A transformer field-circuit coupling time domain simulation method and device

Abstract

The application discloses a transformer field-circuit coupling time domain simulation method and device, and relates to the technical field of power equipment simulation; the scheme is: a corresponding field simulation operation rule is constructed for a first operation included in transformer full-time domain simulation operation, and a corresponding circuit simulation operation rule is constructed for a second operation; the first operation is transformer two-dimensional electromagnetic field simulation operation, the second operation is transformer winding external circuit simulation operation, and the two types of operations provide simulation required excitation conditions and boundary conditions for each other; then, time step alternation iteration operation is carried out according to the two types of operation rules in each single simulation time step, the two operations are alternately executed, and the input of the operation executed later is determined according to the output of the operation executed earlier in the execution process; finally, all single simulation time steps covered by the full-time domain simulation period are sequentially traversed, the full-time domain cyclic iteration is checked, and after the overall simulation result meets the preset self-consistent standard, the transformer full-time domain transient simulation result is output.

Description

Technical Field

[0001] This application relates to the field of power equipment simulation technology, and in particular to a transformer field-circuit coupling time-domain simulation method and apparatus. Background Technology

[0002] Transformers experience typical transient electromagnetic processes under power system equipment switching and fault conditions, such as closing excitation inrush current, winding short-circuit faults, and lightning overvoltage impacts. The electrical characteristics of these transient conditions are jointly determined by the dynamic evolution of the electromagnetic field inside the transformer core and the circuit parameters of the external power supply circuit. Therefore, conducting accurate transformer field-circuit coupling time-domain simulation is a key technical means to evaluate transformer dynamic performance, design relay protection strategies, and conduct fault mechanism analysis.

[0003] Currently, the mainstream solution for transformer field-circuit coupling simulation in the industry adopts a heterogeneous co-simulation architecture consisting of the Finite Element Method (FEM) and the Simulation Program with Integrated Circuit Emphasis (SPICE). Data is exchanged between the two independent simulation tools to complete the field-circuit coupling solution calculation.

[0004] However, existing FEM+SPICE heterogeneous co-simulation schemes suffer from inherent drawbacks such as cumbersome interface interactions, mismatched solution time steps, and difficulty in converging numerical oscillations. Limited by these shortcomings, such schemes struggle to balance computational efficiency, simulation stability, and model universality, and cannot efficiently adapt to the full-time-domain simulation requirements of various transient conditions such as transformer inrush current, equipment short circuits, and lightning strikes. Therefore, overcoming the inherent shortcomings of traditional FEM+SPICE co-simulation and providing a technical solution capable of achieving full-time-domain field-circuit coupling simulation of transformers is a pressing technical challenge in this field. Summary of the Invention

[0005] This application provides a transformer field-circuit coupling time-domain simulation method and apparatus. The main purpose is to provide a technical solution that can realize the full-process field-circuit coupling simulation of transformers in the entire time domain, and effectively improve the fit and calculation accuracy of the field-circuit coupling simulation results under transient conditions.

[0006] To achieve the above objectives, this application mainly provides the following technical solutions: The first aspect of this application provides a time-domain simulation method for transformer field-circuit coupling, the method comprising: For the first task included in the full-time-domain transformer simulation, corresponding field simulation rules are established, and for the second task included in the full-time-domain transformer simulation, corresponding circuit simulation rules are established. The first task is a two-dimensional electromagnetic field simulation of the transformer, used to collect relevant physical quantities of the spatial magnetic field under field simulation to provide excitation basis for circuit simulation. The second task is an external circuit simulation of the transformer winding, used to collect winding circuit parameters under circuit simulation to provide boundary input for field simulation. Within a single simulation time step, time-step alternating iterative operations are performed according to the field simulation operation rules and the circuit simulation operation rules. The time-step alternating iterative operations are performed alternately by executing the first operation and the second operation, and the input of the operation to be executed later is determined based on the output obtained from the operation executed first. The first simulation operation uses the initial input. For each of the other individual simulation time steps included in the full-time simulation cycle, the time step alternation iteration operation corresponding to each individual simulation time step is traversed and executed in sequence until the simulation result corresponding to the full-time simulation cycle meets the preset self-consistency standard, and the full-time transient simulation result of the transformer is output.

[0007] A second aspect of this application provides a transformer field-circuit coupling time-domain simulation device, the device comprising: The construction unit is used to establish corresponding field simulation operation rules for the first operation included in the full-time domain simulation of the transformer, and to establish corresponding circuit simulation operation rules for the second operation included in the full-time domain simulation of the transformer; wherein, the first operation is a two-dimensional electromagnetic field simulation of the transformer, used to collect relevant physical quantities of the spatial magnetic field under field simulation to provide excitation basis for circuit simulation; the second operation is an external circuit simulation of the transformer winding, used to collect winding circuit parameters under the circuit simulation to provide boundary input for the field simulation; The time-step processing unit is used to perform time-step alternating iterative operations according to the field simulation operation rules and the circuit simulation operation rules within a single simulation time step. The time-step alternating iterative operations are to alternately execute the first operation and the second operation, and determine the input of the operation to be executed later based on the output obtained from the operation executed first. The first simulation operation uses the initial input. The full-time-domain processing unit is used to sequentially traverse and execute the time-step alternation iterative operation corresponding to each of the other individual simulation time steps included in the full-time-domain simulation cycle, until the simulation result corresponding to the full-time-domain simulation cycle meets the preset self-consistency standard, and output the full-time-domain transient simulation result of the transformer.

[0008] A third aspect of this application provides a computer-readable storage medium storing a computer program that, when executed by a processor, implements the transformer field-circuit coupling time-domain simulation method as described above.

[0009] A fourth aspect of this application provides an electronic device, the device including at least one processor, and at least one memory and bus connected to the processor; wherein the processor and the memory communicate with each other through the bus; the processor is used to call program instructions in the memory to execute the transformer field-circuit coupling time-domain simulation method as described above.

[0010] By employing the above-described technical solution, the technical solution provided in this application has at least the following advantages: This application provides a method and apparatus for time-domain simulation of transformer field-circuit coupling: First, corresponding field simulation rules are constructed for the first task included in the full-time-domain simulation of the transformer, and corresponding circuit simulation rules are constructed for the second task; wherein, the first task is a two-dimensional electromagnetic field simulation of the transformer, and the second task is an external circuit simulation of the transformer winding. The two types of tasks provide each other with the excitation conditions and boundary conditions required for simulation; then, within each individual simulation time step, time-step alternating iterative operations are carried out according to the two types of operation rules, alternating between the first and second tasks. During the execution, the input of the subsequent task is determined based on the output obtained from the first task, and the first simulation task uses the initial input; finally, all individual simulation time steps covered by the full-time-domain simulation cycle are traversed sequentially according to the time sequence. After full-time-domain iterative verification, once the overall simulation result meets the preset self-consistency standard, the full-time-domain transient simulation result of the transformer is output.

[0011] Compared to the existing FEM+SPICE heterogeneous co-simulation scheme, which has the technical shortcomings of being difficult to efficiently adapt to the full-time-domain simulation of transformer transient conditions, the existing scheme relies on two independent solution software tools to carry out coupled calculations. There are problems such as cumbersome interaction interfaces, mismatched simulation time steps, and difficulty in numerical oscillation convergence between heterogeneous solvers. It is difficult to balance computational efficiency, simulation stability and model universality. It can only achieve loose field-circuit coupling solution and cannot guarantee the overall physical self-consistency of field-circuit parameters under the full-time-domain period. The simulation of severe transient conditions such as short circuits and impacts is obviously distorted.

[0012] To address the aforementioned shortcomings, this application abandons the heterogeneous approach of using two independent software tools in conjunction with a novel integrated simulation architecture. Under this unified architecture, two sets of supporting operational rules are established: one for the field and one for the circuit. Instantaneous operating condition coupling convergence is achieved through alternating closed-loop iterations of the field and circuit within a single time step. This is further superimposed with full-time-domain self-consistent verification to form a hierarchical two-level iterative architecture. Utilizing this full-time-domain step-by-step iterative closed-loop logic, the actual operation process of the transformer throughout its entire lifecycle can be accurately replicated, fully reproducing the changes in circuit parameters and magnetic fields under various transient conditions such as short circuits and faults. Therefore, this application effectively overcomes the inherent drawbacks of the FEM+SPICE heterogeneous co-simulation scheme, providing a technical solution capable of achieving full-time-domain field-circuit coupling simulation of the transformer throughout its entire process, effectively improving the consistency and calculation accuracy of the field-circuit coupling simulation results under transient conditions.

