Multi-physics coupling power equipment real-time simulation method, device, equipment and medium

Abstract

The application discloses a kind of multi-physical field coupling power equipment real-time simulation method, device, equipment and storage medium, for solving the technical problems that significant deviation exists between the simulation results of existing simulation system and the actual operating state of equipment.The present application comprises: when receiving operation starting signal, in each simulation step, the resistivity correction coefficient of last thermal field calculation is obtained to carry out electromagnetic field calculation, and electromagnetic simulation data is generated;Electromagnetic simulation data includes electromagnetic force distribution data and loss distribution data;When the counter value of global time synchronization scheduler reaches the first preset value, the force field calculation is carried out using electromagnetic force distribution data, and the mechanical simulation data of iron core structure is generated;When the counter value reaches the second preset value, the thermal field calculation is carried out using the loss distribution data of second preset value simulation step, and thermal simulation data is generated;Output electromagnetic simulation data, mechanical simulation data and thermal simulation data.

Description

Technical Field

[0001] This invention relates to the field of power equipment simulation technology, and in particular to a method, apparatus, equipment and medium for real-time simulation of power equipment using multi-physics coupling. Background Technology

[0002] As power systems develop towards a higher proportion of new energy sources and greater power electronics, core power equipment such as transformers and motors face more complex electromagnetic, thermal, and mechanical stress conditions. The safe operation and remaining life assessment of this equipment depend not only on electrical parameters but also on multi-physics factors such as insulation material temperature rise and core vibration. Constructing high-fidelity digital twin models of power equipment to achieve real-time joint simulation of multi-physics fields has become a key requirement for intelligent operation and maintenance in the power industry.

[0003] Current power equipment simulation primarily relies on CPU- or GPU-based software platforms. Electromagnetic transient simulation tools such as PSCAD and ATP-EMTP can achieve microsecond-level electromagnetic transient analysis, while finite element software such as ANSYS and COMSOL can perform thermal and mechanical field analyses. However, these tools typically operate independently, making it difficult to meet the requirements for real-time performance and multi-physics closed-loop coupling. FPGAs (Field-Programmable Gate Arrays), with their hardware-level parallelism and deterministic timing advantages, have been used for real-time electromagnetic transient simulation. However, existing FPGA solutions only deploy a single electromagnetic solver core and have not yet integrated heat conduction and mechanical vibration solvers on the FPGA chip to achieve multi-physics coupling.

[0004] Existing technologies propose FPGA-based real-time electromagnetic transient simulation systems. These systems deploy an electromagnetic transient solver core within the FPGA and employ state-space or nodal analysis methods to solve for the voltage and current of power equipment in real-time steps ranging from 1 to 50 microseconds. Thermal and vibration analyses are performed offline by host computer software, with simulation results asynchronously transmitted via files or interfaces. However, the loose or unidirectional coupling between the multiphysics fields fails to accurately reflect the dynamic closed-loop interaction between electromagnetic losses, temperature rise, and material properties, resulting in significant discrepancies between the simulation results and the actual operating conditions of the equipment. Summary of the Invention

[0005] This invention provides a method, apparatus, device, and storage medium for real-time simulation of multi-physics coupled power equipment, which addresses the technical problem of significant deviations between simulation results and actual operating conditions of existing simulation systems.

[0006] This invention provides a real-time simulation method for multi-physics coupled power equipment, applied to FPGA, the method comprising:

[0007] When a start signal is received, the resistivity correction coefficient from the previous thermal field calculation is obtained in each simulation step to perform electromagnetic field calculations and generate electromagnetic simulation data; the electromagnetic simulation data includes electromagnetic force distribution data and loss distribution data.

[0008] When the counter value of the global time synchronization scheduler reaches the first preset value, the electromagnetic force distribution data is used to perform force field calculation and generate mechanical simulation data of the iron core structure.

[0009] When the counter value reaches the second preset value, thermal field calculation is performed using the loss distribution data of the second preset value simulation step size to generate thermal simulation data.

[0010] Output the electromagnetic simulation data, the mechanical simulation data, and the thermal simulation data.

[0011] Optionally, the step of obtaining the resistivity correction coefficient from the previous thermal field calculation for electromagnetic field calculation and generating electromagnetic simulation data in each simulation step when a start signal is received includes:

[0012] When the start signal is received, the resistivity correction coefficient of the previous thermal field calculation is obtained in each simulation step.

[0013] The updated resistance matrix is ​​generated using the resistivity correction coefficient.

[0014] Generate an updated state transition matrix and an updated input matrix based on the updated resistance matrix;

[0015] Electromagnetic simulation data is generated using the updated state transition matrix and the updated input matrix.

[0016] Optionally, the step of using the electromagnetic force distribution data to perform force field calculations and generate mechanical simulation data for the core structure when the counter value of the global time synchronization scheduler reaches a first preset value includes:

[0017] When the counter value of the global time synchronization scheduler reaches the first preset threshold, the mode shape vectors of the multi-mode of the core structure are extracted to form a mode matrix;

[0018] The generalized force equations for each mode of the core structure are calculated using the electromagnetic force distribution data and the modal matrix.

[0019] The generalized force equations for each mode are integrated over time using the constant acceleration method to obtain the updated mode coordinates and updated mode velocities.

[0020] The displacements of each node in the physical coordinates of the core structure are generated using the updated modal coordinates and the modal matrix.

[0021] Mechanical simulation data of the core structure is generated based on the displacement, the updated modal coordinates, and the updated modal velocity.

[0022] Optionally, the power equipment includes multiple thermal nodes; the step of generating thermal simulation data by using loss distribution data of the second preset value step size for a thermal field calculation when the counter value reaches a second preset value includes:

[0023] When the counter value reaches the second preset value, the average loss power is calculated using the loss distribution data of the second preset value simulation step size.

[0024] The average power loss is used to perform temperature update calculations on all the hot nodes to obtain thermal simulation data.

[0025] Optionally, the step of using the average power loss to perform temperature update calculations on all the thermal nodes to obtain thermal simulation data includes:

[0026] Obtain the heat capacity of each hot node, and the thermal resistance between each hot node and its adjacent hot nodes;

[0027] Obtain the time step of the thermal field;

[0028] Obtain the temperature data of each thermal node at the time step of the previous thermal field;

[0029] The current temperature data of each thermal node at the current thermal time step is calculated based on the temperature data, the thermal field time step, the heat capacity, and the thermal resistance.

[0030] By integrating the current temperature data of all the aforementioned hot nodes, thermal simulation data of the power equipment is obtained.

[0031] Optionally, the hot node includes a winding; the method further includes:

[0032] Calculate the average temperature value based on the current temperature data of all windings;

[0033] The updated resistivity correction factor is calculated based on the copper conductor temperature coefficient formula and the average temperature value.

[0034] Optionally, after the step of outputting the electromagnetic simulation data, the mechanical simulation data, and the thermal simulation data, the method further includes:

[0035] Acquire sensor measurement data;

[0036] The electromagnetic simulation data, the mechanical simulation data, the thermal simulation data, and the sensor measurement data are compared, and the parameters of the FPGA are corrected based on the comparison results.

[0037] This invention also provides a real-time simulation device for multiphysics coupled power equipment, applied to FPGA, the device comprising:

[0038] The electromagnetic simulation data generation module is used to obtain the resistivity correction coefficient of the previous thermal field calculation and perform electromagnetic field calculation in each simulation step when a start signal is received, thereby generating electromagnetic simulation data; the electromagnetic simulation data includes electromagnetic force distribution data and loss distribution data.

[0039] The mechanical simulation data generation module is used to perform force field calculations using the electromagnetic force distribution data when the counter value of the global time synchronization scheduler reaches a first preset value, thereby generating mechanical simulation data of the iron core structure.

