matrix from n to d.

[0094] In some embodiments, the inversion matrix is ​​determined based on the discrete point temperature matrix, the reduced-order mode matrix, and the mode coefficient matrix. This can be achieved by first determining the node number of each discrete point based on the discrete point temperature matrix, and then selecting a row vector from the reduced-order mode matrix that matches the node number of the discrete point, based on the discrete point temperature matrix, the node number of each discrete point, and the mode coefficient matrix, to obtain the inversion matrix.

[0095] In this embodiment, assuming the temperatures of m discrete points within the field are known, the discrete point temperature matrix can be represented as T. D =[T D1 ,T D2 ,…,T Dj ,…,T Dm], where 1 to m are the serial numbers of discrete points. It is necessary to link the discrete points with the global temperature field. Therefore, it is necessary to determine the node numbers of the discrete points in the global temperature field.

[0096] See afterward. Figure 4 The diagram shows the construction of the inversion matrix. To represent the discrete point temperature matrix, it can be extracted from the reduced-order mode matrix. Row vectors with the same node numbers as the discrete points form a new matrix, namely the inversion matrix P, which can be expressed as P = [p1, p2, ..., p...]. j ,…,p m ].

[0097] Here, during the extraction process, the discrete point temperature matrix can be substituted into the formulas expressed by the reduced-order mode matrix and the mode coefficient matrix to determine the specific expression of the inversion matrix.

[0098] Optionally, in this embodiment, based on the discrete point temperature matrix, the node number of each discrete point, and the modal coefficient matrix, a row vector with the same node number as the discrete point is selected from the reduced-order modal matrix to obtain the inversion matrix. This can be done according to the expression: Determine the inversion matrix; where T D p represents the discrete-point temperature matrix. j Let α represent the row vector corresponding to the j-th discrete point in the inversion matrix. j u' represents the modal coefficient corresponding to the j-th discrete point in the modal coefficient matrix. k This represents the row vector containing the node number of the k-th discrete point in the reduced-order mode matrix, where j represents the j-th discrete point, m represents the total number of discrete points, and k represents the node number of the k-th node. Dj This represents the row vector corresponding to the j-th discrete point in the discrete-point temperature matrix, index(T) Dj ) represents the node number corresponding to the j-th discrete point, k = index(T) Dj The ) indicates that the node number of the k-th node in the reduced-order mode matrix corresponds to the node number of the j-th discrete point.

[0099] In this embodiment, the row vectors corresponding to each discrete point in the inversion matrix are selected from the reduced-order mode matrix using the above expression, and then these row vectors are combined to form the inversion matrix.

[0100] Optionally, if the number of discrete points m selected is equal to the number of singular values ​​d selected in the reduced-order mode matrix, i.e., m = d, then matrix P is an invertible matrix, such as... Figure 4 As shown, a 3×3 matrix is ​​represented. Therefore, the expression for the winding temperature field analysis model can be obtained as: T=U'α=U'P -1 T DIn the formula, T represents the snapshot matrix, U' represents the reduced-order mode matrix, α represents the mode coefficient matrix, P represents the inversion matrix, and T D This represents the temperature matrix at discrete points.

[0101] In this embodiment, the mode coefficient matrix is ​​represented by the discrete point temperature matrix and the inversion matrix, thereby constructing a winding temperature field analysis model. Using the reduced-order model as a carrier, a proxy relationship is established between the discrete point temperature inside the field and the temperature field distribution of the whole field, thereby skipping the complex nonlinear calculation process and achieving efficient calculation.

[0102] Considering that the location and number of discrete points significantly affect the accuracy of the temperature field distribution, a greedy algorithm can be used to select the discrete points in this embodiment. The search approach can be as follows:

[0103] Input the reduced-order mode matrix from the established winding temperature field analysis model; the reduced-order mode matrix includes multiple first-order reduced-order modes, i.e., the selected singular values.

[0104] The node number containing the element with the largest absolute value in the first reduced mode u'1 of the first group in the reduced mode matrix is ​​determined as the position of the first discrete point, and the first column p1 of the inversion matrix is ​​established.

[0105] For the first reduced-order mode of each group from the second group onwards, the discrete points and inversion matrix determined by the previous group of the first reduced-order mode of that group are used to calculate the second reduced-order mode of that group, and the residual corresponding to each element in the second reduced-order mode is calculated. The node number of the element with the largest residual is determined as the position of the discrete point corresponding to the first reduced-order mode of that group.