[0013] The above description is only an overview of the technical solution of this application. In order to better understand the technical means of this application and to implement it in accordance with the contents of the specification, and to make the above and other objects, features and advantages of this application more obvious and understandable, the following are specific embodiments of this application. Attached Figure Description

[0014] Various other advantages and benefits will become apparent to those skilled in the art upon reading the following detailed description of preferred embodiments. The accompanying drawings are for illustrative purposes only and are not intended to limit the scope of this application. Furthermore, the same reference numerals denote the same parts throughout the drawings. In the drawings: Figure 1 A flowchart of a transformer field-circuit coupling time-domain simulation method provided in this application embodiment; Figure 2 A flowchart illustrating the configuration of field simulation operation rules provided in this application embodiment; Figure 3 A flowchart of the configuration circuit simulation operation rules provided in the embodiments of this application; Figure 4 A flowchart of the time-step alternating iterative calculation method provided in the embodiments of this application; Figure 5 A flowchart for full-time domain iterative iteration to self-consistency provided in the embodiments of this application; Figure 6 A block diagram of a transformer field-circuit coupling time-domain simulation device provided in this application embodiment; Figure 7 A block diagram of another transformer field-circuit coupling time-domain simulation device provided in this application embodiment. Detailed Implementation

[0015] Exemplary embodiments of the present application will now be described in more detail with reference to the accompanying drawings. While exemplary embodiments of the present application are shown in the drawings, it should be understood that the present application may be implemented in various forms and should not be limited to the embodiments set forth herein. Rather, these embodiments are provided so that this application will be thorough and complete, and will fully convey the scope of the present application to those skilled in the art.

[0016] Existing FEM+SPICE heterogeneous co-simulation schemes suffer from inherent drawbacks such as cumbersome interface interactions, mismatched solution time steps, and difficulty in converging numerical oscillations. Detailed explanations are as follows: (1) The interface interaction is complicated. FEM finite element electromagnetic field simulation software and SPICE circuit simulation belong to two independent and heterogeneous simulation programs. Their underlying mathematical architecture and data storage formats are incompatible. The simulation process requires frequent construction of dedicated data interface and manual data format conversion and parameter mapping. There are many data transfer links between programs, the operation process is complicated, and it is easy to introduce additional data errors in the data transfer process.

[0017] (2) The drawbacks of mismatched time steps: FEM electromagnetic field simulation is to finely divide the spatial field, and usually uses a very small simulation time step; SPICE circuit simulation builds the circuit model based on lumped parameter components, and selects a larger simulation time step. The original time step scales of the two types of simulations cannot be naturally unified; forced alignment of time steps either sacrifices the circuit simulation calculation efficiency or causes the electromagnetic field simulation calculation accuracy to be lost, making it difficult to achieve synchronous operation of field and electrical timing.

[0018] (3) The drawbacks of numerical oscillation and difficulty in convergence: When the two types of software perform step-by-step serial alternating calculations, the field output flux / induced voltage and circuit feedback loop parameters are transmitted unidirectionally across the program, lacking a full-time closed-loop self-consistent correction mechanism; affected by the nonlinear characteristics of transformer core magnetization, the single-time step calculation error continues to accumulate along the time axis, inducing full-domain numerical oscillation, making it difficult to converge the iteration and resulting in poor simulation stability.

[0019] Constrained by the three inherent defects mentioned above, the existing FEM+SPICE heterogeneous co-simulation scheme cannot simultaneously achieve computational efficiency, simulation stability, and model universality. For transient conditions such as inrush current, short-circuit faults, and lightning strikes, it is difficult to implement long-term continuous simulation in the entire time domain.

[0020] Based on the numerous technical shortcomings of existing FEM+SPICE heterogeneous co-simulation schemes, this application's embodiments abandon the traditional cross-software heterogeneous coupling architecture and completely replace it with a brand-new integrated simulation architecture, achieving the following: Figure 1 The transformer field-circuit coupling time-domain simulation method shown in this application provides the following specific steps 100-300: Step 100: Establish corresponding field simulation operation rules for the first task included in the full-time domain simulation of the transformer, and establish corresponding circuit simulation operation rules for the second task included in the full-time domain simulation of the transformer; wherein, the first task is the two-dimensional electromagnetic field simulation of the transformer, which is used to collect relevant physical quantities of the spatial magnetic field under field simulation to provide excitation basis for circuit simulation; the second task is the external circuit simulation of the transformer winding, which is used to collect winding circuit parameters under circuit simulation to provide boundary input for field simulation.

[0021] Traditional transformer field-circuit coupling generally uses electromagnetic field finite element software and circuit simulation software for joint simulation. Dedicated data interaction interfaces need to be built between different simulation programs. The simulation time step and convergence judgment criteria used by the two types of simulation operations cannot be unified. Frequent data interaction not only increases the simulation computation overhead, but also easily causes numerical oscillation and iteration non-convergence.

[0022] Based on the differences in physical mechanisms and the mathematical forms of control equations, this application's embodiments divide the overall simulation task into two independent computational dimensions: a first task and a second task. The first task is a two-dimensional electromagnetic field simulation of a transformer, with the corresponding field quantities following the time-domain diffusion equations in partial differential form. The second task is a simulation of the external circuit of the transformer windings, with the corresponding circuit quantities following Kirchhoff's equations in ordinary differential form. It should be noted that in this application's embodiments, the terms "first task" and "second task" are merely identifiers to distinguish between the two types of simulation tasks; they do not imply any order of execution or hierarchy, and there is no ambiguity regarding their sequence.

[0023] By pre-decomposing the simulation dimensions, the two types of physical problems are decoupled and separated, eliminating the need for complex interface connections between heterogeneous simulation programs. This reduces data exchange complexity at the mechanistic level, improves the computational efficiency and numerical stability of subsequent simulations, and provides a basis for establishing two sets of simulation operation rules in subsequent steps. The first and second tasks can mutually supply the data required for simulation. For example, the first task collects relevant physical quantities of the spatial magnetic field under field simulation, providing excitation basis for the circuit simulation of the second task; the second task collects winding circuit parameters under circuit simulation, providing boundary input for the field simulation of the first task.

[0024] Furthermore, leveraging the inherent mathematical model characteristics of the first and second tasks derived from the decomposition, corresponding simulation task operation rules are established for each. The governing equations for the two tasks belong to partial differential equations and ordinary differential equations, respectively, and their operational mechanisms and discretization methods are fundamentally different, making it impossible to share the same set of operational logic. This application's embodiments customize exclusive operation rules for each type of task, capable of matching the operational characteristics of the corresponding equations.

[0025] Specifically, the first assignment uses the algorithm logic of domain decomposition combined with physical constraint neural networks to formulate field operation rules and realize continuous constraints on the boundaries of multiple physical subdomains; the second assignment uses the numerical integration step algorithm to formulate loop operation rules and realize quantitative calculation from input voltage to winding current.

[0026] This application embodiment achieves the independence of two computing systems, field and circuit, by setting up separate dedicated computing rules. This provides a standardized computing benchmark for the data interaction and iteration of the two subsequent simulation operations, and avoids the problems of parameter adaptation difficulties and numerical instability caused by traditional heterogeneous joint computing from the algorithm level.

[0027] Step 200: Within a single simulation time step, execute time-step alternating iterative operations according to the field simulation operation rules and the circuit simulation operation rules. The time-step alternating iterative operations alternately execute the first operation and the second operation, and determine the input of the operation to be executed later based on the output obtained from the operation executed first. The first simulation operation uses the initial input.

[0028] This application embodiment performs timing-coupled operations based on the field simulation operation rules corresponding to the first operation and the circuit simulation operation rules corresponding to the second operation, using an alternating time-step iterative operation to complete the closed-loop interaction of field and circuit data. This alternating time-step iterative operation runs within a single simulation time step, alternately implementing the first and second operations. During the operation, the preceding and following operations form a data transfer relationship, using the output of the preceding operation as the input condition for the following operation. The entire operation process uses the initial input configured in the first simulation operation as the starting point.

[0029] For example, multiple rounds of alternating field-circuit closed-loop operations need to be executed within a single simulation time step. The process of a single round of alternating field-circuit closed-loop operation includes: calculating the winding current from the induced electromotive force using the circuit simulation rules corresponding to the second task, and substituting the obtained current as the field excitation source into the calculation of the first task; then calculating the global magnetic field distribution using the field simulation rules corresponding to the first task and inversely deducing the winding induced electromotive force, and feeding the electromotive force data back to the second task side as the input for the next round of alternating field-circuit closed-loop operation. This alternating execution process is repeated continuously, updating the physical parameters on both sides of the field and circuit until the convergence condition is met, thus completing the solution for the current single simulation time step.

[0030] Step 300: For other individual simulation time steps included in the full-time simulation cycle, traverse and execute the time step alternation iteration operation corresponding to each individual simulation time step in sequence until the simulation result corresponding to the full-time simulation cycle meets the preset self-consistency standard, and output the full-time transient simulation result of the transformer.

[0031] The sequential traversal and execution of the time step alternation iteration operation corresponding to each individual simulation time step refers to the sequential traversal and execution of the time step alternation iteration operation corresponding to each individual simulation time step along the simulation time axis according to the time sequence of each individual simulation time step, so as to realize the orderly deduction of the transformer transient condition along the time process.

[0032] This step achieves full-time domain cyclic iteration to self-consistency. Through a two-layer architecture of full-time domain step-by-step traversal and superimposed full-cycle full-domain self-consistent iterative verification, it fully reproduces the transient evolution process of the transformer's entire operating cycle in a time sequence. Based on the full-cycle overall self-consistent correction mechanism, it offsets the cumulative deviation of field and electrical parameters caused by single-step serial push, accurately replicating the real physical operating characteristics of the transformer under transient conditions such as short circuits and faults. It breaks away from the limitations of traditional segmented truncation simulation, ensures full-time domain, full-field domain, and full-circuit parameter coupling self-consistency, and further improves the accuracy of the full-time domain transient simulation results in matching the actual operating conditions.