[0040] The thermal simulation data generation module is used to perform thermal field calculation using loss distribution data of the second preset value number of simulation steps when the counter value reaches the second preset value, and generate thermal simulation data.

[0041] The output module is used to output the electromagnetic simulation data, the mechanical simulation data, and the thermal simulation data.

[0042] The present invention also provides an electronic device, the device comprising a processor and a memory:

[0043] The memory is used to store program code and transmit the program code to the processor;

[0044] The processor is used to execute the real-time simulation method for multiphysics coupled power equipment as described above, according to the instructions in the program code.

[0045] The present invention also provides a computer-readable storage medium for storing program code for executing the real-time simulation method for multiphysics coupled power equipment as described in any of the preceding claims.

[0046] As can be seen from the above technical solutions, the present invention has the following advantages: the present invention couples electromagnetic fields, force fields and thermal fields on FPGA, thereby constructing a dynamic closed-loop interaction relationship between electromagnetic loss, temperature rise and material properties, thereby improving the accuracy of real-time simulation of power equipment. Attached Figure Description

[0047] To more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0048] Figure 1 A flowchart illustrating the steps of a real-time simulation method for multiphysics-coupled power equipment provided in this embodiment of the invention;

[0049] Figure 2 A schematic diagram of the workflow of the electromagnetic transient solution kernel in each microsecond-level time step;

[0050] Figure 3 A flowchart of the process for solving the mechanical vibration kernel in each sub-millisecond time step;

[0051] Figure 4 Workflow diagram for solving the kernel for heat conduction;

[0052] Figure 5 This is a schematic diagram of the architecture of the real-time simulation method for multi-physics coupled power equipment of the present invention;

[0053] Figure 6 This is a schematic diagram of the structure of a real-time simulation system for multiphysics coupled power equipment provided in an embodiment of the present invention;

[0054] Figure 7 This is a structural block diagram of a real-time simulation device for multi-physics coupled power equipment provided in an embodiment of the present invention. Detailed Implementation

[0055] This invention provides a method, apparatus, device, and storage medium for real-time simulation of multi-physics coupled power equipment, which addresses the technical problem of significant deviations between simulation results and actual operating conditions of existing simulation systems.

[0056] To make the objectives, features, and advantages of this invention more apparent and understandable, the technical solutions of the embodiments of this invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the embodiments described below are only some embodiments of this invention, and not all embodiments. Based on the embodiments of this invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of this invention.

[0057] Please see Figure 1 , Figure 1 The flowchart illustrates the steps of a real-time simulation method for multi-physics coupled power equipment provided in this embodiment of the invention.

[0058] This invention provides a real-time simulation method for multi-physics coupled power equipment, applicable to FPGA;

[0059] Multiphysics Transient Simulation (MPS) refers to a modeling and computational method that simultaneously considers multiple physical effects, such as electromagnetic, temperature, and mechanical fields, within the same simulation framework, and establishes their interrelationships and data exchange mechanisms. MPS can reflect the closed-loop physical process in actual equipment, where current loss leads to temperature rise, temperature rise alters material properties, and material properties, in turn, affect current distribution.

[0060] An FPGA is an integrated circuit chip that can be programmed and configured by the user according to their needs. It has high parallel computing capabilities and deterministic timing characteristics, enabling the synchronous parallel execution of multiple computing tasks at the hardware level. It is widely used in real-time signal processing, high-speed data acquisition, and hardware-in-the-loop simulation.

[0061] In this embodiment of the invention, to simulate the electromagnetic, temperature, and mechanical fields of power equipment using an FPGA, it is first necessary to establish electromagnetic, thermal, and mechanical models based on the physical structure and operating specifications of the power equipment to be simulated, and prepare the parameter data required for each model. For the electromagnetic model, based on design parameters such as the number of turns of the transformer or motor windings, conductor cross-sectional area, core size, and magnetization curve, a state-space equivalent circuit model is established, and the order n of the state equation, the initial inductance matrix L, the initial resistance matrix R, the mutual inductance coefficient matrix M, and the Steinmetz coefficients k, α, and β of the core are determined. For the thermal model, the equipment's thermal domain is divided into several thermal nodes (typically 20 to 100 nodes), and the heat capacity C of each thermal node is determined. i Thermal resistance R between adjacent hot nodes ij These parameters can be calculated from the material and assembly information in the power equipment design drawings, or by referring to typical parameter values ​​in IEC standards. For the mechanical model, modal analysis of the core structure is performed using offline finite element analysis software (such as ANSYS Mechanical) to extract the first p-th order (p ranges from 5 to 20) natural frequencies ω. i Modal shape matrix Φ, modal mass m i And modal damping ratio ξ i All model parameters are encoded in binary fixed-point format and stored in the parameter configuration file.

[0062] Then, based on the model size and real-time requirements, plan and allocate hardware resources on the FPGA chip. Select an FPGA chip with sufficient DSP multipliers (recommended at least 1000 DSP48E2 units), Block RAM (recommended at least 20Mb), and logic units (recommended at least 1 million LUTs), such as the Xilinx Virtex UltraScale+ series or the Intel Stratix 10 series. Deploy the electromagnetic transient solver core in one clock domain of the FPGA, with the operating clock frequency set to 200MHz to 400MHz, to achieve real-time solving in steps of 1 to 50 microseconds. Deploy the heat conduction solver core and the mechanical vibration solver core in the same clock domain or independent clock domains, with the operating clock frequency set to 100MHz to 200MHz; when using independent clock domains, ensure predictable timing of shared data read / write and buffer switching through dual-clock shared memory or deterministic cross-clock domain synchronization mechanisms. In the Block RAM, a shared memory area is allocated, divided into an electromagnetic-thermal coupling data area (stores loss distribution vectors and temperature / resistivity correction vectors, each vector length equal to the number of hot nodes, double-buffered and occupying twice the space), an electromagnetic-mechanical coupling data area (stores electromagnetic force distribution vectors and core deformation vectors, double-buffered and occupying twice the space), and a global state parameter area. Auxiliary modules such as a global time synchronization scheduler, external data interface, parameter configuration and monitoring, and result output and recording are deployed. After FPGA synthesis, placement and routing, and timing constraint verification are completed, a bitstream file is generated.

[0063] Next, the FPGA bitstream file is downloaded to the FPGA chip to complete the hardware configuration. Through the parameter configuration and monitoring module, all model parameters prepared in step one are loaded from the host computer into the parameter storage area inside the FPGA. After loading, the system initializes: the electromagnetic solver kernel initializes the state vector x to a zero state or a specified steady-state operating point; the heat conduction solver kernel initializes the temperature vector T to the ambient temperature or a specified initial temperature distribution; and the mechanical vibration solver kernel initializes the modal coordinates q and modal velocities. Initialize to zero. All buffers in the shared memory area are cleared to zero, and the counter of the global time synchronization scheduler is reset to zero. Configure the initial values ​​of the electromagnetic-thermal step ratio N and the electromagnetic-mechanical step ratio M. After system initialization, it enters the standby state.

[0064] Based on the FPGA configured above, the embodiments of the present invention may specifically include the following steps:

[0065] Step 101: When the start signal is received, in each simulation step, the resistivity correction coefficient of the previous thermal field calculation is obtained to perform electromagnetic field calculation and generate electromagnetic simulation data; the electromagnetic simulation data includes electromagnetic force distribution data and loss distribution data.

[0066] In this embodiment of the invention, the electromagnetic simulation data includes the current waveform, magnetic flux state, electromagnetic force distribution data, and loss distribution data of each branch of the power equipment.