[0106] Discrete points are selected based on the positions of the discrete points corresponding to the first reduced-order modes in each group.

[0107] Here, the larger the residual, the higher the importance of the point. Choosing this point as the discrete point can significantly improve the accuracy of the final determined temperature field distribution.

[0108] The previous section introduced the process of establishing a winding temperature field analysis model and selecting discrete points. However, determining the temperature field distribution also requires obtaining the temperatures at discrete points. The following section will continue to introduce the process of establishing a discrete point temperature model.

[0109] In some embodiments, before determining multiple discrete point temperatures of the oil-immersed transformer winding under the target operating condition based on the target operating condition and the pre-established discrete point temperature model, various different types of sampling operating conditions and the corresponding discrete point temperatures within the field under each sampling operating condition can be obtained first; then, based on the various sampling operating conditions, the discrete point temperatures corresponding to each sampling operating condition, and the response surface model, a discrete point temperature model between the sampling operating conditions and the discrete point temperatures of the oil-immersed transformer winding can be established.

[0110] In this embodiment, the discrete point temperature is obtained using the response surface methodology. A response surface model is used to establish the relationship between the sampling operating conditions of the winding and the discrete point temperature.

[0111] Here, the response surface model is a functional expression representing the relationship between influencing factors and the objective solution quantity. The polynomial response surface function is simple in structure and convenient to calculate. The following section uses the polynomial response surface function as an example for introduction.

[0112] Polynomial response surface models are generally quadratic, so the expression for the quadratic polynomial response surface function reflecting the relationship between the parameters of the sampled operating conditions of n windings and the discrete point temperature can be:

[0113]

[0114] In the formula, T Dx The temperature at a discrete point under the parameter x of the sampling condition is represented, where x represents the parameter of the sampling condition of the winding. i The i-th parameter, x, represents the sampling condition of the winding. j The first parameter, β0, represents the sampling condition of the winding. i β ii and β ij Let represent the polynomial coefficients, n represent the set of sampling conditions for different types of windings, i represent the i-th sampling condition, and j represent the j-th sampling condition.

[0115] Assuming that s trials are needed to obtain each coefficient, the above formula can be written in the following matrix form:

[0116]

[0117] Or:

[0118] T Dx =Xβ+e;

[0119] In the formula, T Dx1 This represents the temperature of the x-th discrete point in the first experiment. and The parameters β0, β1, and β2 represent the sampling conditions in the first test. n β 11 β nn β 12 and β (n-1)n Equations represent the polynomial coefficients, e1, e2, and e... s Equation T represents the error between the calculated and actual values ​​in each trial. DxLet X represent the matrix corresponding to the temperature of the x-th discrete point formed in each experiment, let X represent the parameter matrix corresponding to the parameters of the sampling conditions of the winding, let β represent the coefficient matrix of the polynomial coefficients, and let e represent the error matrix between the calculated value and the actual value.

[0120] To obtain the polynomial coefficients, we need to minimize the sum of squares of the errors in the above equation, that is:

[0121] mine T e = min(T) Dx -Xβ) T (T Dx -Xβ);

[0122] Taking the partial derivative of the above equation with respect to the coefficient matrix β, we get:

[0123] 2XXβ-2XT Dx =0;

[0124] Therefore, 2XXβ-2XT Dx The least squares solution for =0 is:

[0125] β=(X T X) -1 X T T Dx .

[0126] This yields the polynomial coefficients in the quadratic polynomial response surface function. Consequently, a surface model of the temperature at each discrete point with respect to the winding operating parameters can be established, i.e., a discrete point temperature model. The discrete point temperature can then be quickly obtained using the parameters of the sampled winding operating conditions.

[0127] Optionally, the expression for the discrete-point temperature model in this embodiment is: T Dx =Xβ; where T Dx Let represent the temperature at the x-th discrete point, X represent the parameter matrix corresponding to the sampling operating conditions of the winding, and β represent the polynomial coefficient matrix.

[0128] In this embodiment, the polynomial coefficient matrix corresponding to each discrete point can be solved by the above method, thereby establishing a discrete point temperature model corresponding to each discrete point.