[0033] The preset self-consistency standard provided in this application includes two layers of judgment: first, the temporal variation law of field physical quantities such as winding flux linkage and induced electromotive force at each moment in the whole time domain converges and stabilizes; second, the temporal variation law of circuit parameters such as loop current and port voltage in the whole time domain converges and stabilizes. When the change amplitude of field data and circuit parameters in the whole time domain with each iteration is lower than the preset convergence threshold, and the bidirectional coupling data of field and circuit match each other and satisfy the physical constraints corresponding to the coupling of electromagnetic field and circuit, it is determined that the preset self-consistency standard has been reached.

[0034] It should be noted that steps 200 and 300 constitute a two-level iterative architecture with an outer large iteration nested within an inner small iteration. Step 200 corresponds to the inner small iteration, which, within a single simulation time step, performs alternating time-step iterations according to the field simulation rules and circuit simulation rules. Based on the interaction between field simulation and circuit simulation, it conducts multiple rounds of field-circuit alternating closed-loop operations corresponding to the current instantaneous operating condition, achieving convergence of the solution within a single simulation time step. Step 300 corresponds to the outer large iteration, which sequentially traverses all individual simulation time steps within the full-time-domain simulation cycle, executing the alternating time-step iterations corresponding to each individual simulation time step in turn, completing one round of full-time-domain simulation calculation. If the result of one round of full-time-domain simulation does not meet the preset self-consistency standard, all individual simulation time steps are traversed again, and the entire full-time-domain calculation is performed again until the simulation result meets the preset self-consistency standard, at which point the full-time-domain transient simulation result of the transformer is output.

[0035] Based on steps 100-300 above, the transformer field-circuit coupling time-domain simulation method provided in this application completely replaces the original integrated simulation architecture. By setting up two types of matching operation rules under a unified architecture—field and circuit—it completes alternating closed-loop iteration of field and circuit within a single time step. Then, it superimposes a layered iterative architecture with full-time-domain self-consistent verification. This effectively overcomes the inherent drawbacks of the FEM+SPICE heterogeneous joint simulation scheme (e.g., cumbersome heterogeneous tool interaction interfaces, mismatched simulation time steps, difficulty in numerical oscillation convergence, and difficulty in balancing computational efficiency, simulation stability, and model universality). This application embodiment utilizes full-time-domain step-by-step iterative closed-loop logic to accurately replicate the actual operation process of the transformer throughout its entire lifecycle, completely reproducing the circuit parameters and magnetic field changes under various transient conditions such as short circuits and faults. Therefore, it provides a technical solution that can realize full-time-domain field-circuit coupling simulation of the transformer, effectively improving the fit and calculation accuracy of the field-circuit coupling simulation results under transient conditions.

[0036] Furthermore, after completing the solution output of the transformer full-time transient simulation results (as described in steps 100-300 above), the embodiments of this application can further conduct engineering application analysis based on high-precision, fully self-consistent simulation data, including the following steps (1)-(2), used to extract and organize the simulation results and conduct working condition analysis and engineering evaluation: (1) Extract the field timing data and circuit timing data corresponding to the full-time-domain simulation cycle from the full-time-domain transient simulation results of the transformer, and construct the transformer transient operating condition feature dataset.

[0037] The field time-series data consists of temporal information on spatial magnetic field-related physical quantities recorded sequentially along the simulation time axis, including spatial field quantities such as magnetic field distribution and flux density in the transformer core and winding regions at different times. The circuit time-series data consists of temporal information on winding circuit electrical parameters recorded sequentially along the simulation time axis, including electrical parameters such as winding induced voltage and circuit current at different times. The two are bidirectionally coupled: the field time-series data is obtained from the field simulation, converted into winding induced voltage through flux integration, and serves as the excitation input for the circuit simulation, directly affecting the evolution of the circuit time-series data; the corresponding circuit electrical parameters are then fed back to the field solution as boundary conditions, changing the spatial magnetic field distribution and thus updating the field time-series data. The two sets of time-series data correspond one-to-one and evolve synchronously over time, jointly and completely characterizing the transient electromagnetic process of transformer field-circuit coupling.

[0038] This application's embodiments, based on high-precision, globally self-consistent physical simulation results obtained through two-level nested iterative calculations, comprehensively extract continuous time-series information such as magnetic field distribution, winding flux linkage, induced voltage, and loop current across the entire time domain, and uniformly organize it into a standardized transient operating condition feature dataset. Compared to existing FEM+SPICE heterogeneous co-simulation schemes, traditional heterogeneous co-simulation is limited by solver data interaction formats and loose coupling mechanisms, making it easy for time-series data to become discontinuous and offset, and difficult to obtain a complete set of time-series data that meets global self-consistent constraints. The data collected in this step can completely carry electromagnetic evolution information under various transient operating conditions such as closing excitation inrush current, winding short-circuit faults, and lightning overvoltage impacts, ensuring the consistency of the physical logic within the dataset and providing a reliable and continuous data foundation for subsequent quantitative analysis of operating conditions.

[0039] (2) Analyze the dynamic operating characteristics of the transformer based on the transient operating condition feature dataset, generate transient operating condition evaluation conclusions, and store and visualize the simulation results and evaluation conclusions for use in transformer insulation verification, structural optimization or fault protection scheme verification.

[0040] Based on the constructed transient operating condition feature dataset, quantitative analysis is conducted, and the embodiments of this application can fully leverage the advantages of the obtained high-precision full-time-domain transient simulation results. Existing FEM+SPICE heterogeneous co-simulation has large simulation deviations under severe transient conditions such as short circuits and voltage surges, making it difficult to accurately characterize the dynamic evolution process of the bidirectional coupling between the field and the circuit.

[0041] The method and flow of this application embodiment (as described in steps 100-300 above) can accurately reproduce the dynamic operation process of a transformer across the entire time domain. Taking the inrush current condition during closing as a typical application scenario, it can completely reproduce the full-time evolution process of asymmetric inrush current caused by core deep saturation during the no-load or load closing process of the transformer, providing precise support for the design of inrush current suppression schemes and the setting of relay protection devices. Simultaneously, it can be extended to simulation analysis of various typical transient conditions such as winding short-circuit faults and lightning overvoltage impacts, as well as transformer fault retrospection and accident inversion scenarios. It can reliably conduct transient electromagnetic stress assessment of windings and local field strength verification of insulation, guiding the optimization of transformer insulation structure and winding mechanical structure. This provides a reliable basis for the study of power transformer operating characteristics, equipment optimization design, and fault mechanism analysis, avoiding engineering problems such as conservative design and unreasonable protection strategy settings caused by simulation distortion. Simulation data and evaluation conclusions are uniformly stored and visualized, facilitating designers to intuitively assess operating conditions.

[0042] Before executing steps 100-300 in this embodiment of the application, it is necessary to complete the construction division of the first job and the second job in advance. The specific construction operations include the following (A1) and (A2): (A1) For the field simulation task corresponding to the full time domain simulation of the transformer, establish the control equation represented by the time domain magnetic vector potential; construct the two-dimensional electromagnetic field simulation task of the transformer based on the control equation, as the first task.

[0043] Based on the differences between the physical mechanism and the mathematical form of the control equation, the full-time domain simulation of the transformer is divided into field simulation and circuit simulation.

[0044] Specifically, control equations corresponding to the time-domain magnetic vector potential representation are established for the field simulation task. For example, the physical laws of the two-dimensional electromagnetic field are represented by the time-domain magnetic vector potential diffusion equation, and the control equation of the two-dimensional electromagnetic field is obtained, which is expressed as the following formula (1): Formula (1); In formula (1): x and y Represents spatial coordinates, t Represents the time coordinate; μ ( x , y () is the permeability related to spatial location. σ ( x, y () is the conductivity related to spatial location. J z ( x, y, t ) is the current density source term. A z The potential is a scalar magnetic vector; the field boundary adopts a homogeneous Dirichlet boundary condition. A z | Ω = 0, initial condition is A z ( t =0) = 0.

[0045] Based on the control equations corresponding to the field simulation task, the two-dimensional electromagnetic field simulation task of the transformer is defined in the full-time domain simulation task of the transformer. When the field modeling is carried out in the future, the two-dimensional electromagnetic field region can be divided into four sub-domains: left iron core, right iron core, winding, and air. The operation rules of the first task are built according to the physical constraint neural network of the domain decomposition form.

[0046] (A2) Based on Kirchhoff's voltage law, for the circuit simulation task corresponding to the full-time domain simulation of the transformer, build the ordinary differential control equations corresponding to the primary winding and the secondary winding respectively; construct the external circuit simulation task of the transformer winding based on the ordinary differential control equations, as the second task.

[0047] Based on Kirchhoff's voltage law, ordinary differential control equations are established for the primary and secondary windings respectively for the circuit simulation task. For example, the external circuit corresponding to the winding is modeled according to Kirchhoff's voltage law, and ordinary differential control equations are established for the primary and secondary windings respectively. Specifically, the circuit equation of the primary winding is expressed as formula (2) and the circuit equation of the secondary winding is expressed as formula (3). Formula (2); Formula (3); In the two formulas above: t Represents time; This is the input voltage of the primary winding; i 1 (t) and These represent the current and resistance of the primary winding, respectively. For the flux linkage of the primary winding; E sec (t) The induced electromotive force of the secondary winding; i 2 (t) , R 0 These represent the current and resistance of the secondary winding, respectively. β The coefficient characterizing the nonlinear loss characteristics of the secondary winding.