[0067] In this embodiment of the invention, after the FPGA receives the start signal, the electromagnetic transient solver core runs continuously in microsecond-level simulation steps. Within each step, it sequentially performs the following: reads the latest current resistivity correction coefficient and core deformation data from the front buffer area of ​​the shared memory region; updates the state equation parameters based on the current resistivity correction coefficient and core deformation data; and performs state advancement and loss calculation to obtain electromagnetic simulation data containing electromagnetic force distribution data and loss distribution data. The electromagnetic force distribution and loss distribution data are then written to the back buffer area of ​​the shared memory region.

[0068] In one example, step 101 may include the following sub-steps:

[0069] S11, When the start signal is received, the resistivity correction coefficient of the previous thermal field calculation is obtained in each simulation step;

[0070] S12, using resistivity correction coefficient to generate updated resistance matrix;

[0071] S13, Generate the update state transition matrix and update input matrix based on the update resistance matrix;

[0072] S14 uses the updated state transition matrix and updated input matrix to generate electromagnetic simulation data.

[0073] In practical implementation, taking a transformer as an example, the transformer winding is equivalent to multiple inductor-resistor branches connected in series or parallel, with each branch coupled through mutual inductance. Let the system state vector x(t) contain the current and flux linkage variables of each branch, and the input vector u(t) contain the externally applied voltage source and load impedance. Then the continuous-time state equation is: dx / dt = A·x(t) + B·u(t), where A is the system matrix and B is the input matrix. Discretizing the above equation using the trapezoidal integral method yields the recursive relation:

[0074] x(k+1)=F·x(k)+G·u(k);

[0075] Where F = (I - Δt / 2·A) -1 ·(I +Δt / 2·A) is the state transition matrix, G=(I-Δt / 2·A) -1·Δt·B is the input matrix, Δt is the time step, and I is an n×n identity matrix with the same order n as the state equation, i.e., a square matrix with all diagonal elements being 1 and all other elements being 0. The resistance parameter R in matrices F and G is not a fixed constant, but is updated in real time based on the latest temperature field data read from the shared memory region. Specifically, at the beginning of each step, the electromagnetic solver kernel reads the resistivity correction coefficients of each winding segment from the electromagnetic-thermal coupling data area of ​​the shared memory region, updates the resistance matrix R, recalculates the state transition matrix F and the input matrix G, obtaining the updated state transition matrix and the updated input matrix (or using an incremental update strategy to reduce the computational load), and then performs a matrix-vector multiplication to complete the state vector advancement.

[0076] Loss calculation is performed immediately after state progression is complete, with copper loss calculated according to... Calculate, where Ii is the current value of the i-th winding branch, Ri is the temperature-corrected resistance value of the i-th winding branch, and the summation symbol Σ is applied to all winding branches. Iron loss is calculated according to the improved Steinmetz formula. Calculate, where k, α, β are the Steinmetz coefficients of the core material, f is the frequency, and B... m This represents the peak magnetic flux density. The calculated loss distribution data is written to the back buffer of the shared storage area and read by the front-end heat conduction solver kernel during the next synchronization trigger.

[0077] like Figure 2 As shown, Figure 2 This diagram illustrates the workflow of the electromagnetic transient solver kernel within each microsecond-level time step. First, the updated temperature data from the heat conduction solver kernel is read from the on-chip shared memory and converted into resistivity correction coefficients. Then, the resistance parameter items in the state transition matrix are updated. Next, matrix-vector multiplication is performed to advance the time of the current and flux states. Subsequently, the copper losses of each winding segment and the iron losses of each core segment are calculated. Finally, the loss distribution data is written to the back buffer in the shared memory area. The entire process is executed in a hardware pipeline within the FPGA, with single-step computation latency controlled within 500 nanoseconds to ensure microsecond-level real-time requirements are met.

[0078] On-Chip Shared Memory: A data storage area consisting of Block RAM or distributed RAM resources integrated within the FPGA chip. Multiple parallel computing modules can simultaneously access this area for data reading and writing through predefined address mapping and arbitration mechanisms. It features nanosecond-level access latency and deterministic timing characteristics, avoiding the high latency and bandwidth bottleneck problems of off-chip memory.

[0079] Step 102: When the counter value of the global time synchronization scheduler reaches the first preset value, the electromagnetic force distribution data is used to calculate the force field and generate mechanical simulation data of the iron core structure.

[0080] In this embodiment of the invention, the mechanical simulation data includes the vibration displacement and acceleration time history of key nodes in the core, vibration spectrum, A-weighted sound pressure level, and deformation at key locations in the core.

[0081] In this embodiment of the invention, the counter of the global time synchronization scheduler increments with the step size of the electromagnetic transient solver kernel. When the counter value reaches M (a first preset value, which can be 10-100, corresponding to 0.01 to 5 milliseconds. This time range is determined by the product of the step size ratio M and the electromagnetic simulation step size Δt: when M takes the minimum value of 10 and Δt takes the minimum value of 1 microsecond, the mechanical field update cycle is 10 × 1 microsecond = 0.01 milliseconds; when M takes the maximum value of 100 and Δt takes the maximum value of 50 microseconds, the mechanical field update cycle is 100 × 50 microseconds = 5 milliseconds. Therefore, the range of 0.01 to 5 milliseconds is the mechanical field update cycle of the complete interval covered by the Cartesian product of the independent value ranges of M and Δt), the global time synchronization scheduler issues a mechanical field trigger signal: firstly, it performs a buffer switch (foreground and background pointer exchange) of the electromagnetic-mechanical coupling data area, and then starts the mechanical vibration solver kernel to perform mechanical field update calculation. The mechanical vibration solver reads electromagnetic force distribution data from the switched front buffer. This data can be instantaneous values ​​or short-time statistical values ​​formed by configuration (e.g., the root mean square value of electromagnetic force over M consecutive electromagnetic steps, or the time average value of electromagnetic force over M steps). It then performs Newmark-β integration to obtain the mechanical simulation data of the core structure. Based on the mechanical simulation data, the core deformation is calculated and written to the back buffer for correction of the air gap length in the electromagnetic solver.

[0082] In one example, step 102 may include the following sub-steps:

[0083] S21, when the counter value of the global time synchronization scheduler reaches the first preset threshold, extract the mode shape vectors of the multi-mode of the iron core structure to form a mode matrix;

[0084] S22, using electromagnetic force distribution data and modal matrix to calculate the generalized force equations for each mode of the iron core structure;

[0085] S23, the generalized force equations of each mode are integrated over time using the constant acceleration method to obtain the updated mode coordinates and updated mode velocities;

[0086] S24, using updated modal coordinates and modal matrices to generate the displacement of each node in the physical coordinates of the iron core structure;

[0087] S25 generates mechanical simulation data of the core structure based on displacement, updated modal coordinates, and updated modal velocities.

[0088] In practical implementation, the core structure is discretized into a finite element model with n degrees of freedom. Modal analysis is used to extract the first p natural frequencies ωi and mode shape vectors φi (i=1,2,…,p). The vibration equations in physical coordinates are... Where M0 is the mass matrix of the iron core structure (n×n order, describing the inertial characteristics of each degree of freedom; here M0 is the mechanical model symbol). Let be the acceleration vector of each node (the second derivative of displacement u with respect to time), and C be the damping matrix (n×n order, describing the energy dissipation characteristics of the system). Let u be the velocity vector of each node (the first derivative of displacement u with respect to time), K be the stiffness matrix (n×n, describing the elastic restoring force characteristics of the structure), u be the displacement vector of each node, and F(t) be the time-varying external excitation force vector, i.e., the electromagnetic force distribution data. Through modal coordinate transformation, u = Φ·q is decoupled into p independent single-degree-of-freedom vibration equations: ,in, Let q be the second derivative of the modal coordinate q with respect to time, i.e., the modal acceleration; correspondingly, Let q be the first derivative of the i-th modal coordinates with respect to time, i.e., the modal velocity. i Let φ be the coordinates of the i-th modal. i Let be the i-th order mode shape vector, describing the relative displacement shape of each node when the core structure vibrates at the i-th natural frequency. Its dimension is equal to the number of degrees of freedom n of the finite element model, and it is pre-extracted from offline finite element modal analysis. All p-th order mode shape vectors are arranged in columns to form the modal matrix. . mi is the mass of the i-th modal, ξi is the modal damping ratio, and F(t) is the electromagnetic force vector.