[0129] In some embodiments, before determining the temperature field distribution of the oil-immersed transformer winding based on multiple discrete point temperatures and a pre-established winding temperature field analysis model, the measured temperatures of multiple discrete points of the oil-immersed transformer winding can be obtained under the target operating condition using multiple sensors installed at the oil-immersed transformer winding; then, the temperature difference between the measured temperature and the discrete point temperature corresponding to each discrete point is determined; it is determined whether the temperature difference is greater than a preset threshold; if the temperature difference is greater than the preset threshold, the discrete point temperature model is adjusted based on the measured temperature.

[0130] In this embodiment, sensors can be installed at discrete points on the oil-immersed transformer to measure the temperature corresponding to each discrete point, obtaining the measured temperature. The measured temperature is then compared with the temperature calculated by the discrete point temperature model. If the difference between the two is greater than a preset threshold, it indicates that the currently used discrete point temperature model may be inaccurate, possibly due to wear and tear of the components caused by long-term use.

[0131] Therefore, in order to reduce the inaccuracy of the discrete-point temperature model caused by wear changes during long-term use of oil-immersed transformers, the accuracy of the discrete-point temperature model can be improved by adjusting the polynomial coefficients in the model, and then the discrete-point temperature and temperature field distribution can be determined.

[0132] The above describes the process of determining the discrete-point temperature corresponding to a discrete point using a discrete-point temperature model, as well as the process of establishing the relevant model. Alternatively, the discrete-point temperature can also be obtained directly by setting up a sensor.

[0133] In other embodiments, sensors can be placed at locations corresponding to discrete points in the winding to obtain the measured temperature of the corresponding discrete points, which can then be used as the discrete point temperature to directly determine the temperature field distribution, thereby further improving the accuracy and precision of the determined temperature field distribution.

[0134] In some feasible embodiments, with Figure 1 The experiment was conducted using an oil-immersed transformer as an example. The model of this oil-immersed transformer, from left to right, consists of low-voltage, medium-voltage, and high-voltage windings. A non-turn-by-turn modeling method was adopted, and an equivalent thermal conductivity was set to simplify the influence of the insulating paper on the heat dissipation of the windings. The low-voltage, medium-voltage, and high-voltage windings contain 80, 96, and 90 coils, respectively. The transformer as a whole adopts a natural oil circulation heat dissipation method.

[0135] A non-invasive order reduction model was established using the methods described in the above embodiments. Two factors that significantly affect the winding temperature, namely load rate and ambient temperature, were selected to form the sample space. The load rate ranged from 50% to 140%, and the ambient temperature ranged from 278K to 313K.

[0136] 125 sample points are selected in the sample space, and the transformer temperature field distribution corresponding to these 125 sample points is calculated. They are then combined into a snapshot matrix and a POD reduced-order calculation model is established to obtain the transformer winding temperature field distribution model.

[0137] The order of the reduction can be determined by the following expression: In the formula, ε(d) represents the percentage of the physical system energy that the selected d-th order mode can represent, and σ id represents the i-th singular value of the diagonal elements of the singular value matrix, d represents the order of reduction, and n represents the set of selected sample points, that is, the set of sampling conditions.

[0138] When the selected order of reduction d is such that ε(d) > 99.9%, the reduced order model can be considered to cover most of the features of the full-order model and can be used for order reduction calculation.

[0139] Different modal orders correspond to different singular values ​​and energy ratios, see [link to relevant documentation]. Figure 5 The diagram illustrates the distribution of singular values ​​at different modal orders, where the horizontal axis represents the reduced order and the vertical axis represents the magnitude of the singular values. Figure 5 It can be seen that the singular values ​​of the snapshot matrix gradually decrease with the increase of the modal order, and the singular value corresponding to the first-order mode is 1.5 × 10⁻⁶. 6 The singular values ​​corresponding to the 10th mode are all less than 236.

[0140] as well as Figure 6 The diagram shows the energy ratios at different modal orders, where the horizontal axis represents the reduced order and the vertical axis represents the energy ratio. Figure 6 It can be seen that when the reduced mode order is greater than 10, the energy ratio is greater than 99.9%, which meets the accuracy requirements of the reduced mode order calculation model.