[0048] This application embodiment completes the pre-construction of two independent simulation tasks based on the above (A1) and (A2): First, it establishes the control equations representing the time-domain magnetic vector potential for the field simulation task, and uses these equations to construct the transformer two-dimensional electromagnetic field simulation task as the first task. Second, it constructs the ordinary differential control equations of the primary and secondary windings based on Kirchhoff's voltage law, and uses these equations to construct the winding external circuit simulation task as the second task. This method decouples and divides tasks based on the differences in the mathematical models corresponding to the field and circuit physical problems, with each type of task possessing an independent and complete mathematical model system. This pre-construction scheme eliminates the need to build complex interactive interfaces between heterogeneous simulation programs, reducing data exchange complexity from a mechanistic perspective. It effectively avoids the defects of traditional FEM+SPICE co-simulation, such as cumbersome interfaces, time-step mismatch, and difficulty in converging numerical oscillations. It provides a standardized computational carrier for subsequent hierarchical iterative field-circuit coupling simulation, ensuring the orderly bidirectional interaction of field-circuit data and improving the overall simulation's computational efficiency and numerical stability.

[0049] Furthermore, for the first and second tasks mentioned above, two independent simulation task rules are constructed for each task, specifically executed using the following two branches (B1) and (B2): (B1) For the first task, the field simulation task rules corresponding to the first task are obtained by dividing the physical subdomains into multiple types and using loss constraints and subdomain interface continuity constraints to construct the corresponding field operation rules.

[0050] (B2) For the second task, the circuit simulation task rules corresponding to the second task are obtained by using the induced voltage as the input parameter, combining the circuit equation with the numerical integration algorithm to construct the corresponding circuit operation rules.

[0051] For the two branches (B1) and (B2) mentioned above, which respectively construct field and circuit simulation operation rules, this application embodiment implements the regularized and standardized definition of the two types of simulations. (B1) Based on subdomain partitioning and physical constraints, the operation rules for field simulation are constructed to ensure the accuracy and convergence of magnetic field solution; (B2) Based on circuit equations and numerical integration algorithms, the operation rules for circuit simulation are constructed to ensure the stability of current waveform solution. This regularized construction method clarifies the input and output interfaces and operation logic on both sides of the field and circuit, eliminates the coupling between heterogeneous simulation modules, and provides a standardized and interactive operation carrier for the subsequent round of field-circuit alternating closed-loop operation, effectively improving the overall efficiency and numerical stability of transformer full-time domain simulation.

[0052] Specifically, in one optional embodiment, the operation of constructing the field simulation operation rules corresponding to the first task for the above-mentioned branch (B1) is as follows: Figure 2 As shown, the specific implementation is carried out using the following steps 110-112.

[0053] Step 110: Construct a basic field simulation model based on the physical constraint neural network of domain decomposition. The physical constraint neural network is the computational carrier for subdomain partitioning and constraint application.

[0054] A field-based simulation model is built using a physical constraint neural network for domain decomposition. This neural network serves as the computational carrier for subdomain partitioning and physical constraint application. The traditional finite element mesh generation architecture is abandoned, and the physical constraint neural network is used to embed constraints into the electromagnetic field equations. This provides a unified underlying computational framework for subsequent subdomain partitioning and parameter configuration, thus avoiding simulation errors caused by the cumbersome traditional FEM mesh generation and mesh distortion from the model's bottom layer.

[0055] Step 111: According to the physical structure of the transformer, the two-dimensional electromagnetic field region of the transformer is divided into four physical subdomains: the left iron proton domain, the right iron proton domain, the winding subdomain, and the peripheral air subdomain.

[0056] The two-dimensional electromagnetic field region is divided into four categories based on the actual physical structure of the transformer: the left iron proton domain, the right iron proton domain, the winding subdomain, and the peripheral air subdomain. The electromagnetic properties of different regions are processed in blocks to match the electromagnetic distribution characteristics of the transformer core, winding, and air domains, so as to achieve a one-to-one correspondence between the physical space and the simulation subdomain, which facilitates the subsequent accurate configuration of dielectric parameters and boundary constraints.

[0057] Step 112: Configure the corresponding medium property parameters within each type of physical subdomain, set continuous constraints at the interface between adjacent subdomains, and combine global loss constraints to complete the construction of field operation rules.

[0058] Within each subdomain, corresponding medium property parameters are configured, and continuous field constraints are added at the interface between adjacent subdomains. These are superimposed with global loss constraints to ultimately form complete field operation rules. This ensures that the electromagnetic properties of the medium in the same subdomain conform to the physical parameters, and also ensures the continuous connection of magnetic field physical quantities at the interface between adjacent subdomains. By using multiple physical constraints to lock the evolution law of the electromagnetic field, the physical fit and calculation accuracy of the field simulation are improved.

[0059] The above steps 110-112 are illustrated with examples. The field operation rules for the first task are as follows: the field operation is implemented by using a domain decomposition combined with a physical constraint neural network architecture. The global electromagnetic field is divided into four physical subdomains: left iron core, right iron core, winding and air. The global field loss function is used to constrain the accuracy of the global model operation. The expression of the field loss function is as follows (4): Formula (4); In formula (4): L pde L is the residual loss term in the physical equation. iface L is the loss term for the continuity condition at the subdomain interface. bc For the boundary condition loss term, L ic This is the initial condition loss term; , , , These are the weight hyperparameters corresponding to each loss, and all are preset values.

[0060] Furthermore, the electromagnetic parameters at the subdomain interface are guaranteed by two physical constraints: one is the continuity condition of the scalar magnetic vector potential, corresponding to the following mathematical constraint formula (5); the other is the continuity condition of the tangential component of the magnetic field strength, corresponding to the following mathematical constraint formula (6).

[0061] Formula (5); Formula (6); In formulas (5) and (6): subscripts 1 and 2 represent two adjacent different material subdomains, A z1 A z2 These are the scalar magnetic vector bits corresponding to the adjacent subdomains. 、 These are the magnetic permeabilities of the materials in the two subdomains, respectively. Here, is a partial differential operator; n is the normal direction of the subdomain interface. Based on the aforementioned loss function and interface constraints, the calculation rules for the entire field are defined.

[0062] Steps 110-112 above complete the construction of the field operation rules corresponding to the first task. A physical constraint neural network in the form of domain decomposition is used to build the underlying simulation framework, abandoning the traditional finite element mesh generation architecture and avoiding simulation errors caused by cumbersome mesh generation and mesh distortion. Four physical subdomains are divided according to the transformer's physical structure, achieving a one-to-one correspondence between the physical space and the simulation subdomains, facilitating precise configuration of dielectric property parameters within each domain. Complete field operation rules are built through the configuration of physical property parameters within subdomains and the superposition of continuous constraints on field quantities at subdomain interfaces with global loss constraints. The smooth transition of electromagnetic parameters between subdomains is ensured by relying on two types of interface continuity conditions: scalar magnetic vector potential and tangential component of magnetic field strength. The entire scheme locks in the electromagnetic field evolution law based on multiple physical constraints, improving the fit between field simulation and actual operating conditions and the accuracy of calculations. It provides a high-precision field solution carrier for subsequent field-circuit alternating closed-loop iterations, ensuring reliable excitation data output from the field to the circuit side.

[0063] Furthermore, in some preferred embodiments, an electromagnetic field simulation mechanism with domain-specific differentiated constraint adaptation capabilities can be provided. Specifically, this includes configuring differentiated physical constraint conditions for different physical subdomains of the transformer based on a physical constraint neural network in the form of domain decomposition. For example, combining the strong electromagnetic nonlinearity of the core subdomain, the lumped electromagnetic characteristics of the winding subdomain, and the weak field gradient characteristics of the air subdomain, corresponding electromagnetic constitutive relations and computational constraints are matched in each subdomain respectively; simultaneously, continuous boundary constraints of field quantities are uniformly applied at the interfaces of adjacent subdomains, taking into account both the electromagnetic operation law of a single subdomain and the continuity of field quantity transmission across the entire region. During the iterative calculation of a single simulation time step, the field calculation results are dynamically corrected based on the domain-specific differentiated constraints to adapt to the electromagnetic distribution characteristics of different structural regions of the transformer.

[0064] Unlike simulation modes that use uniform constraints across the entire domain, this mechanism can accurately adapt to the differences in electromagnetic properties in different regions, weaken the calculation deviations caused by the nonlinear effects of the iron core, further improve the matching degree between the field simulation results and the electromagnetic properties of the actual object, and at the same time reduce the convergence resistance caused by a single constraint rule to the global operation.

[0065] Specifically, in one optional embodiment, the operation of constructing the circuit simulation task rule corresponding to the second task for the above branch (B2) is as follows: Figure 3 As shown, this is specifically achieved using the following steps 120-122.

[0066] Step 120: Based on Kirchhoff's voltage law, establish the ordinary differential equations for the primary winding and the secondary winding respectively.