[0089] At the start of each mechanical field step, the mechanical vibration solution kernel reads the electromagnetic force distribution data F(t) from the electromagnetic-mechanical coupling data area of ​​the shared storage region, and calculates the generalized force of each mode through the modal matrix Φ. Then, the Newmark-β method (taking β=1 / 4, γ=1 / 2, i.e., the constant acceleration method) is used to perform time integration on each modal equation to obtain the modal coordinates q and modal velocities. The update value is obtained through the following process: Let the time step of the mechanical field be Δt_m, and the modal coordinates qi(k) and modal velocities at the current time k be known. and modal acceleration And the generalized force fi(k+1) at the next moment, first calculate the equivalent stiffness coefficient. Then calculate the equivalent load. Next, solve for and update the modal coordinates q.i (k+1) = f_eff_i / k_eff_i; then according to q i (k+1) Calculate the updated modal acceleration Finally, calculate the update mode velocity. The above integration process is performed independently for each mode, and each mode can be calculated synchronously in the FPGA using parallel multiply-accumulate units.

[0090] Finally, the displacement of each node in physical coordinates is obtained through inverse modal transformation u = Φ·q. The deformation at key locations in the core (such as the joint between the core column and the yoke) is extracted and written into the shared storage area for the electromagnetic solver kernel to correct the air gap geometry parameters. The acceleration response of each node is used to calculate the A-weighted sound pressure level and assess the equipment noise level.

[0091] Modal parameters (natural frequencies, mode shapes, modal masses, damping ratios) are loaded via the parameter configuration module before simulation begins, and are typically pre-calculated and extracted by offline finite element analysis software. Only the first p modal parameters need to be stored in the FPGA, with p typically ranging from 5 to 20 orders, which is sufficient to cover the frequency range of interest for power equipment noise and vibration analysis (usually 100Hz to 2000Hz).

[0092] like Figure 3 As shown, Figure 3 The workflow of the mechanical vibration solving kernel within each sub-millisecond time step is described. First, the electromagnetic force distribution data output by the electromagnetic solving kernel is read from the on-chip shared memory area. Then, the generalized excitation force of each mode is calculated through modal matrix transformation. Next, the single-degree-of-freedom vibration equations of each mode are solved in time using the Newmark-β numerical integration method. Subsequently, the modal coordinate results are converted into displacement and acceleration responses of each node in physical coordinates through inverse modal transformation. Finally, the deformation data of key locations in the core are written to the shared memory area for the electromagnetic solving kernel to read. The entire process utilizes parallel multiply-accumulate units in the FPGA to achieve synchronous solving of multiple modes, achieving computational efficiency that meets the sub-millisecond real-time requirements.

[0093] Step 103: When the counter value reaches the second preset value, the thermal field is calculated using the loss distribution data of the second preset value simulation step size to generate thermal simulation data.

[0094] In this embodiment of the invention, when the counter value reaches N (a second preset value, which can be from 100 to 10000, corresponding to a thermal field update cycle of 0.1 to 500 milliseconds). This time range is determined by the product of the step size ratio N and the electromagnetic simulation step size Δt: when N takes the minimum value of 100 and Δt takes the minimum value of 1 microsecond, the thermal field update cycle is 100 × 1 microsecond = 0.1 milliseconds; when N takes the maximum value of 10000 and Δt takes the maximum value of 50 microseconds, the thermal field update cycle is 10000 × 50 microseconds = 500 milliseconds. In the same system configuration, the electromagnetic simulation step size Δt is a fixed value, and N and M share the same Δt. Therefore, N > M can ensure that the thermal field update cycle is greater than the mechanical field update cycle. Generally, the thermal field timescale is greater than the mechanical field timescale. In the field timescale, N can be set to an integer multiple of M in the engineering configuration to ensure periodic alignment of trigger times. When N and M do not satisfy the integer multiple relationship, the scheduler can still trigger independently when each reaches its threshold, and use a fixed priority or parallel triggering strategy to process multiple triggering events at the same time to ensure deterministic timing. When the global time synchronization scheduler issues a thermal field trigger signal, it first performs a buffer switch of the electromagnetic-thermal coupling data area, and then starts the thermal conduction solver to perform thermal field update calculations. The thermal conduction solver reads the accumulated loss data (the time average power value of N steps of electromagnetic calculation between two adjacent thermal field triggers) from the front buffer after the switch, performs temperature iteration and resistivity correction coefficient calculations, and writes the results to the back buffer.

[0095] In one example, step 103 may include the following sub-steps:

[0096] S31, When the counter value reaches the second preset value, the average power loss is calculated using the loss distribution data of the second preset value simulation step size.

[0097] S32 uses the average power loss to perform temperature update calculations on all hot nodes, obtaining thermal simulation data.

[0098] In practical implementation, hot nodes can include winding sections, core sections, insulation layers, cooling oil channels, etc. Each hot node is described by thermal capacity and thermal resistance parameters.

[0099] The thermal simulation data includes the time curves of the temperature at each thermal node, the temperature of the winding hot spots, and the cumulative thermal aging life consumption of the insulation material.

[0100] When the calculator value reaches the second preset value, the average loss power can be calculated using the loss distribution data of the second preset value simulation step size; then the average loss power is used to perform temperature update calculations on all hot nodes to obtain thermal simulation data.

[0101] In one example, S32 may include the following sub-steps:

[0102] S321, obtain the heat capacity of each hot node, and the thermal resistance between each hot node and its adjacent hot nodes;

[0103] S322, obtain the time step of the thermal field;

[0104] S323, obtain the temperature data of each thermal node at the previous thermal field time step;

[0105] S324 calculates the current temperature data of each thermal node at the current thermal time step based on temperature data, thermal field time step, heat capacity and thermal resistance;

[0106] S325 integrates the current temperature data of all hot nodes to obtain thermal simulation data of the power equipment.

[0107] In the specific implementation, the thermal domain of the device is divided into several thermal nodes, and each thermal node i has a heat capacity C. i And the thermal resistance Rij between the adjacent hot node j. The temperature equation is: C i × dT i / dt = Σ[(T j - T i ) / R ij ] + Q i T i Let T be the temperature value of hot node i at the current moment. j R represents the temperature value of heat node j adjacent to heat node i at the current moment. ij Let be the thermal resistance between hot node i and its adjacent hot node j. The summation symbol Σ iterates through all hot nodes j adjacent to hot node i. Qi is the heat source power of node i, derived from the loss distribution data output by the electromagnetic solution kernel. After discretization using the explicit Euler method, we obtain: T i (k+1) = T i (k) + Δt_th / C i × {Σ[(T j (k) - T i (k)) / R ij ] + Q i (k)}, where T i (k+1) represents the updated temperature value of thermal node i at the (k+1)th thermal time step, T i (k) represents the current temperature value of thermal node i at the k-th thermal field time step, T j (k) represents the temperature value of the adjacent thermal node j at the k-th thermal time step, Q i(k) represents the heat source power of thermal node i within the k-th thermal field time step. Δt_th represents the thermal field time step. At the beginning of each thermal field step, the heat conduction solution kernel reads the latest loss distribution data Qi from the front buffer of the electromagnetic-thermal coupling data area in the shared memory region, then performs temperature update calculations in parallel for all thermal nodes, and writes the updated temperature vector to the back buffer of the shared memory region.

