A small-step real-time simulation method and system for power electronic converters

By adopting a CPU-centric small-step simulation architecture in the power system, combined with external FPGA hardware for signal acquisition and transmission, the problems of low efficiency in traditional CPU simulation and high cost in FPGA simulation are solved, realizing high-precision real-time simulation of large-scale power electronic systems and improving simulation accuracy and flexibility.

CN122332337APending Publication Date: 2026-07-03CHENGDU MITT INTELLIGENT POWER TECH CO LTD
View PDF 0 Cites 0 Cited by

Patent Information

Application Number
CN202610269799.7
Authority / Receiving Office
CN · China
Patent Type
Applications(China)
Current Assignee / Owner
Filing Date
2026-03-06
Publication Date
2026-07-03

AI Technical Summary

Technical Problem

Existing technologies struggle to achieve high-precision real-time simulation of power electronic converters in power systems, especially at high switching frequencies. Traditional CPU simulation is limited by computational efficiency, FPGA simulation is costly and lacks flexibility, and hybrid simulation schemes suffer from communication overhead and synchronization delay issues.

Method used

A CPU-centric small-step simulation architecture is adopted, combined with external FPGA hardware for high-speed signal acquisition and transmission. The CPU performs small-step calculations within the main step size, and the electrical dynamics of high-switching-frequency devices are equivalently reproduced through software algorithms. The converter model and the external power grid model are integrated to achieve a balance between high simulation accuracy and scale.

Benefits of technology

It enables high-precision real-time simulation of large-scale power electronic systems on a single CPU platform with high efficiency and low cost, avoiding communication bottlenecks in hybrid simulation and improving simulation accuracy and flexibility.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN122332337A_ABST
    Figure CN122332337A_ABST
Patent Text Reader

Abstract

The application belongs to the technical field of power electronic real-time simulation, and aims at the problems of insufficient simulation accuracy of traditional pure CPU simulation, limited simulation scale of FPGA simulation, communication bottleneck and poor flexibility of "CPU+FPGA" hybrid simulation, and proposes a power electronic converter small step real-time simulation method and system. The core is to build a multi-port equivalent model based on the switching function method or the improved node method, and to equivalent the nonlinear part of the converter into a controlled source and fuse the linear passive network; adopt the "CPU unique calculation + FPGA special communication" architecture, FPGA nanosecond level acquisition PWM pulse sequence transmission to CPU, CPU through the main step nesting small step recursive calculation, in the microsecond level main step, reproduce the hundred nanosecond level electrical dynamic. The matching special system deeply optimizes the CPU multi-core parallel ability, the communication cost is low, realizes the accuracy, the efficiency and the scale unification, the cost is low, the flexibility is high, can support the large-scale high switching frequency system high precision simulation, is applicable to the new energy grid connection, direct current transmission and other scenes.
Need to check novelty before this filing date? Find Prior Art

Description

Technical Field

[0001] This invention belongs to the field of real-time simulation technology of power electronics, specifically relating to a small-step method and system for high-precision real-time simulation of power electronics, which is particularly suitable for the simulation of large-scale complex power grid systems containing high-switching-frequency power electronic equipment. Background Technology

[0002] In modern power systems such as renewable energy grid connection and DC transmission, the high-frequency switching characteristics of power electronic converters require electromagnetic transient simulations with step sizes in the microsecond or even sub-microsecond range to accurately capture their transient processes. Traditional pure CPU simulations are limited by operating system scheduling and computational efficiency, making it difficult to stably achieve real-time simulation step sizes of less than 5 microseconds, thus failing to meet high-precision requirements. While FPGA-based simulations can achieve nanosecond-level step sizes, they are complex to develop, costly, and limited by FPGA logic resources, resulting in a very limited scale of systems that can be simulated.

[0003] In existing technologies, the "CPU+FPGA" hybrid simulation scheme typically uses an FPGA to perform calculations on the entire power electronic converter or the fast dynamic part. This method has significant drawbacks: First, the floating-point computing power of an FPGA is far lower than that of a modern multi-core CPU, and the complexity of implementing large-scale systems on an FPGA increases exponentially; second, the frequent exchange of a large number of state variables (such as all inductor currents and capacitor voltages) between the CPU and the FPGA brings huge communication overhead and synchronization delays, becoming a bottleneck for simulation speed; third, the model is hardware-bound, lacks flexibility, and is difficult to quickly adapt to different topologies.