[0067] Based on Kirchhoff's voltage law, the ordinary differential equations of the circuit corresponding to the primary winding and the secondary winding are constructed. The mathematical relationship between the winding port voltage and the electrical quantity of the circuit is established from the level of the basic physical law of the circuit. This establishes the underlying mathematical benchmark for subsequent circuit simulation and provides the theoretical basis for circuit operation rules.

[0068] Step 121: Receive the induced voltage waveform of the winding as the input parameter for calculation, and retrieve the known primary winding input voltage. Based on the ordinary differential equations of the circuits corresponding to the primary winding and the secondary winding, construct a set of ordinary differential equations containing the primary and secondary winding currents.

[0069] Using the winding induced voltage and the external input voltage of the winding output from the field simulation as known input quantities, the differential equations of each winding are combined to form a complete set of equations; the data transfer of the electrical quantities output from the field to the circuit simulation is realized, the field and circuit data parameters are connected, and a complete mathematical relationship for solving the winding current from the voltage is constructed.

[0070] Step 122: Using the numerical integration method, the system of ordinary differential equations obtained by the simultaneous construction is calculated step by step along the simulation time to obtain the primary winding current waveform and secondary winding current waveform in the full time domain, so as to complete the circuit operation rules.

[0071] The numerical integration method is used to discretely solve the equation set along the simulation time step, and the waveforms of the primary and secondary winding current changes in the full time domain are obtained step by step, and finally the entire circuit operation rules are formed; the accurate solution of the full time sequence loop current is realized, and the obtained winding current can be used as the external excitation of the field simulation in reverse, supporting the bidirectional alternating iterative operation of the field and circuit.

[0072] The above steps 120-122 are illustrated with examples. For the second task, the circuit operation rules are constructed as follows: the winding induced voltage waveform output by the field simulation is used as the circuit operation input. The external circuit forms the primary winding circuit equation and the secondary winding circuit equation according to Kirchhoff's voltage law. The operation includes: (1) receiving the given winding induced voltage waveform and obtaining the known primary winding input voltage; (2) establishing a set of ordinary differential equations that simultaneously contain the primary winding current and the secondary winding current according to the primary winding circuit equation and the secondary winding circuit equation; (3) then using the numerical integration algorithm to discretize and solve step by step along the time axis, and obtain the changing waveforms of the primary winding current and the secondary winding current in the full time domain, and completely reproduce the electrical dynamic change law of the winding circuit. The scheme formed by the above steps 120-122 standardizes the data transmission link from the field to the circuit, realizes the orderly connection of electrical quantities, and the winding current obtained by the solution can be supplied to the field simulation as an external source excitation to support the bidirectional alternating iterative operation of the field and circuit.

[0073] The two sets of simulation rules define the modeling methods, constraints, and numerical calculation methods for the field and circuit, respectively. The field simulation operation rules and circuit simulation operation rules are determined as the execution criteria to be followed when carrying out field-circuit data interaction iteration within a single simulation time step, and are used to complete the implementation of the operation logic of the two simulation operations.

[0074] In one optional embodiment, the transformer full-time domain simulation operation is discretized along the time axis into multiple consecutive individual simulation time steps. Step 200 is a general operation that needs to be performed in each individual simulation time step within the full-time domain simulation cycle; that is, in each individual simulation time step, multiple rounds of field-circuit alternating closed-loop iterations under the current instantaneous operating condition are carried out according to the field simulation operation rules and circuit simulation operation rules constructed in step 100.

[0075] Specifically, in a multi-round iterative process within a single simulation time step, the iteration termination determination utilizes iterative loop-driven logic as follows: iterations within a single simulation time step terminate upon convergence of key physical quantities in the field circuit. The iteration process continuously reuses the predetermined simulation rules from step 100 to cyclically update the induced voltage. Once the parameter deviation reaches the target, the iteration ends and proceeds to the next simulation time step. Specifically, as follows... Figure 4 This step 200 can be further refined to include the following implementation steps 210-230: Step 210: Complete the initial parameter configuration within the current single simulation time step to start the iteration loop corresponding to this simulation time step.

[0076] In the iterative initialization phase, this embodiment completes the initial parameter configuration for the current single simulation time step, assigns initial boundary conditions and calculation benchmarks to the field path iterative calculation of this time step, smoothly starts the iterative cycle of this time step, and realizes the smooth continuation of parameters between adjacent simulation time steps.

[0077] Step 220: According to the circuit simulation operation rules corresponding to the second task, the winding loop current is obtained by calculating the input induced voltage. The winding loop current is used as the external excitation input to the first task to perform magnetic field calculation. The winding induced voltage is obtained by reverse conversion based on the magnetic field calculation result, thus forming a round of field-circuit alternating closed-loop operation.

[0078] In the closed-loop operation of alternating field and circuit, the embodiments of this application complete the bidirectional data interaction between field and electricity based on two sets of operating rules: the circuit current is solved by the induced voltage, and the magnetic field is solved by the winding current as the external excitation. The induced voltage is deduced from the magnetic field result, forming a closed-loop data coupling link of field → circuit → field in a single time step. This restores the real physical coupling relationship between the electromagnetic field and the external circuit, and gets rid of the drawback of the software-segmented data transmission between FEM and SPICE.

[0079] Step 230: Determine the input parameters for the next round of field-circuit alternating closed-loop operation based at least on the winding induced voltage obtained by reverse conversion, and iteratively execute the field-circuit alternating closed-loop operation; when the difference in winding induced voltage obtained based on the two adjacent rounds of field-circuit alternating closed-loop operation meets the preset convergence threshold, terminate the time step alternating iterative operation corresponding to the current single simulation time step.

[0080] In the cyclic convergence determination stage, the difference in winding induced voltage corresponding to two adjacent rounds of alternating closed-loop field circuit operations is used as the convergence criterion. If the difference in winding induced voltage obtained from adjacent iterations does not reach the preset convergence threshold, the updated winding induced voltage is continuously reused, and the alternating closed-loop field circuit operation is repeated. When the difference in winding induced voltage obtained from two adjacent rounds of alternating closed-loop field circuit operations meets the preset convergence threshold, the iteration process within the current single simulation time step is immediately terminated, and the operation is switched to the next single simulation time step. This determination method can effectively ensure that the field data and circuit parameters within a single simulation time step meet the physical self-consistency requirements, accurately determine the single time step iteration termination and time step switching nodes, and provide basic computational data with good convergence and reliability for subsequent full-time domain continuous simulation.

[0081] It should be noted that the winding induced voltage is a key connecting physical quantity between field simulation and circuit simulation: the field simulation solves for the magnetic field distribution, and the winding induced voltage is derived through flux linkage integration, which directly serves as the excitation input for the circuit simulation; conversely, the electrical parameters obtained from the circuit simulation affect the boundary conditions of the field simulation. The winding induced voltage centrally reflects the energy transfer and parameter interaction between the magnetic field and the circuit, and can simultaneously reflect the changing states of parameters on both sides of the field and the circuit. Therefore, the difference in winding induced voltage between two adjacent rounds of alternating closed-loop operation of the field and circuit is selected as the convergence criterion. This can simultaneously characterize the fluctuation degree of the coupling results on both sides of the field and the circuit. A single characteristic quantity can be used to comprehensively evaluate whether the field-circuit coupling system tends to be stable, avoiding the problems of complex judgment and difficulty in matching standards caused by setting multiple convergence conditions for field and circuit quantities separately.

[0082] Furthermore, in an optional embodiment, step 230, which determines the input parameters for the next round of field-circuit alternation closed-loop operation based at least on the winding induced voltage obtained by reverse conversion, and iteratively executes the field-circuit alternation closed-loop operation, further includes: introducing a preset relaxation factor, weighting the winding induced voltage obtained in the current round of field-circuit alternation closed-loop operation to generate an updated winding induced voltage; and using the updated winding induced voltage as the input parameter for the next round of field-circuit alternation closed-loop operation.

[0083] Its purpose is to correct the parameter weights by using a relaxation factor to correct the winding induced voltage obtained by iteration, thereby smoothing the amplitude of single-step parameter changes, suppressing the drastic parameter jumps caused by transformer core nonlinearity, effectively reducing numerical oscillations during the iteration process, improving the computational stability of the single-step iteration process, and accelerating the iteration convergence speed.

[0084] In a preferred embodiment, an iterative optimization mechanism with time-step adaptive relaxation factor adjustment capability can be provided. Specifically, it includes: during the closed-loop iterative operation of field-circuit alternation in a single simulation time step, abandoning the fixed numerical form of relaxation factor configuration, and dynamically and adaptively adjusting the relaxation factor value according to the fluctuation amplitude of field physical quantities and circuit parameters in the current iteration round; when the parameter fluctuation amplitude is large, increasing the relaxation factor to smooth the parameter change rate; when the parameters tend to stabilize, decreasing the relaxation factor to improve calculation accuracy, and completing the dynamic control of the weighted update strategy based on the real-time operation status.

[0085] Compared to the iterative mode using a fixed relaxation factor, this adaptive adjustment mechanism can specifically suppress numerical oscillations caused by transient transformer conditions and deep core saturation conditions, taking into account both the stability and accuracy of iterative operations, effectively shortening the convergence time of a single time step iteration, and improving the overall computational efficiency and operational reliability of full-time domain simulation.