[0108] Furthermore, in this embodiment of the invention, the hot node includes a winding, and step 103 may further include the following sub-steps:

[0109] S326, calculate the average temperature value based on the current temperature data of all windings;

[0110] S327, calculate and update the resistivity correction factor based on the copper conductor temperature coefficient formula and average temperature value.

[0111] The most crucial temperature data is the average temperature of each winding segment. The heat conduction solver simultaneously calculates the corresponding resistivity correction coefficient based on the copper conductor temperature coefficient formula and writes it into the shared storage area. This allows the electromagnetic solver to directly use it in subsequent steps, avoiding redundant temperature-resistivity conversion calculations within the electromagnetic solver. The heat conduction solver also performs cumulative statistics on the temperature of insulation material nodes, calculating the equivalent thermal lifetime loss due to insulation aging. This data is used for predicting the remaining life of the equipment.

[0112] The thermal capacity and thermal resistance parameters of the thermal network model are loaded through the parameter configuration module before the simulation starts. It supports selecting preset parameter templates or manual configuration according to the equipment type (oil-immersed transformer, dry-type transformer, asynchronous motor, etc.). For oil-immersed transformers, the thermal network model includes typical thermal nodes such as high-voltage winding, low-voltage winding, core column, core yoke, upper and lower oil passages, tank wall, and radiator. The total number of nodes is generally between 20 and 100.

[0113] like Figure 4 As shown, Figure 4 The flowchart illustrates the workflow of the heat conduction solver core. First, the loss distribution data output by the electromagnetic solver core is read from the on-chip shared memory area as the heat source power input for each hot node. Then, a thermal conductivity matrix for the thermal network is constructed based on the current thermal property parameters. Next, temperature iteration update calculations are performed in parallel for all hot nodes. Subsequently, the resistivity correction coefficient for the copper conductor is calculated based on the winding temperature value. Finally, the updated temperature field data and resistivity correction coefficient are written to the back buffer in the shared memory area. This process utilizes a parallel multiplier array in the FPGA to achieve synchronous calculation of all hot nodes, with single-step calculation latency controlled within tens of microseconds.

[0114] Step 104: Output electromagnetic simulation data, mechanical simulation data, and thermal simulation data.

[0115] In this embodiment of the invention, the electromagnetic simulation data includes the current of each branch, the magnetic flux density distribution of the core (i.e., electromagnetic force distribution data), and the loss density of each winding segment, output with the electromagnetic step size as the sampling rate. The thermal simulation results include the time curves of the temperature of each hot node, the temperature of the winding hot spots, and the cumulative thermal aging life consumption of the insulation material, output with the thermal field step size as the sampling rate. The mechanical simulation results include the vibration displacement and acceleration time history of key nodes in the core, the vibration spectrum, and the A-weighted sound pressure level, output with the mechanical field step size as the sampling rate. The above results can be transmitted to the host computer monitoring system in real time through an external data interface for the following application scenarios: real-time status detection, comparing the simulation results with the actual sensor measurements to detect abnormal deviations and achieve early fault warning; life prediction, based on the insulation cumulative thermal aging model (such as the Arrhenius aging model) and the mechanical fatigue cumulative model, assessing the remaining life of the equipment based on the temperature and vibration history output by the simulation; and operation optimization, by adjusting the operating parameters in the simulation model (such as load level and cooling method), pre-simulating the impact of different operating strategies on the temperature rise and vibration of the equipment, and assisting in the decision-making of operating schemes.

[0116] In this embodiment of the invention, step 104 is followed by:

[0117] Step 105: Acquire sensor measurement data;

[0118] Step 106: Compare electromagnetic simulation data, mechanical simulation data, thermal simulation data, and sensor measurement data, and correct the FPGA parameters based on the comparison results.

[0119] During long-term operation of the simulation system, when persistent deviations occur between the actual sensor measurement data (such as winding temperature, oil temperature, and vibration acceleration) and the simulation results, correction values ​​for parameters such as thermal resistance, thermal capacity, or damping ratio can be calculated on the host computer based on data-driven parameter identification algorithms (such as least squares method or Kalman filtering). The corrected parameters are then written into the FPGA to achieve continuous calibration between the digital twin model and the physical entity. The parameter update process does not require stopping the simulation; a parameter buffering switching mechanism is used to apply the new parameters at the next synchronization trigger, ensuring the continuity of the simulation process.

[0120] This invention couples electromagnetic, force, and thermal fields on an FPGA to construct a dynamic closed-loop relationship between electromagnetic loss, temperature rise, and material properties, thereby improving the accuracy of real-time simulation of power equipment.

[0121] Please see Figure 5 , Figure 5This is a schematic diagram of the architecture of the multi-physics coupled real-time simulation method for power equipment according to the present invention. Its core idea is to unify the three physical field simulation tasks—electromagnetic transient solution, heat conduction solution, and mechanical vibration solution—which were originally located on different computing platforms, onto the same FPGA chip. Utilizing the hardware-level parallel computing capabilities and deterministic timing characteristics of the FPGA, the three physical field solution cores are run synchronously and in parallel. A real-time data exchange and closed-loop feedback mechanism between the multiple physical fields is established through a shared on-chip memory area of ​​the FPGA. Specifically, it can include the following five steps:

[0122] The first step is the design and deployment of the electromagnetic transient solver core. Based on the state-space method, the electromagnetic transient solver core establishes an equivalent electromagnetic transient circuit model containing multiple branches for transformer or motor windings. It uses the Trapezoidal Rule to discretize the state equations and implements matrix-vector multiplication and linear equation solving in a hardware pipeline within the FPGA. This solver core operates in time steps ranging from 1 to 50 microseconds, completing a full calculation and update of voltage, current, magnetic flux, and loss distribution within each step. The loss distribution includes the copper loss (I²R loss) of the windings and the iron loss of the core (hysteresis loss and eddy current loss). The winding resistivity parameter used in the copper loss calculation is read in real-time from the on-chip shared memory area, and this resistivity value is dynamically updated by the thermal conductivity solver core based on the latest temperature field calculation results.

[0123] The second step involves the design and deployment of the heat conduction solver kernel. Based on the Fourier heat conduction equation, the heat conduction solver kernel uses either the Finite Difference Method or the Lumped Parameter Thermal Network method to establish discretized thermal models of the windings, core, insulation materials, and cooling media of the power equipment. The heat conduction solver kernel operates in time steps ranging from 0.1 to 10 milliseconds. Within each step, it reads the loss distribution data output by the electromagnetic solver kernel from the on-chip shared memory area as a heat source term, calculates the temperature values ​​of each node in each region, and writes the updated temperature field data back to the shared memory area. The temperature field data includes the average temperature values ​​of each section of the winding. The electromagnetic solver kernel updates the resistivity parameters of the copper conductor based on this in the next step, thus forming a closed-loop feedback loop of current-loss-temperature-resistivity. The heat conduction solver kernel also outputs temperature time history data of the insulation material for insulation aging life assessment.

[0124] The third step is the design and deployment of the mechanical vibration solver. Based on a lumped mass-spring-damped system model or the finite element modal superposition method, the mechanical vibration solver establishes a vibration dynamics model for transformer core laminations or motor stator structures. This solver runs with a time step of 0.05 to 1 millisecond, reading electromagnetic force distribution data (including Maxwell stress and magnetostrictive force) output from the electromagnetic solver as excitation force from the on-chip shared memory area to solve for the displacement, velocity, and acceleration responses of each node in the core. The vibration response data is written back to the shared memory area, used to calculate the noise radiation level on the equipment casing surface and to feed back the core deformation to the electromagnetic solver to correct the air gap length parameter, thus forming a mechanical closed loop of electromagnetic force-vibration-air gap change.