[0004] Therefore, an innovative simulation architecture is urgently needed to overcome the limitations of existing hybrid simulation schemes. Its core lies in: fully leveraging the powerful floating-point computing capabilities of the CPU to handle large-scale linear network equations, while simultaneously designing a simulation algorithm with nested small steps within a main step size. This allows the CPU to perform equivalent calculations of the electrical dynamics within the power electronic converter at the nanosecond level within its main simulation step size (e.g., 5 microseconds), based on externally acquired refined switching sequences, ultimately achieving a balance between accuracy, efficiency, and simulation scale. Summary of the Invention

[0005] This invention aims to solve the aforementioned technical problems by proposing a small-step real-time simulation method and system for power electronic converters. The core of this method is to construct a CPU-centric small-step simulation framework: the CPU acts as the sole core computing engine, integrating refined switching information from external sources within its main simulation step. Through efficient algorithms, it equivalently reproduces the high-frequency electrical dynamics of high-switching-frequency devices and solves the network equations of the entire large-scale system.

[0006] The first aspect of this invention provides a standardized multi-port equivalent modeling method for power electronic converters, laying the foundation for small-step simulation. The nonlinear components of each converter, including switching devices, are uniformly equivalently represented as a multi-port controlled source driven by real-time switching states, using a switching function model or an improved nodal method. This controlled source can accurately reproduce the modulation effect of switching actions on AC and DC side electrical quantities. This controlled source is combined with the converter's linear passive network (such as filter inductors and capacitors) to form a continuously linear time-varying model solvable within each CPU main step. Through clearly defined electrical ports, this model can seamlessly interconnect with external power grid models.

[0007] The second aspect of this invention provides a small-step simulation architecture and process, which is the core of this invention. This architecture includes the following key steps: Ultra-high-speed signal acquisition and transmission: Utilizing external hardware (such as an FPGA) with precise timing characteristics, PWM pulse signals from an external controller are acquired in real time with extremely small fixed step sizes (e.g., 100 nanoseconds). This hardware does not perform any simulation calculations; it is only responsible for packaging and transmitting the pulse sequence (e.g., forming a 64-bit bitstream) acquired within a CPU main step size cycle (e.g., 5 microseconds) to the CPU. CPU main step size task: The CPU periodically runs simulation tasks with a fixed main step size (e.g., 5 microseconds), including solving the network equations of the entire power system and updating the grid-side voltage, current, and other states. CPU small-step calculation: Within each main step size cycle, the CPU receives the pulse sequence from the external hardware. Based on the equivalent converter model, the CPU recursively calculates the detailed electrical dynamics of the converter under multiple switching states using equivalent small step sizes internally. This process is entirely implemented within the CPU using software algorithms. Data Integration and Output: The CPU averages and integrates the effective changes in converter port voltage and current calculated in small steps within a main step, using this as the simulation result for that main step. Finally, the CPU sends the electrical quantities (voltage / current) obtained from the simulation calculation to the FPGA high-speed communication board, which then feeds back the data to the physical control device through its analog / digital interface at the main step cycle, thus achieving simulation closed loop.

[0008] A third aspect of this invention provides a real-time simulation system for implementing the above-described method. This system includes real-time simulation software running on a high-performance multi-core CPU, and external hardware (FPGA board) dedicated to high-speed data acquisition. The software architecture is deeply optimized, fully utilizing the CPU's multi-core parallel capabilities and real-time operating system features to handle concurrent tasks such as parallel solving of large-scale network equations and small-step calculations of numerous converters. Since the CPU handles all calculations and only transmits compressed pulse sequences to the acquisition hardware, communication overhead is extremely low, enabling this system to achieve real-time simulation of large-scale power electronic systems on a single CPU platform, far exceeding the capabilities of traditional FPGA solutions.

[0009] The beneficial effects of this invention are as follows: It creatively proposes a small-step simulation architecture with the CPU as the sole computing core, strictly limiting the function of external FPGA hardware to high-speed PWM signal acquisition and transmission, rather than undertaking any simulation calculation tasks. By fusing nanosecond-level switching transient details within the CPU's main step size through software algorithms, a balance between simulation accuracy and computational efficiency is achieved. This method avoids the cumbersome data synchronization problems in hybrid simulations, fully leverages the general computing advantages of the CPU, and provides a lower-cost, more flexible, and high-precision real-time power electronics simulation solution capable of handling ultra-large-scale systems. Attached Figure Description

[0010] To more clearly illustrate the technical solutions of the embodiments of the present invention, the accompanying drawings used in the description of the embodiments will be briefly introduced below. It should be understood that the following drawings only show some embodiments of the present invention and should not be considered as a limitation on the scope of protection. For those skilled in the art, other related drawings can be obtained from these drawings without creative effort.