[0141] To further illustrate the feasibility of the number of reduced-order modes selected, taking a single sampling condition as an example, the temperature field distribution results obtained from the inversion of different reduced-order modes are as follows: Figure 7 As shown, Figure 7 In the diagram, A represents the temperature field distribution results for mode order 1. Figure 7 In the diagram, B represents the temperature field distribution results for mode order 5. Figure 7 C in the diagram represents the temperature field distribution results for mode order 10. Figure 7 D in the diagram is a schematic diagram of the temperature field distribution results for mode order 15, where the left side represents the temperature field distribution and the right side represents the error of the temperature field distribution.

[0142] Depend on Figure 7 It can be seen that as the number of reduced modes increases, the error in temperature field inversion further decreases: the maximum calculation error is approximately 7.1K for the first-order mode; the maximum calculation error decreases to about 4.2K for the fifth-order mode; and when the mode order is greater than 10, the maximum reduction error decreases to below 2K and gradually stabilizes. Therefore, a 10th-order mode can be selected to ensure the overall calculation accuracy of the reduction calculation.

[0143] In other feasible embodiments, the effectiveness of the discrete-point temperature model established by the response surface methodology can also be verified.

[0144] One hundred test conditions can be selected from the sample space as sampling conditions to verify the fitting of the response surface model to the temperature at 10 discrete points under the 100 test conditions. The error of the discrete points under each test condition is:

[0145]

[0146]

[0147] In the formula, T represents the simulated value of the discrete point temperature, which is used as the reference value. RSM N represents the discrete-point temperature determined by the discrete-point temperature model established using the response surface methodology. D This indicates the total number of sampling conditions.

[0148] See Figure 8 The diagram shown illustrates the error in calculating discrete point temperatures. The horizontal axis represents different test conditions, the left vertical axis represents the average temperature error, and the right vertical axis represents the average relative error. Figure 8 It can be seen that the discrete-point temperature model determined by the polynomial response surface method calculates the temperature of each discrete point under the test conditions. The maximum average temperature error is 1.29K and the maximum average relative error is 3.37%, both of which meet the error requirements for temperature calculation under normal circumstances, demonstrating the accuracy of the polynomial response surface model established in this invention.

[0149] After establishing the response surface model between discrete point temperatures and winding operating conditions, the winding temperature field distribution under the corresponding operating conditions can be quickly calculated. To verify the calculation accuracy of this method, 100 sets of test conditions were randomly selected from the sample space as sampling conditions. The winding temperature field distribution corresponding to the 100 test points was calculated using Fluent simulation software and the method described in the above embodiments of the present invention. To quantify the temperature field calculation error, the following error analysis indicators were set:

[0150]

[0151]

[0152] In the formula, T ROM T represents the temperature field distribution corresponding to the reduced-order calculation result obtained using the method in the above embodiments of the present invention. FOM N represents the temperature field distribution corresponding to the full-order calculation results obtained through Fluent simulation software. node This represents the number of nodes in the field.

[0153] The mean absolute error and mean relative error under 100 test conditions were calculated respectively, and the results are as follows: Figure 9As shown, the horizontal axis represents different test conditions, the left vertical axis represents the average temperature error, and the right vertical axis represents the average relative error. Figure 9 It can be seen that the maximum average absolute error in the temperature field distribution calculation is 0.49 K, and the maximum average relative error is 2.69%. The former occurs when the load rate is 50% and the ambient temperature is 303 K, while the latter occurs when the load rate is 80% and the ambient temperature is 283 K. Apart from this, the average absolute error is less than 0.03 K and the average relative error is less than 3% for the vast majority of other operating conditions. This demonstrates that the method provided by this invention has high accuracy in calculating the temperature across the entire field.

[0154] The method provided by this invention has the advantage of obtaining the winding temperature distribution with higher accuracy compared to traditional fast algorithms. Taking the operating condition with the maximum mean absolute error as an example, a comparison is made between the winding temperature field analysis model determined by the non-intrusive POD algorithm provided by this invention and the temperature field distribution obtained by Fluent simulation calculation. Figure 10 As shown, from left to right, the results are the calculation results of the winding temperature field analysis model provided by this invention, the calculation results of the Fluent full-order model, and the error cloud diagram between the two. Figure 10 It can be seen that the maximum calculation error of each node in the field does not exceed 3K, which meets the requirements of engineering applications and further illustrates the effectiveness of the method provided by this invention for calculating the temperature field of transformer windings.