[0125] The fourth aspect is the design and arbitration management of the on-chip shared memory region. This invention allocates a dedicated shared memory region within the FPGA's on-chip BlockRAM resources, employing a double buffering mechanism and a deterministic timing arbitration strategy to ensure that data exchange between the three solver cores at different time steps does not generate read-write conflicts, nor introduce additional waiting delays caused by mutexes / software handshakes. The address space of the shared memory region is divided into three sub-regions: an electromagnetic-thermal coupled data region, an electromagnetic-mechanical coupled data region, and a global state parameter region. Each sub-region is configured with an independent double-buffered register set to achieve lock-free data exchange.

[0126] The fifth component is a multi-timescale synchronization scheduling mechanism. Due to the order-of-magnitude differences in the time steps of the three physical field solvers (microseconds for electromagnetic field, milliseconds for thermal field, and sub-milliseconds for mechanical field), this invention designs a global time synchronization scheduler. This scheduler achieves multi-step alignment and synchronization based on counters and trigger signals: the electromagnetic solver triggers a calculation update for the thermal conduction solver after every N steps, and the mechanical vibration solver triggers a calculation update after every M steps. The values ​​of N and M are dynamically configured according to the device simulation accuracy requirements and the FPGA clock frequency. The synchronization scheduler ensures that at each thermal or mechanical field update, the coupled data in the shared memory area has completed buffer switching, and the three solvers always use a consistent coupled dataset; when thermal and mechanical field triggers occur simultaneously at the same electromagnetic step boundary, the scheduler uses a predefined approach.

[0127] Through the organic coordination of the above five steps, this invention realizes real-time closed-loop coupled simulation of the electromagnetic, thermal, and mechanical physical fields of power equipment on a single FPGA chip. This eliminates the communication delay and synchronization difficulties of multi-platform data exchange in traditional solutions. All coupled data exchanges in the simulation system are completed on the FPGA chip with nanosecond-level access delays, ensuring the real-time performance and determinism of multi-physical field coupling. This provides a solid technical foundation for the construction of high-fidelity digital twin models of power equipment.

[0128] like Figure 5 As shown, the electromagnetic transient solver core acts as the main driving source, calculating electromagnetic force distribution data and loss distribution data in real time with microsecond-level steps, and writing the coupled data into the shared memory area on the FPGA chip. The heat conduction solver core and the mechanical vibration solver core read the loss distribution data and electromagnetic force distribution data from the shared memory area, respectively, to complete the calculation and update of the temperature field and vibration response, and write the updated temperature field parameters and vibration deformation back to the shared memory area for the electromagnetic transient solver core to read and use in subsequent steps. A global time synchronization scheduler realizes the alignment and synchronization of multiple steps. The three solver cores realize bidirectional data exchange through the on-chip shared memory area, forming a complete closed-loop feedback link of current-loss-temperature-resistivity and electromagnetic force-vibration-air gap change. All data exchange is completed inside the FPGA chip, and the delay is controlled at the nanosecond level.

[0129] Please see Figure 6 , Figure 6 This is a schematic diagram of the structure of a real-time simulation system for multiphysics coupled power equipment provided in an embodiment of the present invention. It consists of the following eight modules: electromagnetic transient solution kernel module, heat conduction solution kernel module, mechanical vibration solution kernel module, on-chip shared memory management module, global time synchronization scheduler module, external data interface module, parameter configuration and monitoring module, and result output and recording module.

[0130] The electromagnetic transient solution kernel module is one of the core computing engines of the system, responsible for the real-time solution of electromagnetic transient processes in power equipment. This module contains three sub-units: a state equation discretization unit, a matrix operation pipeline unit, and a loss calculation unit. The state equation discretization unit constructs the discretized state transition matrix and input matrix based on pre-configured equipment parameters (including inductance matrix, resistance matrix, mutual inductance coefficient, etc.) and the current resistivity parameter (obtained in real-time from the shared memory area). The matrix operation pipeline unit uses a fully pipelined architecture to implement matrix-vector multiplication, supporting real-time solution of state equations up to order 64, with a single-step calculation latency of no more than 500 nanoseconds. The loss calculation unit calculates the copper loss of each winding segment and the iron loss of each core segment based on the current distribution. Copper loss is calculated segment by segment using the I²R formula, and iron loss is calculated using a modified Steinmetz formula based on the magnetic flux density amplitude and frequency. The calculated loss distribution data is written to the electromagnetic-thermal coupling data area of ​​the shared memory area.

[0131] The heat conduction solution module is responsible for solving the equipment's temperature field in real time. This module uses the lumped-parameter thermal network method to establish a discretized thermal model of the equipment, dividing it into thermal nodes such as winding sections, core sections, insulation layers, and cooling oil channels. Each node is described by heat capacity and thermal resistance parameters. The module contains a thermal network matrix construction unit and a temperature iteration solution unit. The thermal network matrix construction unit constructs a thermal conduction matrix based on pre-configured heat capacity and thermal resistance parameters and thermal property correction coefficients at the current temperature. The temperature iteration solution unit uses either the explicit Euler method or the implicit Crank-Nicolson method to advance the temperature equation over time, reading the loss distribution as a heat source term from the shared memory area to calculate the updated temperature values ​​for each node. The updated temperature data is written to the shared memory area, where winding temperature data is used for resistivity correction calculations, and insulation temperature data is used for lifetime assessment. The temperature-to-resistivity conversion uses the linear relationship of the copper conductor temperature coefficient: ,in Reference temperature The resistivity is given by α, where α is the temperature coefficient.

[0132] The mechanical vibration solution kernel module is responsible for the real-time solution of the vibration response of the equipment structure. Based on the modal superposition method, this module decomposes the vibration response of the core structure into a linear superposition of a finite number of low-order modes. The module internally includes a modal force calculation unit and a modal response solution unit. The modal force calculation unit reads electromagnetic force distribution data from the shared storage area and calculates the generalized force of each mode through modal matrix transformation. The modal response solution unit uses the Newmark-β method to perform time integration on the single-degree-of-freedom vibration equations of each mode, calculating the displacement and velocity values ​​of each modal coordinate. Then, through inverse modal matrix transformation, it obtains the displacement and acceleration responses of each node in physical coordinates. The vibration response data is written to the shared storage area, where the core deformation is used to correct the air gap length in the electromagnetic solution kernel, and the acceleration response data is used for noise assessment.

[0133] The on-chip shared memory management module is the central hub for data exchange among the three solver cores. This module allocates three independent data sub-regions within the FPGA's Block RAM resources: an electromagnetic-thermal coupling data region stores loss distribution and temperature / resistivity data; an electromagnetic-mechanical coupling data region stores electromagnetic force distribution and core deformation data; and a global state parameter region stores device topology parameters and simulation control states. Each sub-region employs a dual-buffer structure, including a front buffer and a back buffer. Write operations are performed in the back buffer, and read operations are performed in the front buffer. Buffer switching is uniformly controlled by a global time synchronization scheduler to ensure data consistency. To guarantee deterministic timing, the shared memory preferably operates in a unified system clock domain. When different solver cores use independent clock domains, the shared memory management module uses dual-clock RAM or introduces deterministic cross-clock domain synchronization at the read / write ports (such as a handshake-enabled synchronization bridge or an asynchronous FIFO for event / pointer synchronization) to ensure that buffer switching and data visibility meet deterministic constraints.

[0134] The global time synchronization scheduler module is responsible for coordinating the runtime and data exchange timing of the three solver cores. This module includes a master clock counter, a step ratio configuration register, and a trigger signal generator. The master clock counter increments based on the time step of the electromagnetic solver core. When the count reaches the electromagnetic-thermal step ratio N, a thermal field update trigger signal is generated; when the count reaches the electromagnetic-mechanical step ratio M, a mechanical field update trigger signal is generated. The trigger signals simultaneously control the double-buffer switching operation of the shared memory area.