[0011] Figure 1 This is a CPU-based small-step simulation system architecture diagram, which shows a closed-loop simulation system consisting of a high-performance multi-core CPU, an FPGA high-speed communication board, and physical control devices.

[0012] Figure 2 This diagram illustrates the coordination between external hardware signal acquisition and the CPU's main step size cycle, demonstrating the timing relationship between the FPGA acquiring PWM signals in nanosecond-level steps and transmitting pulse sequences within the CPU's main step size.

[0013] Figure 3 A detailed computation flowchart of the CPU small-step simulation method within a main step is provided, showing the complete process from data reception, small-step recursive computation to main-step network solution.

[0014] Figure 4 This diagram illustrates the interconnection between the multi-port equivalent model of a power electronic converter and an external system, showcasing the bridge arm topology of a two-level converter and the equivalent model of the controlled source driven by its switching function.

[0015] Figure 5 The diagram shows an example of engineering testing for a small-step simulation architecture. Through comparison of three typical scenarios, the core advantages of this invention in terms of accuracy, efficiency, and large-scale system simulation are verified. Detailed Implementation

[0016] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only a part of the embodiments of the present invention, and not all of them. All other embodiments obtained by those skilled in the art based on the embodiments of the present invention without creative effort are within the scope of protection of the present invention.

[0017] See Figure 1 The simulation system architecture provided by this invention mainly consists of a high-performance multi-core CPU 11, an FPGA high-speed communication board 12, and a physical control device 13. The high-performance multi-core CPU 11 undertakes all computational tasks under the small-step simulation architecture and is the system's sole core computing engine, running the real-time operating system and the simulation algorithm of this invention. The FPGA high-speed communication board 12 is connected to the high-performance multi-core CPU 11 via a PCIe high-speed bus and possesses nanosecond-level precision timing sampling capability. Its core function is to sample all high-speed PWM pulse signals from the physical control device 13 at a fixed high rate of 100 nanoseconds and perform signal reconstruction and synchronization operations. Subsequently, the pulse sequence acquired within each CPU main step cycle (5µs) is compressed and transmitted to the high-performance multi-core CPU 11 via the PCIe high-speed bus. The FPGA board itself does not participate in any power system-related calculations. The physical control device 13 is an inverter controller responsible for generating high-speed PWM pulse signals. The specific signal flow of the system is as follows: The high-performance multi-core CPU 11 sends the electrical quantities (voltage / current) obtained from the simulation calculation to the FPGA high-speed communication board 12 at a period of 5µs; the FPGA high-speed communication board 12 feeds back the signal to the physical control device 13 through its analog / digital interface; the high-speed PWM pulse signal (100ns) generated by the physical control device 13 is acquired by the FPGA high-speed communication board 12 at high speed; the FPGA high-speed communication board 12 transmits the compressed pulse sequence to the high-performance multi-core CPU 11 through the PCIe high-speed bus at a period of 5µs, thus forming a closed loop.

[0018] See Figure 2 This illustrates the step-size coordination relationship between the CPU and the external acquisition unit. The CPU simulation uses a fixed main step size Δ. T CPU (For example, 5 microseconds) cycle. In each Δ T CPU During the period, the FPGA high-speed communication board acquires data in steps Δ T FPGA The PWM signal is continuously sampled (e.g., 100 nanoseconds). Let... N =Δ T CPU / Δ T FPGA(For example, 5μs / 100ns=50), then the FPGA high-speed communication board in each Δ T CPU During the cycle, a pulse sequence bitstream of length N bits (e.g., 50 bits) is generated for each PWM channel. At the start of the CPU's main step cycle, the FPGA high-speed communication board sends this bitstream data packet to the CPU all at once.

[0019] See Figure 3 This document describes in detail the small-step simulation process within a CPU main step cycle. This process is entirely executed by the simulation engine on the CPU. The specific steps are as follows:

[0020] S301: Main step size calculation initiated. The simulation engine is triggered by a high-precision real-time clock to begin the i-th main step size (Δ). T CPU The calculation cycle is adjusted to ensure that the simulation timing is strictly synchronized with the hardware clock.

[0021] S302: High-frequency data reception and parsing. The CPU receives a compressed pulse sequence bit stream from the FPGA high-speed communication board via the PCIe high-speed bus, completing data reception and parsing. This data packet contains all PWM channels within the current main step size cycle, acquired by the FPGA at a step size (Δ). T FPGA This provides a precise history of switching events at resolution, offering raw data for subsequent small-step calculations.