[0155] Considering that the winding hot spot temperature plays a crucial role in the safe operation of the transformer, the calculated results of the winding hot spot temperature under various test conditions are compared, such as... Figure 11 As shown, the horizontal axis represents different test conditions, and the vertical axis represents the hot spot temperature; the hot spot temperature error is calculated, as follows: Figure 12 As shown in the figure, the horizontal axis represents different test conditions, and the vertical axis represents the hot spot temperature error. It can be seen that under 100 test conditions, the hot spot temperature results calculated at all order levels are approximately the same as those calculated by the reduced-order winding temperature field analysis model provided by this method, with a maximum error of 1.72K, which meets the engineering calculation error requirements. Therefore, this method still has good accuracy in calculating winding hot spot temperatures.

[0156] In summary, whether considering full-field calculation or winding hotspot calculation, the method provided by this invention has high calculation accuracy.

[0157] In addition, compared with traditional algorithms, the most significant advantage of the method provided in the embodiments of the present invention is that it has extremely high computational efficiency, performing full-order calculations (using Fluent software) under 100 test conditions and the order reduction calculations provided by the present invention.

[0158] The computation time of the method provided in the embodiments of the present invention is 0.78s, while the computation time of full-order calculation using Fluent software is 3675s.

[0159] It can be seen that by using the non-invasive POD order reduction method to calculate 100 sets of test conditions, the calculation time can be reduced to 0.78s, which is about a thousand times more efficient than the full-order calculation method. At the same time, it can realize real-time or near real-time monitoring of winding temperature under different winding conditions, meeting the requirements of digital online monitoring of power equipment.

[0160] This invention establishes a discrete point temperature model by sampling different operating conditions of an oil-immersed transformer and multiple discrete point temperatures within the field under each sampling condition. This model establishes the relationship between the sampling operating conditions and discrete point temperatures of the oil-immersed transformer. Furthermore, by using the temperature field distribution under different sampling operating conditions of the oil-immersed transformer and the discrete point temperatures within the field under the target operating condition, a winding temperature field analysis model is established. This takes into account the significant correlation between discrete point temperatures and external operating conditions, as well as the significant correlation between discrete point temperatures and temperature field distribution. Finally, by using multiple discrete point temperatures of the oil-immersed transformer windings under the target operating condition and the pre-established winding temperature field analysis model, the temperature field distribution of the oil-immersed transformer windings is determined. This allows for rapid and accurate derivation of the entire field temperature from discrete point temperatures, improving the accuracy of the obtained oil-immersed transformer winding temperature field distribution. Specifically, establishing a discrete-point temperature model using a response surface methodology allows for temperature calculations at discrete points, enabling direct use of these discrete-point temperatures for subsequent analysis. Furthermore, establishing a winding temperature field analysis model using a reduced-order model allows for the deriving of the relationship between discrete-point temperatures and the global temperature field, enabling rapid and accurate determination of the global temperature field distribution from discrete-point temperatures, thus improving computational efficiency. Additionally, discrete-point temperatures of the winding can be acquired using sensors to determine the temperature field distribution, further enhancing the accuracy of the winding temperature field distribution.

[0161] It should be understood that the sequence number of each step in the above embodiments does not imply the order of execution. The execution order of each process should be determined by its function and internal logic, and should not constitute any limitation on the implementation process of the embodiments of the present invention.

[0162] The following are device embodiments of the present invention. For details not described in detail, please refer to the corresponding method embodiments described above.

[0163] Figure 13 A schematic diagram of the structure of the oil-immersed transformer winding temperature field determination device provided in an embodiment of the present invention is shown. For ease of explanation, only the parts related to the embodiment of the present invention are shown, and are described in detail below:

[0164] like Figure 13 As shown, the oil-immersed transformer winding temperature field determination device 130 includes:

[0165] The first determining module 1301 is used to determine multiple discrete point temperatures of the oil-immersed transformer winding under the target operating condition based on the target operating condition and a pre-established discrete point temperature model; wherein, the discrete point temperature model is determined based on different types of sampling operating conditions and multiple discrete point temperatures corresponding to the field under each sampling operating condition.