[0135] The external data interface module is responsible for data communication between the FPGA and external systems, including receiving externally injected power supply voltage waveforms, load disturbance signals, and ambient temperature parameters, as well as outputting simulation result data to external systems. This module supports both high-speed serial interfaces (such as Aurora) and general-purpose interfaces (such as SPI and UART) for communication.

[0136] The parameter configuration and monitoring module provides online configuration functions for equipment model parameters and simulation control parameters. It supports modifying configuration items such as equipment electromagnetic parameters, thermal property parameters, mechanical modal parameters and time step ratio through the host computer, while monitoring the running status and resource utilization of each solver core.

[0137] The results output and recording module is responsible for outputting key variables (current waveform, temperature-time curve, vibration spectrum, etc.) in a fixed format during the simulation process. It supports both real-time output and batch storage modes for subsequent condition monitoring analysis and life prediction calculations.

[0138] like Figure 6 As shown, the parameter configuration and monitoring module distributes equipment model parameters and simulation control parameters to the electromagnetic transient solution kernel module, the heat conduction solution kernel module, and the mechanical vibration solution kernel module. The three solution kernel modules exchange data bidirectionally with the on-chip shared memory management module, achieving coupled data transmission of loss-temperature-resistivity and electromagnetic force-vibration-deformation. The global time synchronization scheduler module interacts with the on-chip shared memory management module to synchronously control the trigger timing and buffer switching of the three solution kernels and obtain necessary synchronization status information. The external data interface module is responsible for communicating with external systems and receiving excitation signals and environmental parameters. The result output and recording module outputs the simulation results in a standard format to support subsequent condition monitoring and life assessment applications.

[0139] Please see Figure 7 , Figure 7 This is a structural block diagram of a real-time simulation device for multi-physics coupled power equipment provided in an embodiment of the present invention.

[0140] This invention provides a real-time simulation device for multiphysics-coupled power equipment, applied to FPGA, comprising:

[0141] The electromagnetic simulation data generation module 701 is used to obtain the resistivity correction coefficient of the previous thermal field calculation in each simulation step when a start signal is received, perform electromagnetic field calculation, and generate electromagnetic simulation data; the electromagnetic simulation data includes electromagnetic force distribution data and loss distribution data.

[0142] The mechanical simulation data generation module 702 is used to generate mechanical simulation data of the iron core structure by performing force field calculation using electromagnetic force distribution data when the counter value of the global time synchronization scheduler reaches the first preset value.

[0143] The thermal simulation data generation module 703 is used to perform thermal field calculation using the loss distribution data of the second preset value simulation step when the counter value reaches the second preset value, and generate thermal simulation data.

[0144] Output module 704 is used to output electromagnetic simulation data, mechanical simulation data and thermal simulation data.

[0145] In this embodiment of the invention, the electromagnetic simulation data generation module 701 includes:

[0146] The resistivity correction coefficient acquisition submodule is used to acquire the resistivity correction coefficient of the previous thermal field calculation in each simulation step when a start signal is received.

[0147] The updated resistance matrix generation submodule is used to generate an updated resistance matrix using resistivity correction coefficients.

[0148] The update state transition matrix and update input matrix generation submodule is used to generate the update state transition matrix and update input matrix based on the update resistance matrix.

[0149] The electromagnetic simulation data generation submodule is used to generate electromagnetic simulation data by updating the state transition matrix and updating the input matrix.

[0150] In this embodiment of the invention, the mechanical simulation data generation module 702 includes:

[0151] The modal matrix generation submodule is used to extract the mode shape vectors of the multi-mode iron core structure and form a modal matrix when the counter value of the global time synchronization scheduler reaches the first preset threshold.

[0152] The generalized force equation calculation submodule is used to calculate the generalized force equations for each mode of the iron core structure using electromagnetic force distribution data and mode matrix.

[0153] The submodule for generating updated modal coordinates and updated modal velocities is used to perform time integration on the generalized force equations of each mode using the constant acceleration method to obtain the updated modal coordinates and updated modal velocities.

[0154] The displacement calculation unit is used to generate the displacement of each node in the physical coordinates of the iron core structure by updating the modal coordinates and modal matrix;

[0155] The mechanical simulation data generation submodule is used to generate mechanical simulation data for the iron core structure based on displacement, updated modal coordinates, and updated modal velocities.

[0156] In this embodiment of the invention, the power equipment includes multiple thermal nodes; the thermal simulation data generation module 703 includes:

[0157] The average loss power calculation submodule is used to calculate the average loss power using the loss distribution data of the second preset value and simulation step size when the counter value reaches the second preset value.

[0158] The thermal simulation data generation submodule is used to perform temperature update calculations on all thermal nodes using average power loss to obtain thermal simulation data.

[0159] In this embodiment of the invention, the thermal simulation data generation submodule includes:

[0160] The heat capacity and thermal resistance acquisition unit is used to acquire the heat capacity of each thermal node and the thermal resistance between each thermal node and its adjacent thermal nodes.

[0161] The thermal field time step acquisition unit is used to acquire the thermal field time step.

[0162] Temperature data acquisition unit is used to acquire temperature data of each thermal node at the previous thermal field time step;

[0163] The current temperature data calculation unit is used to calculate the current temperature data of each thermal node at the current thermal field time step based on the temperature data, thermal field time step, heat capacity and thermal resistance.

[0164] The thermal simulation data generation unit is used to integrate the current temperature data of all thermal nodes to obtain thermal simulation data of the power equipment.

[0165] In this embodiment of the invention, the hot node includes a winding; the thermal simulation data generation submodule further includes:

[0166] The average temperature calculation unit is used to calculate the average temperature value based on the current temperature data of all windings.

[0167] The resistivity correction factor calculation unit is used to calculate the updated resistivity correction factor based on the copper conductor temperature coefficient formula and average temperature value.

[0168] In this embodiment of the invention, it further includes:

[0169] Sensor measurement data acquisition module, used to acquire sensor measurement data;

[0170] The correction module is used to compare electromagnetic simulation data, mechanical simulation data, thermal simulation data, and sensor measurement data, and correct the FPGA parameters based on the comparison results.

[0171] This invention also provides an electronic device, which includes a processor and a memory:

[0172] The memory is used to store program code and transfer the program code to the processor;

[0173] The processor is used to execute the real-time simulation method for multiphysics coupled power equipment according to the instructions in the program code of this invention.

[0174] This invention also provides a computer-readable storage medium for storing program code, which is used to execute the multiphysics-coupled power equipment real-time simulation method of this invention.

[0175] Those skilled in the art will clearly understand that, for the sake of convenience and brevity, the specific working processes of the systems, devices, and units described above can be referred to the corresponding processes in the foregoing method embodiments, and will not be repeated here.

[0176] The various embodiments in this specification are described in a progressive manner, with each embodiment focusing on the differences from other embodiments. The same or similar parts between the various embodiments can be referred to each other.

[0177] Those skilled in the art will understand that embodiments of the present invention can be provided as methods, apparatus, or computer program products. Therefore, embodiments of the present invention can take the form of entirely hardware embodiments, entirely software embodiments, or embodiments combining software and hardware aspects. Furthermore, embodiments of the present invention can take the form of computer program products implemented 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.