[0086] The above steps 210-230 are explained exemplarily. Specifically, within a single simulation time step, the simulation job rules corresponding to the first and second jobs established in step 100 are invoked throughout the process, and the time step alternation iteration job is executed according to the seven-step process (S1)-(S7): (S1) Iterative initialization: Given the initial induced voltage waveform corresponding to the time step of this simulation, complete the data initialization configuration for the time step iteration.

[0087] (S2) Start iteration counting: Enter the kth iteration loop of the current simulation time step.

[0088] (S3) Solve the winding current based on circuit simulation rules: Take the induced voltage obtained from the previous iteration as the input parameter of the second task, follow the established simulation task rules, substitute it into the primary and secondary winding circuit equations to carry out the calculation, and obtain the current response corresponding to the primary and secondary circuits.

[0089] (S4) Current mapping excitation + solving the magnetic field solution according to the field rules: The current waveform obtained in the previous step is combined with the preset number of turns and cross-sectional area parameters of the winding and mapped to the current density source term corresponding to the two-dimensional electromagnetic field. The current density source term is used as the external excitation. All field operations strictly follow the simulation operation rules corresponding to the first task. The network optimization training is carried out based on the physical constraint neural network of domain decomposition, the preset multi-component loss function, and the continuous constraint conditions of the subdomain interface. The field-circuit coupling loss function is minimized through the optimization algorithm until the neural network model converges. Finally, the magnetic field solution corresponding to this iteration is output.

[0090] (S5) Inversely calculate the induced voltage by integrating and differentiating the magnetic field solution: Using the magnetic field solution output by (S4), calculate the induced voltage of the corresponding winding circuit in reverse. The actual calculation is divided into two steps: First, perform spatial integration on the magnetic vector potential within the corresponding subdomain of the winding to solve for the total magnetic flux of the winding. Then, differentiate the total magnetic flux with respect to time to obtain the induced voltage.

[0091] In the engineering implementation stage, the spatial integral is approximated by taking the spatial average value of the magnetic vector position at multiple points in the winding subdomain, and the time differential is solved with high precision by using the automatic differential function of the neural network. Finally, the inverse calculation of the induced voltage is completed, and the following formulas (7) and (8) are used.

[0092] Formula (7); Formula (8); In the two formulas above: This represents the total flux linkage obtained in the k-th iteration; The induced voltage waveform is obtained by back-reasoning in the k-th iteration; A k (x,y,t) It is the k-th iteration magnetic field solution output by (S4); S_winding The region is the integration area of ​​the winding cross-sectional area; N is the number of turns in the winding; S is the cross-sectional area of ​​the winding. x, y t represents spatial coordinates, and t represents time.

[0093] (S6) Relaxation factor weighted update of induced voltage: This is achieved using the following formula (9).

[0094] Formula (9); In formula (9): a preset relaxation factor α is introduced to adjust the induced voltage obtained by reverse calculation. Compared with the voltage E of the previous iteration k-1 The weighted average generates the induced voltage E after this round of iteration. k .

[0095] (S7) Iterative parameter feedback and cyclic iteration: The weighted new induced voltage is used as the input induced voltage waveform for the (k+1)th iteration. Jump to (S3) and continue to carry out the next alternating iteration within the simulation time step. Repeat the entire closed-loop operation process until the iteration converges.

[0096] In the multi-round iterative process of a single simulation time step, the iteration termination determination is implemented using iterative loop driving logic: the updated induced voltage is obtained by weighting correction with relaxation factor in (S6), and is sent to (S7) as the input parameter for the next round of iteration and then back to (S3) to restart the iterative operation; when the numerical deviation of the winding induced voltage obtained by the alternating closed-loop operation of the field circuit in two adjacent iteration rounds meets the preset convergence determination threshold, all iterative operations within this simulation time step are terminated, the field distribution and circuit parameters of this time step are locked, and the entire set of alternating iterative operations for the next simulation time step is entered.

[0097] Steps 210-230 above construct a complete field-circuit alternating closed-loop iterative computation system within a single simulation time step. Based on the full-process logic of standardized initialization, closed-loop iterative computation, parameter relaxation correction, and round convergence determination, the time-sequential, adaptive, and self-converging computation of transformer field-circuit coupled simulation is realized. In each simulation time step, this application completes bidirectional iterative updates of circuit and field data through multiple rounds of field-circuit alternating closed-loop operations, abandoning the limitations of traditional fixed-parameter iterative modes. By using relaxation factors to weight and correct the iterative voltage parameters, numerical oscillations caused by core nonlinearity are effectively suppressed, significantly improving the stability of single-time-step iterative computation. Simultaneously, the difference in winding induced voltage between adjacent iteration rounds of field-circuit alternating closed-loop operations is used as the convergence criterion to accurately control the iteration termination node. While ensuring the physical self-consistency of single-time-step field-circuit parameters, the simulation calculation accuracy and iterative convergence efficiency are balanced, providing stable, accurate, and self-converging time-step computation support for continuous full-time-domain transformer simulation, significantly improving the overall consistency and reliability of the simulation results.

[0098] Specifically, in one optional embodiment, step 200 completes the time-step alternation iteration within a single simulation time step; step 300, based on the results of each single time step calculation, advances the entire time-domain simulation cycle along the time sequence. Specifically, as... Figure 5 As shown, step 300 can be further refined to include the following implementation steps 310-330: Step 310: For the current single simulation time step, obtain the field data and circuit parameters after the corresponding time step alternation iteration operation has been completed and converged, and use them as the initial input conditions for the next single simulation time step.

[0099] This step is a timing sequence configuration. The field and circuit parameters after the convergence of a single time step are used as the initial boundary of the next simulation step. This enables seamless data inheritance between adjacent time steps, conforms to the physical law of continuous change of electrical and magnetic field parameters in actual transformer operation, avoids data gaps between time steps and distortion of initial conditions, and provides a reliable timing sequence connection basis for continuous simulation in the full time domain.

[0100] Step 320: Based on the initial input conditions, iteratively execute the time step alternation iterative operation corresponding to the next single simulation time step until the individual simulation time steps contained in the full time domain simulation cycle are traversed one by one in time sequence.

[0101] This step is a time-by-time sequential operation. Using the initial parameters obtained from the sequential operation, it performs single-step alternating iterations one by one, traversing the entire time-domain simulation cycle along the time axis, and reproducing each instantaneous operating condition in the order of real time evolution. This achieves step-by-step refined simulation of transient processes such as faults and short circuits, and completely restores the dynamic operation process of the transformer throughout its entire cycle.

[0102] Step 330: After a single complete traversal of the full-time domain simulation cycle, verify whether the changes in the field data and circuit parameters corresponding to the full-time domain simulation cycle meet the preset self-consistency standard. If they do, output the full-time domain transient simulation result of the transformer. If they do not meet the standard, iterate and execute the operation of a complete traversal of the full-time domain simulation cycle again until the preset self-consistency standard is met, and output the full-time domain transient simulation result of the transformer.

[0103] This step is a global self-consistent determination. After a full-time domain traversal, the overall field and circuit parameter change patterns are checked. If the parameters do not meet the standards, the entire cycle is re-iterated to eliminate the field-electric coupling deviation caused by the accumulation of serial operations step by step. This ensures that the entire set of simulation data in the full-time domain meets the physical self-consistency constraints. The final simulation results closely match the actual operating characteristics of the transformer, greatly improving the accuracy of transient simulation.

[0104] Steps 310-330 above construct the timing progression and global convergence control flow within the full-time-domain simulation cycle. Based on the closed-loop logic of seamless inheritance of single-step convergence data, step-by-step sequential calculation, and whole-cycle self-consistency verification, continuous and self-consistent calculation of transformer full-time-domain transient simulation is realized. This embodiment uses the single-step iterative convergence result as the initial boundary condition for the next step, ensuring the temporal continuity of electrical and magnetic field parameters between adjacent steps, effectively avoiding data discontinuities and jump distortions, and laying a reliable foundation for continuous full-time-domain simulation. Simultaneously, through whole-cycle self-consistency verification after full-time-domain traversal, the field-circuit coupling deviation accumulated by step-by-step serial calculation can be identified and eliminated. While maintaining the precision of single-step refined simulation, it ensures the overall physical self-consistency of the field and circuit parameters throughout the entire cycle, significantly improving the realism and reliability of the simulation results for transient conditions such as transformer faults and short circuits.