[0166] The second determining module 1302 is used to determine the temperature field distribution of the oil-immersed transformer winding based on multiple discrete point temperatures and a pre-established winding temperature field analysis model. The winding temperature field analysis model is determined based on the temperature field distribution of the oil-immersed transformer under a first preset number of different types of sampling conditions and the second preset number of discrete point temperatures within the field under the target condition.

[0167] In one possible implementation, the second determining module 1302 is further configured to:

[0168] Obtain the temperature field distribution under a first preset number of different types of sampling conditions;

[0169] Based on the temperature field distribution under all sampling conditions, a snapshot matrix is ​​established;

[0170] Based on the snapshot matrix, determine the reduced-order mode matrix and the mode coefficient matrix;

[0171] Select a second preset number of discrete points;

[0172] Construct a discrete point temperature matrix based on the discrete point temperatures corresponding to the discrete points under the target operating conditions;

[0173] The inversion matrix is ​​determined based on the discrete point temperature matrix, the reduced-order mode matrix, and the mode coefficient matrix.

[0174] A winding temperature field analysis model is established based on the discrete point temperature matrix, the inversion matrix, and the reduced-order mode matrix.

[0175] In one possible implementation, the second determining module 1302 is specifically used for:

[0176] Singular value decomposition is performed on the snapshot matrix to obtain its singular value representation.

[0177] Based on the singular value representation of the snapshot matrix, an eigenorthogonal decomposition is performed to obtain the reduced-order modal representation of the snapshot matrix;

[0178] Based on the reduced-order modal representation of the snapshot matrix, the reduced-order modal matrix and modal coefficient matrix are determined.

[0179] In one possible implementation, the second determining module 1302 is specifically used for:

[0180] The node number of each discrete point is determined based on the discrete point temperature matrix.

[0181] Based on the discrete point temperature matrix, the node number of each discrete point, and the modal coefficient matrix, a row vector with the same node number as the discrete point is selected from the reduced-order modal matrix to obtain the inversion matrix.

[0182] In one possible implementation, the first determining module 1301 is further configured to:

[0183] Obtain various sampling conditions and the corresponding discrete point temperatures within the field under each sampling condition;

[0184] Based on multiple sampling conditions, the discrete point temperature corresponding to each sampling condition, and the response surface model, a discrete point temperature model between the sampling conditions and the discrete point temperature of the oil-immersed transformer winding is established.

[0185] In one possible implementation, the first determining module 1301 is further configured to:

[0186] Under the target operating conditions, the measured temperature of multiple discrete points of the oil-immersed transformer winding is obtained by using multiple sensors installed at the winding of the oil-immersed transformer.

[0187] Determine the temperature difference between the measured temperature and the discrete point temperature for each discrete point;

[0188] Determine if the temperature difference is greater than a preset threshold;

[0189] If the temperature difference exceeds the preset threshold, the discrete point temperature model will be adjusted based on the measured temperature.

[0190] In one possible implementation, the second determining module 1302 is specifically used for:

[0191] According to the expression: Determine the inversion matrix;

[0192] In the formula, T D p represents the discrete-point temperature matrix. j Let α represent the row vector corresponding to the j-th discrete point in the inversion matrix. j u' represents the modal coefficient corresponding to the j-th discrete point in the modal coefficient matrix. k Let T be the row vector representing the node number of the discrete point in the reduced-order mode matrix, where j represents the j-th discrete point, d represents the total number of discrete points, and k represents the node number of the k-th node. Dj This represents the row vector corresponding to the j-th discrete point in the discrete-point temperature matrix, index(T)Dj ) represents the node number corresponding to the j-th discrete point, k = index(T) Dj The ) indicates that the node number of the k-th node in the reduced-order mode matrix corresponds to the node number of the j-th discrete point.

[0193] In one possible implementation, the expression for the winding temperature field analysis model is:

[0194] T = U'P -1 T D ;

[0195] In the formula, T represents the snapshot matrix, U' represents the reduced mode matrix, P represents the inversion matrix, and T D This represents the temperature matrix at discrete points.

[0196] In one possible implementation, the discrete-point temperature model is expressed as:

[0197] T Dx =Xβ;

[0198] In the formula, T Dx Let represent the temperature at the x-th discrete point, X represent the parameter matrix corresponding to the sampling conditions of the winding, and β represent the polynomial coefficient matrix.