[0178] Embodiments of the present invention are described with reference to flowchart illustrations and / or block diagrams of methods, terminal devices (systems), and computer program products according to embodiments of the invention. It will be understood that each block of the flowchart illustrations and / or block diagrams, and combinations of blocks in the flowchart illustrations and / or block diagrams, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, special-purpose computer, embedded processor, or other programmable data processing terminal device to produce a machine, such that the instructions, which execute via the processor of the computer or other programmable data processing terminal device, generate instructions for implementing the flowchart illustrations and / or block diagrams. Figure 1 One or more processes and / or boxes Figure 1 A device that provides the functions specified in one or more boxes.

[0179] These computer program instructions may also be stored in a computer-readable storage medium that can direct a computer or other programmable data processing terminal device to operate in a particular manner, such that the instructions stored in the computer-readable storage medium produce an article of manufacture including instruction means, which are implemented in a process Figure 1 One or more processes and / or boxes Figure 1 The function specified in one or more boxes.

[0180] These computer program instructions can also be loaded onto a computer or other programmable data processing terminal equipment, causing a series of operational steps to be performed on the computer or other programmable terminal equipment to produce a computer-implemented process, thereby providing instructions that execute on the computer or other programmable terminal equipment for implementing the process. Figure 1 One or more processes and / or boxes Figure 1 The steps of the function specified in one or more boxes.

[0181] Although preferred embodiments of the present invention have been described, those skilled in the art, upon learning the basic inventive concept, can make other changes and modifications to these embodiments. Therefore, the appended claims are intended to be interpreted as including the preferred embodiments as well as all changes and modifications falling within the scope of the present invention.

[0182] It should be noted that the user information (including but not limited to user device information, user personal information, etc.) and data (including but not limited to data used for analysis, data stored, data displayed, etc.) involved in this application are all information and data authorized by the user or fully authorized by all parties. Furthermore, the collection, use and processing of the relevant data must comply with the relevant laws, regulations and standards of the relevant countries and regions, and corresponding operation entry points are provided for users to choose to authorize or refuse.

[0183] Finally, it should be noted that in this document, relational terms such as "first" and "second" are used only to distinguish one entity or operation from another, and do not necessarily require or imply any such actual relationship or order between these entities or operations. Furthermore, the terms "comprising," "including," or any other variations thereof are intended to cover non-exclusive inclusion, such that a process, method, article, or terminal device that comprises a list of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such a process, method, article, or terminal device. Without further limitations, an element defined by the phrase "comprising one..." does not exclude the presence of other identical elements in the process, method, article, or terminal device that includes said element.

[0184] The above-described embodiments are only used to illustrate the technical solutions of the present invention, and are not intended to limit it. Although the present invention has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that modifications can still be made to the technical solutions described in the foregoing embodiments, or equivalent substitutions can be made to some of the technical features. Such modifications or substitutions do not cause the essence of the corresponding technical solutions to deviate from the spirit and scope of the technical solutions of the embodiments of the present invention.

Claims

1. A real-time simulation method for multiphysics coupled power equipment, characterized in that, Applied to FPGA, the method includes: When a start signal is received, the resistivity correction coefficient from the previous thermal field calculation is obtained in each simulation step to perform electromagnetic field calculations and generate electromagnetic simulation data; the electromagnetic simulation data includes electromagnetic force distribution data and loss distribution data. When the counter value of the global time synchronization scheduler reaches the first preset value, the electromagnetic force distribution data is used to perform force field calculation and generate mechanical simulation data of the iron core structure. When the counter value reaches the second preset value, thermal field calculation is performed using the loss distribution data of the second preset value simulation step size to generate thermal simulation data. Output the electromagnetic simulation data, the mechanical simulation data, and the thermal simulation data.

2. The method according to claim 1, characterized in that, When a start signal is received, the resistivity correction coefficient from the previous thermal field calculation is obtained in each simulation step to perform electromagnetic field calculation and generate electromagnetic simulation data. The steps for obtaining electromagnetic simulation data, including electromagnetic force distribution data and loss distribution data, include: When the start signal is received, the resistivity correction coefficient of the previous thermal field calculation is obtained in each simulation step. The updated resistance matrix is ​​generated using the resistivity correction coefficient. Generate an updated state transition matrix and an updated input matrix based on the updated resistance matrix; Electromagnetic simulation data is generated using the updated state transition matrix and the updated input matrix.

3. The method according to claim 1, characterized in that, The step of generating mechanical simulation data for the core structure by performing force field calculations using the electromagnetic force distribution data when the counter value of the global time synchronization scheduler reaches a first preset value includes: When the counter value of the global time synchronization scheduler reaches the first preset threshold, the mode shape vectors of the multi-mode of the core structure are extracted to form a mode matrix; The generalized force equations for each mode of the core structure are calculated using the electromagnetic force distribution data and the modal matrix. The generalized force equations for each mode are integrated over time using the constant acceleration method to obtain the updated mode coordinates and updated mode velocities. The displacements of each node in the physical coordinates of the core structure are generated using the updated modal coordinates and the modal matrix. Mechanical simulation data of the core structure is generated based on the displacement, the updated modal coordinates, and the updated modal velocity.

4. The method according to claim 1, characterized in that, The power equipment includes multiple thermal nodes; the step of generating thermal simulation data by using loss distribution data of the second preset value and a simulation step size when the counter value reaches a second preset value includes: When the counter value reaches the second preset value, the average loss power is calculated using the loss distribution data of the second preset value simulation step size. The average power loss is used to perform temperature update calculations on all the hot nodes to obtain thermal simulation data.

5. The method according to claim 4, characterized in that, The step of using the average power loss to perform temperature update calculations on all the hot nodes to obtain thermal simulation data includes: Obtain the heat capacity of each hot node, and the thermal resistance between each hot node and its adjacent hot nodes; Obtain the time step of the thermal field; Obtain the temperature data of each thermal node at the time step of the previous thermal field; The current temperature data of each thermal node at the current thermal time step is calculated based on the temperature data, the thermal field time step, the heat capacity, and the thermal resistance. By integrating the current temperature data of all the aforementioned hot nodes, thermal simulation data of the power equipment is obtained.

6. The method according to claim 5, characterized in that, The hot node includes a winding; the method further includes: Calculate the average temperature value based on the current temperature data of all windings; The updated resistivity correction factor is calculated based on the copper conductor temperature coefficient formula and the average temperature value.

7. The method according to claim 1, characterized in that, After the step of outputting the electromagnetic simulation data, the mechanical simulation data, and the thermal simulation data, the method further includes: Acquire sensor measurement data; The electromagnetic simulation data, the mechanical simulation data, the thermal simulation data, and the sensor measurement data are compared, and the parameters of the FPGA are corrected based on the comparison results.

8. A real-time simulation device for multiphysics coupled power equipment, characterized in that, The device, applied to an FPGA, includes: The electromagnetic simulation data generation module is used to obtain the resistivity correction coefficient of the previous thermal field calculation and perform electromagnetic field calculation in each simulation step when a start signal is received, thereby generating electromagnetic simulation data; the electromagnetic simulation data includes electromagnetic force distribution data and loss distribution data. The mechanical simulation data generation module is used to perform force field calculations using the electromagnetic force distribution data when the counter value of the global time synchronization scheduler reaches a first preset value, thereby generating mechanical simulation data of the iron core structure. The thermal simulation data generation module is used to perform thermal field calculation using loss distribution data of the second preset value number of simulation steps when the counter value reaches the second preset value, and generate thermal simulation data. The output module is used to output the electromagnetic simulation data, the mechanical simulation data, and the thermal simulation data.

9. An electronic device, characterized in that, The device includes a processor and a memory: The memory is used to store program code and transmit the program code to the processor; The processor is used to execute the real-time simulation method for multiphysics coupled power equipment according to any one of claims 1-7 according to the instructions in the program code.

10. A computer-readable storage medium, characterized in that, The computer-readable storage medium is used to store program code for executing the real-time simulation method for multiphysics coupled power equipment as described in any one of claims 1-7.

Images