[0105] Furthermore, as a response to the above Figure 1 , Figure 2 , Figure 3 and Figure 4 , Figure 5 To implement the method shown, this application provides a transformer field-circuit coupling time-domain simulation device. This device embodiment corresponds to the aforementioned method embodiment. For ease of reading, this device embodiment will not repeat the details of the aforementioned method embodiment, but it should be understood that the device in this embodiment can implement all the contents of the aforementioned method embodiment. This device is applied to provide a technical solution that can realize full-time-domain transformer field-circuit coupling simulation, specifically as follows... Figure 6 As shown, the device includes: The construction unit 410 is used to establish corresponding field simulation operation rules for the first operation included in the full-time domain simulation operation of the transformer, and to establish corresponding circuit simulation operation rules for the second operation included in the full-time domain simulation operation of the transformer; wherein, the first operation is a two-dimensional electromagnetic field simulation operation of the transformer, which is used to collect relevant physical quantities of the spatial magnetic field under the field simulation to provide excitation basis for the circuit simulation; the second operation is an external circuit simulation operation of the transformer winding, which is used to collect winding circuit parameters under the circuit simulation to provide boundary input for the field simulation. The time-step processing unit 420 is used to perform time-step alternating iterative operations according to the field simulation operation rules and the circuit simulation operation rules within a single simulation time step. The time-step alternating iterative operations are to alternately execute the first operation and the second operation, and determine the input of the operation to be executed later based on the output obtained from the operation executed first. The first simulation operation uses the initial input. The full-time domain processing unit 430 is used to sequentially traverse and execute the time step alternation iteration operation corresponding to each of the other individual simulation time steps included in the full-time domain simulation cycle until the simulation result corresponding to the full-time domain simulation cycle meets the preset self-consistency standard, and output the full-time domain transient simulation result of the transformer.

[0106] In an optional embodiment, the timing processing unit 420 is further specifically used for: The initial parameter configuration is completed within the current single simulation time step; according to the circuit simulation operation rules corresponding to the second operation, the winding loop current is calculated from the input induced voltage, and the winding loop current is used as an external excitation input to the first operation to perform magnetic field calculation. The winding induced voltage is calculated in reverse based on the magnetic field calculation result, thus forming a round of field-circuit alternating closed-loop operation. The input parameters for the next round of field-circuit alternating closed-loop operation are determined based at least on the winding induced voltage obtained by the reverse conversion, and the field-circuit alternating closed-loop operation is executed iteratively. When the difference in winding induced voltage obtained based on two adjacent rounds of field-circuit alternating closed-loop operation meets the preset convergence threshold, the time-step alternating iterative operation corresponding to the current single simulation time step is terminated.

[0107] In an optional embodiment, during the execution of the time-stepping processing unit 420, the input parameters for the next round of field-path alternation closed-loop operation are determined at least based on the winding induced voltage obtained by the reverse conversion, and the field-path alternation closed-loop operation is executed iteratively, including: introducing a preset relaxation factor, weighting the winding induced voltage obtained in the current round of the field-path alternation closed-loop operation to generate an updated winding induced voltage; and using the updated winding induced voltage as the input parameter for the next round of field-path alternation closed-loop operation.

[0108] In an optional embodiment, the full-time domain processing unit 430 is further specifically used for: For the current single simulation time step, the field data and circuit parameters after the corresponding time step alternation iteration operation has been completed and converged are obtained as the initial input conditions for the next single simulation time step. Based on the initial input conditions, the time step alternation iteration operation corresponding to the next single simulation time step is iteratively executed until each single simulation time step contained in the full-time domain simulation cycle is traversed sequentially. After a single complete traversal of the full-time domain simulation cycle, it is verified whether the change law of the field data and the change law of the circuit parameters corresponding to the full-time domain simulation cycle meet the preset self-consistency standard. If they meet the standard, the full-time domain transient simulation result of the transformer is output. If they do not meet the standard, the operation of completely traversing the full-time domain simulation cycle is iterated again until the preset self-consistency standard is met, and the full-time domain transient simulation result of the transformer is output.

[0109] In one alternative embodiment, such as Figure 7 As shown, the transformer field-circuit coupling time-domain simulation device further includes a determination unit 440; the determination unit 440 includes a first construction module 441 and a second construction module 442; The first construction module 441 is specifically used to: before establishing the field simulation operation rules and the circuit simulation operation rules, establish control equations characterized by time-domain magnetic vector potential for the field simulation operation content corresponding to the full-time-domain simulation of the transformer; and construct a two-dimensional electromagnetic field simulation operation of the transformer based on the control equations, as the first operation. The second construction module 442 is specifically used for: before establishing the field simulation operation rules and the circuit simulation operation rules, based on Kirchhoff's voltage law, for the circuit simulation operation content corresponding to the full-time domain simulation of the transformer, respectively building the ordinary differential control equations corresponding to the primary winding and the secondary winding; and constructing the external circuit simulation operation of the transformer winding based on the ordinary differential control equations as the second operation.

[0110] In one alternative embodiment, such as Figure 7 As shown, the construction unit 410 further includes a third construction module 411 and a fourth construction module 412; the third construction module 411 is used to: for the first task, divide multiple physical subdomains and construct corresponding field operation rules using loss constraints and subdomain interface continuity constraints to obtain the field simulation operation rules corresponding to the first task; the fourth construction module 412 is used to: for the second task, construct corresponding circuit operation rules using induced voltage as input parameters, combined with circuit equations and numerical integration algorithms to obtain the circuit simulation operation rules corresponding to the second task.

[0111] In an optional embodiment, during the execution of the third construction module 411, the step of constructing corresponding field operation rules for the first task by dividing multiple physical subdomains and utilizing loss constraints and continuous constraints at subdomain interfaces includes: constructing a basic field simulation model based on a physical constraint neural network of domain decomposition, wherein the physical constraint neural network is the computational carrier for subdomain division and constraint application; dividing the two-dimensional electromagnetic field region of the transformer into four physical subdomains according to the physical structure of the transformer: left iron subdomain, right iron subdomain, winding subdomain, and peripheral air subdomain; configuring corresponding dielectric property parameters within each physical subdomain, setting continuous constraint conditions at the interface of adjacent subdomains, and combining global loss constraints to complete the construction of field operation rules.

[0112] In an optional embodiment, during the execution of the fourth construction module 412, for the second task, the corresponding circuit operation rules are constructed by using the induced voltage as an input parameter and combining circuit equations with a numerical integration algorithm. This includes: establishing the circuit ordinary differential equations corresponding to the primary winding and the secondary winding respectively based on Kirchhoff's voltage law; receiving the winding induced voltage waveform as the operation input parameter and retrieving the known primary winding input voltage; constructing a set of ordinary differential equations containing the primary and secondary winding currents based on the circuit ordinary differential equations corresponding to the primary winding and the secondary winding respectively; and using a numerical integration method to perform step-by-step calculations on the obtained set of ordinary differential equations along the simulation time to obtain the primary winding current waveform and the secondary winding current waveform in the full time domain, thereby completing the construction of the circuit operation rules.

[0113] In summary, the transformer field-circuit coupling time-domain simulation device provided in this application includes a processor and a memory. The aforementioned construction unit, time-step processing unit, and full-time-domain processing unit are all stored in the memory as program units, and the processor executes the aforementioned program units stored in the memory to realize the corresponding functions.

[0114] The processor contains a kernel, which retrieves the corresponding program units from memory. One or more kernels can be configured, and by adjusting kernel parameters, a technical solution can be provided to achieve full-time-domain field-circuit coupling simulation of transformers, effectively improving the fit and calculation accuracy of field-circuit coupling simulation results under transient conditions.

[0115] This application provides a computer-readable storage medium storing a computer program, which, when executed by a processor, implements the transformer field-circuit coupling time-domain simulation method as described above.

[0116] This application provides an electronic device, which includes at least one processor, at least one memory and a bus connected to the processor; wherein the processor and the memory communicate with each other through the bus; the processor is used to call program instructions in the memory to execute the transformer field-circuit coupling time-domain simulation method as described above.

[0117] This application also provides a computer program product that, when executed on a data processing device, is suitable for executing the steps of initializing a time-domain simulation method for transformer field-circuit coupling.

[0118] This application is described with reference to flowchart illustrations and / or block diagrams of methods, apparatus (systems), and computer program products according to embodiments of this application. It will be understood that each block of the flowchart illustrations and / or block diagrams, and combinations of blocks in the flowchart illustrations and / or block diagrams, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, special-purpose computer, embedded processor, or other programmable data processing apparatus to produce a machine, such that the instructions, which execute via the processor of the computer or other programmable data processing apparatus, generate instructions for implementing the flowchart... Figure 1 One or more processes and / or boxes Figure 1 A device that provides the functions specified in one or more boxes.

[0119] In a typical configuration, the device includes one or more processors (CPUs), memory, and a bus. The device may also include input / output interfaces, network interfaces, etc.

[0120] Memory may include non-persistent memory in computer-readable media, such as random access memory (RAM) and / or non-volatile memory, such as read-only memory (ROM) or flash RAM, and memory includes at least one memory chip. Memory is an example of computer-readable media.

[0121] Computer-readable media includes both permanent and non-permanent, removable and non-removable media that can store information using any method or technology. Information can be computer-readable instructions, data structures, modules of programs, or other data. Examples of computer storage media include, but are not limited to, phase-change memory (PRAM), static random access memory (SRAM), dynamic random access memory (DRAM), other types of random access memory (RAM), read-only memory (ROM), electrically erasable programmable read-only memory (EEPROM), flash memory or other memory technologies, CD-ROM, digital versatile optical disc (DVD) or other optical storage, magnetic tape, magnetic magnetic disk storage or other magnetic storage devices, or any other non-transferable medium that can be used to store information accessible by a computing device. As defined herein, computer-readable media does not include transient computer-readable media, such as modulated data signals and carrier waves.

[0122] It should also be noted that the terms "comprising," "including," or any other variations thereof are intended to cover non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such process, method, article, or apparatus. Unless otherwise specified, an element defined by the phrase "comprising one..." does not exclude the presence of other identical elements in the process, method, article, or apparatus that includes that element.