[0199] In the above embodiments, the descriptions of each embodiment have different focuses. For parts that are not described in detail or recorded in a certain embodiment, please refer to the relevant descriptions of other embodiments.

[0200] Those skilled in the art will recognize that the templates, units, and algorithm steps of the various examples described in conjunction with the embodiments disclosed herein can be implemented in electronic hardware, or a combination of computer software and electronic hardware. Whether these functions are implemented in hardware or software depends on the specific application and design constraints of the technical solution. Those skilled in the art can use different methods to implement the described functions for each specific application, but such implementations should not be considered beyond the scope of this invention.

[0201] If a module / unit is implemented as a software functional unit and sold or used as an independent product, it can be stored in a computer-readable storage medium. Based on this understanding, all or part of the processes in the methods of the above embodiments of the present invention can also be implemented by a computer program instructing related hardware. The computer program can be stored in a computer-readable storage medium, and when executed by a processor, it can implement the steps of the various method embodiments described above. The computer program includes computer program code, which can be in the form of source code, object code, executable files, or certain intermediate forms. The computer-readable medium can include: any entity or device capable of carrying computer program code, recording media, USB flash drives, portable hard drives, magnetic disks, optical disks, computer memory, read-only memory, random access memory, electrical carrier signals, telecommunication signals, and software distribution media, etc.

[0202] The above embodiments are only used to illustrate the technical solutions of the present invention, and are not intended to limit it. Although the present invention has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that modifications can still be made to the technical solutions described in the foregoing embodiments, or equivalent substitutions can be made to some of the technical features. Such modifications or substitutions do not cause the essence of the corresponding technical solutions to deviate from the spirit and scope of the technical solutions of the embodiments of the present invention, and should all be included within the protection scope of the present invention.

Claims

1. A method for determining the temperature field of an oil-immersed transformer winding, characterized in that, include: Based on the target operating condition and the pre-established discrete point temperature model, the temperatures of multiple discrete points of the oil-immersed transformer winding under the target operating condition are determined; wherein, the discrete point temperature model is determined based on different types of sampling operating conditions and the multiple discrete point temperatures corresponding to the field under each sampling operating condition. Based on the multiple discrete point temperatures and the pre-established winding temperature field analysis model, the temperature field distribution of the oil-immersed transformer winding is determined; wherein, the winding temperature field analysis model is determined based on the temperature field distribution of the oil-immersed transformer under a first preset number of different types of sampling conditions, and the temperature of a second preset number of discrete points within the field under the target condition. Before determining the temperature field distribution of the oil-immersed transformer winding based on the multiple discrete point temperatures and the pre-established winding temperature field analysis model, the process also includes: Obtain the temperature field distribution under a first preset number of different types of sampling conditions; Based on the temperature field distribution under all sampling conditions, a snapshot matrix is ​​established; Based on the snapshot matrix, the reduced-order mode matrix and the mode coefficient matrix are determined; Select a second preset number of discrete points; Construct a discrete point temperature matrix based on the discrete point temperatures corresponding to the discrete points under the target operating conditions. The inversion matrix is ​​determined based on the discrete point temperature matrix, the reduced-order mode matrix, and the mode coefficient matrix. A winding temperature field analysis model is established based on the discrete point temperature matrix, the inversion matrix, and the reduced-order mode matrix.

2. The method for determining the winding temperature field of an oil-immersed transformer according to claim 1, characterized in that, Based on the snapshot matrix, the reduced-order mode matrix and mode coefficient matrix are determined, including: Singular value decomposition is performed on the snapshot matrix to obtain the singular value representation of the snapshot matrix; Based on the singular value representation of the snapshot matrix, an eigenorthogonal decomposition is performed to obtain the reduced-order modal representation of the snapshot matrix; Based on the reduced-order modal representation of the snapshot matrix, the reduced-order modal matrix and modal coefficient matrix are determined.

3. The method for determining the winding temperature field of an oil-immersed transformer according to claim 1, characterized in that, The inversion matrix is ​​determined based on the discrete point temperature matrix, the reduced-order mode matrix, and the mode coefficient matrix, including: Based on the discrete point temperature matrix, determine the node number of each discrete point; Based on the discrete point temperature matrix, the node number of each discrete point, and the modal coefficient matrix, a row vector with the same node number as the discrete point is selected from the reduced-order modal matrix to obtain the inversion matrix.

4. The method for determining the winding temperature field of an oil-immersed transformer according to any one of claims 1-3, characterized in that, Before determining the multiple discrete-point temperatures of the oil-immersed transformer windings under the target operating condition based on the target operating condition and a pre-established discrete-point temperature model, the process also includes: Obtain various sampling conditions and the corresponding discrete point temperatures within the field under each sampling condition; Based on multiple sampling conditions, the discrete point temperature corresponding to each sampling condition, and the response surface model, a discrete point temperature model between the sampling conditions and the discrete point temperature of the oil-immersed transformer winding is established.

5. The method for determining the winding temperature field of an oil-immersed transformer according to any one of claims 1-3, characterized in that, Before determining the temperature field distribution of the oil-immersed transformer winding based on the multiple discrete point temperatures and the pre-established winding temperature field analysis model, the process also includes: Under the target operating condition, the measured temperature of multiple discrete points of the oil-immersed transformer winding is obtained by using multiple sensors installed at the winding of the oil-immersed transformer. Determine the temperature difference between the measured temperature and the discrete point temperature for each discrete point; Determine whether the temperature difference is greater than a preset threshold; If the temperature difference is greater than the preset threshold, the discrete point temperature model is adjusted according to the measured temperature.

6. The method for determining the winding temperature field of an oil-immersed transformer according to claim 3, characterized in that, Based on the discrete point temperature matrix, the node number of each discrete point, and the modal coefficient matrix, a row vector with the same node number as the discrete point is selected from the reduced-order modal matrix to obtain the inversion matrix, including: According to the expression: Determine the inversion matrix; In the formula, Represents the discrete-point temperature matrix. In the inversion matrix, the first... Row vectors corresponding to discrete points Represents the modal coefficient matrix of the th The modal coefficients corresponding to each discrete point This represents the first discrete point in the reduced-order mode matrix. A row vector with node numbers, Indicates the first discrete points, This represents the total number of discrete points. Indicates the first Each node number Represents the first point in the discrete point temperature matrix. Row vectors corresponding to discrete points Indicates the first The node number corresponding to each discrete point. The reduced-order mode matrix represents the first... The node number and the first The node numbers of each discrete point correspond to the node numbers.

7. The method for determining the winding temperature field of an oil-immersed transformer according to any one of claims 1-3, characterized in that, The expression for the winding temperature field analysis model is as follows: ； In the formula, Represents a snapshot matrix. Represents the reduced-order mode matrix. Represents the inversion matrix. This represents the temperature matrix at discrete points.

8. The method for determining the winding temperature field of an oil-immersed transformer according to claim 4, characterized in that, The expression for the discrete-point temperature model is: ； In the formula, Indicates the first Temperature at discrete points This represents the parameter matrix corresponding to the sampling conditions of the winding. This represents the coefficient matrix of the polynomial.

9. A device for determining the temperature field of an oil-immersed transformer winding, characterized in that, include: The first determining module is used to determine multiple discrete point temperatures of the oil-immersed transformer winding under the target operating condition based on the target operating condition and a pre-established discrete point temperature model; wherein, the discrete point temperature model is determined based on different types of sampling operating conditions and multiple discrete point temperatures corresponding to the field under each sampling operating condition. The second determining module is used to determine the temperature field distribution of the oil-immersed transformer winding based on the multiple discrete point temperatures and a pre-established winding temperature field analysis model; wherein, the winding temperature field analysis model is determined based on the temperature field distribution of the oil-immersed transformer under a first preset number of different types of sampling conditions, and the second preset number of discrete point temperatures within the field under the target condition. The second determining module is specifically used for: Obtain the temperature field distribution under a first preset number of different types of sampling conditions; Based on the temperature field distribution under all sampling conditions, a snapshot matrix is ​​established; Based on the snapshot matrix, the reduced-order mode matrix and the mode coefficient matrix are determined; Select a second preset number of discrete points; Construct a discrete point temperature matrix based on the discrete point temperatures corresponding to the discrete points under the target operating conditions. The inversion matrix is ​​determined based on the discrete point temperature matrix, the reduced-order mode matrix, and the mode coefficient matrix. A winding temperature field analysis model is established based on the discrete point temperature matrix, the inversion matrix, and the reduced-order mode matrix.

Images