[0123] Those skilled in the art will understand that embodiments of this application can be provided as methods, systems, or computer program products. Therefore, this application can take the form of a completely hardware embodiment, a completely software embodiment, or an embodiment combining software and hardware aspects. Furthermore, this application can take the form of a computer program product embodied on one or more computer-usable storage media (including, but not limited to, disk storage, CD-ROM, optical storage, etc.) containing computer-usable program code.

[0124] The above are merely embodiments of this application and are not intended to limit the scope of this application. Various modifications and variations can be made to this application by those skilled in the art. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of this application should be included within the scope of the claims of this application.

Claims

1. A time-domain simulation method for transformer field-circuit coupling, characterized in that, The method includes: For the first task included in the full-time-domain transformer simulation, corresponding field simulation rules are established, and for the second task included in the full-time-domain transformer simulation, corresponding circuit simulation rules are established. The first task is a two-dimensional electromagnetic field simulation of the transformer, used to collect relevant physical quantities of the spatial magnetic field under field simulation to provide excitation basis for circuit simulation. The second task is an external circuit simulation of the transformer winding, used to collect winding circuit parameters under circuit simulation to provide boundary input for field simulation. Within a single simulation time step, time-step alternating iterative operations are performed according to the field simulation operation rules and the circuit simulation operation rules. The time-step alternating iterative operations alternately execute the first operation and the second operation, and determine the input of the subsequent operation based on the output obtained from the first operation. The first simulation operation uses the initial input. The alternating execution of the first operation and the second operation includes: calculating the winding loop current from the input induced voltage according to the circuit simulation operation rules corresponding to the second operation; using the winding loop current as an external excitation input to the first operation to perform magnetic field calculation; and calculating the winding induced voltage in reverse based on the magnetic field calculation result. For each of the other individual simulation time steps included in the full-time simulation cycle, the time step alternation iteration operation corresponding to each individual simulation time step is traversed and executed in sequence until the simulation result corresponding to the full-time simulation cycle meets the preset self-consistency standard, and the full-time transient simulation result of the transformer is output.

2. The method according to claim 1, characterized in that, The step of performing alternating iterative operations within a single simulation time step, according to the field simulation operation rules and the circuit simulation operation rules, includes: Complete the initial parameter configuration within the current single simulation time step; By performing the first and second tasks alternately once, a round of field-road alternating closed-loop operation is formed; The input parameters for the next round of field-circuit alternation closed-loop operation are determined based at least on the winding induced voltage obtained by the reverse conversion, and the field-circuit alternation closed-loop operation is executed iteratively. When the difference in winding induced voltage obtained from the alternating closed-loop operation of two adjacent field circuits meets the preset convergence threshold, the alternating time step iteration operation corresponding to the current single simulation time step is terminated.

3. The method according to claim 2, characterized in that, The step of determining the input parameters for the next round of field-circuit alternation closed-loop operation based at least on the winding induced voltage obtained by the reverse conversion, and iteratively executing the field-circuit alternation closed-loop operation, includes: A preset relaxation factor is introduced to weight the winding induced voltage obtained from the current round of field circuit alternation closed-loop operation, and an updated winding induced voltage is generated. The updated winding induced voltage is used as the input parameter for the next round of field-circuit alternation closed-loop operation.

4. The method according to claim 1, characterized in that, For each individual simulation time step included in the full-time simulation cycle, the time step alternation iteration operation corresponding to each individual simulation time step is traversed and executed in sequence until the simulation result corresponding to the full-time simulation cycle meets the preset self-consistency standard, and the full-time transient simulation result of the transformer is output, including: For the current single simulation time step, obtain the field data and circuit parameters after completing the corresponding time step alternation iteration operation and convergence, and use them as the initial input conditions for the next single simulation time step. Based on the initial input conditions, the time step alternation iteration operation corresponding to the next single simulation time step is executed iteratively until each single simulation time step contained in the full time domain simulation cycle is traversed sequentially. After a single complete traversal of the full-time domain simulation cycle, it is verified whether the changes in the field domain data and the changes in the circuit parameters corresponding to the full-time domain simulation cycle meet the preset self-consistency standard. If satisfied, output the full-time transient simulation results of the transformer; If the conditions are not met, the operation of completely traversing the full-time domain simulation cycle is re-executed until the preset self-consistency criterion is met, and the full-time domain transient simulation results of the transformer are output.

5. The method according to any one of claims 1 to 4, characterized in that, Before establishing the field simulation operation rules and the circuit simulation operation rules, determine the first and second operations included in the transformer full-time domain simulation operation, including: For the field simulation tasks corresponding to the full-time-domain simulation of transformers, a control equation represented by the time-domain magnetic vector potential is established. A two-dimensional electromagnetic field simulation task for a transformer is constructed based on the aforementioned control equations, which serves as the first task. Based on Kirchhoff's voltage law, for the circuit simulation task corresponding to the full-time domain simulation of the transformer, the ordinary differential control equations corresponding to the primary winding and the secondary winding are respectively constructed. The second task is to construct a simulation of the external circuit of the transformer winding based on the ordinary differential control equation.

6. The method according to claim 1, characterized in that, The establishment of corresponding field simulation rules for the first task included in the full-time-domain transformer simulation task, and the establishment of corresponding circuit simulation rules for the second task included in the full-time-domain transformer simulation task, include: For the first task, the field simulation task rules corresponding to the first task are obtained by dividing the physical subdomains into multiple types and constructing the corresponding field operation rules using loss constraints and subdomain interface continuity constraints. For the second task, the circuit simulation task rules corresponding to the second task are obtained by using the induced voltage as the input parameter, combining the circuit equation with the numerical integration algorithm to construct the corresponding circuit operation rules.

7. The method according to claim 6, characterized in that, For the first task, the corresponding field operation rules are constructed by dividing the physical subdomains into multiple categories and utilizing loss constraints and subdomain interface continuity constraints, including: A field-based simulation model is constructed based on a physical constraint neural network for domain decomposition, wherein the physical constraint neural network serves as the computational carrier for subdomain partitioning and constraint application. According to the physical structure of the transformer, the two-dimensional electromagnetic field region of the transformer is divided into four physical subdomains: the left iron proton domain, the right iron proton domain, the winding subdomain, and the peripheral air subdomain. By configuring corresponding medium property parameters within various physical subdomains, setting continuous constraints at the interfaces of adjacent subdomains, and combining global loss constraints, field operation rules are constructed.

8. The method according to claim 6, characterized in that, For the second task, the corresponding circuit operation rules are constructed by using the induced voltage as an input parameter, combining circuit equations with numerical integration algorithms, including: Based on Kirchhoff's voltage law, establish the ordinary differential equations for the primary winding and the secondary winding respectively; The received winding induced voltage waveform is used as the calculation input parameter, and the known primary winding input voltage is retrieved. Based on the ordinary differential equations of the circuits corresponding to the primary winding and the secondary winding, a set of ordinary differential equations containing the primary winding current and the secondary winding current is constructed simultaneously. The numerical integration method is used to perform step-by-step calculations on the system of ordinary differential equations obtained by the simultaneous construction along the simulation time to obtain the primary winding current waveform and secondary winding current waveform in the full time domain, so as to construct the circuit operation rules.

9. A transformer field-circuit coupling time-domain simulation device, characterized in that, The device includes: The construction unit is used to establish corresponding field simulation operation rules for the first operation included in the full-time domain simulation of the transformer, and to establish corresponding circuit simulation operation rules for the second operation included in the full-time domain simulation of the transformer; wherein, the first operation is a two-dimensional electromagnetic field simulation of the transformer, used to collect relevant physical quantities of the spatial magnetic field under field simulation to provide excitation basis for circuit simulation; the second operation is an external circuit simulation of the transformer winding, used to collect winding circuit parameters under the circuit simulation to provide boundary input for the field simulation; The time-step processing unit is used to perform time-step alternating iterative operations according to the field simulation operation rules and the circuit simulation operation rules within a single simulation time step. The time-step alternating iterative operations alternately execute the first operation and the second operation, and determine the input of the operation to be executed later based on the output obtained from the operation executed first. The first simulation operation uses the initial input. The alternating execution of the first operation and the second operation includes: calculating the winding loop current from the input induced voltage according to the circuit simulation operation rules corresponding to the second operation; using the winding loop current as an external excitation input to the first operation to perform magnetic field calculation; and calculating the winding induced voltage in reverse based on the magnetic field calculation result. The full-time-domain processing unit is used to sequentially traverse and execute the time-step alternation iterative operation corresponding to each of the other individual simulation time steps included in the full-time-domain simulation cycle, until the simulation result corresponding to the full-time-domain simulation cycle meets the preset self-consistency standard, and output the full-time-domain transient simulation result of the transformer.

10. A computer-readable storage medium, characterized in that, The computer-readable storage medium stores a computer program that, when executed by a processor, implements the transformer field-circuit coupling time-domain simulation method as described in any one of claims 1-8.

11. An electronic device, characterized in that, The device includes at least one processor, and at least one memory and bus connected to the processor; The processor and the memory communicate with each other via the bus. The processor is used to call program instructions in the memory to execute the transformer field-circuit coupling time-domain simulation method as described in any one of claims 1-8.

Images