[0022] S303: Small step size calculation initialization. Entering the small step size calculation phase, starting "step j small step size calculation", initializing the small step size index j=1, preparing to begin calculation based on Δ. T FPGA Recursive fine calculation.

[0023] S304: Switching state sequence analysis. For the current j-th small step, extract all converter states in the Δ state from the analyzed pulse sequence. T FPGA The combination of switching states within the time slice clarifies the on / off state of the switching transistors of each converter, providing a basis for the switching function in subsequent equivalent model calculations.

[0024] S305: Boundary condition recursive solution. Based on the converter port voltage obtained in the previous main step, the port voltage boundary conditions of the equivalent converter model under the current small step are recursively obtained through a linear extrapolation algorithm, providing initial support for the iteration of internal state variables.

[0025] S306: Iterative solution of internal state variables of the converter. Based on the standardized multi-port equivalent circuit model of the converter, under the current fixed switching state, with Δ... T FPGAUsing the step size, the instantaneous values ​​of state variables such as inductor current and capacitor voltage inside the converter are recursively and iteratively solved by numerical integration methods such as the backward Euler method.

[0026] S307: Equivalent Controlled Source Parameter Mapping. Based on the current small-step switching state function or the improved nodal method, the iteratively obtained internal state variables are mapped to the instantaneous parameters of the controlled voltage / current sources in the equivalent circuit of the converter, thus completing the calculation of the equivalent model parameters under small-step conditions.

[0027] S308: Small step size loop judgment and termination. After completing the calculation of the j-th small step size, check whether all loops within the current main step size have been traversed. N Small step size ( N =Δ T CPU / Δ T FPGA For example, 5μs / 100ns = 50). If j < N If the condition is met, then set j = j + 1 and return to the "Switch Sequence Analysis" step to continue the calculation of the next small step; if all small steps have been completed, then end the "Small Step Subnet Calculation".

[0028] S309: Equivalent parameter averaging solution for converter. The instantaneous parameters of the controlled source calculated by all N small steps within a main step period are weighted and averaged to obtain the equivalent parameters of the converter at the main step level, which are used for interaction with the external main grid model.

[0029] S310: Solving the main step-size network equations. Integrate the equivalent admittances of all converters and linear network elements (such as lines and transformers) to construct a global constant admittance matrix [G]. Combine this with the equivalent injected current source [I] of the converters and solve the equation [G]×[U]=[I] using the nodal voltage method to update the nodal voltages of the entire network.

[0030] S311: Data Type Conversion and Hardware-in-the-Loop Communication. The grid state variables and converter port variables obtained from the main step-size network solution are converted into hardware-in-the-loop interface-compatible data types. This data is then sent to the FPGA high-speed communication board via Aurora high-speed serial communication, enabling real-time interaction between the simulation system and external physical control devices.

[0031] S312: Simulation cycle judgment and stepping. Determine whether the simulation task has ended: if not, set i=i+1, return to "Calculate the main step size of step i" and start the next cycle; if completed, end the simulation process, output the result data file, and complete the entire simulation loop.

[0032] See Figure 4The converter multi-port equivalent model of this invention takes a two-level converter bridge arm as a typical object, and achieves accurate equivalence of nonlinear switching behavior through a controlled source network driven by a switching function. Its technical implementation is deeply coupled with the small step size simulation process. Figure 4 (a) is an example of the physical topology of a two-level converter bridge arm, with the core switching network consisting of upper and lower bridge arm IGBTs and anti-parallel diodes. The AC side is filtered by an inductor. L Connect to grid node V abc The DC side is connected to the positive and negative busbars. V p , V n . Figure 4 (b) is an example of the equivalent effect of this bridge arm. Figure 4 The IGBT switching network in (a) is equivalent to one controlled voltage source and two controlled current sources. V con The output is driven by the switching function m, satisfying... V con =m· V p +(1−m)· V n Where m is the upper bridge arm switching function (0 / 1 corresponds to turn-off / turn-on), directly replicating the modulation effect of the switching action on the AC side voltage; controlled current source I p , I n The output is determined by the AC side inductor current. I abc Mapped to the switching function m, it satisfies I p =m· I abc , I n =(1-m)· I abc , respectively, represent the equivalent current flowing through the positive and negative DC busbars.

[0033] In the small-step simulation process, the equivalent model operates as follows: During the small-step calculation stage, the PWM pulse sequence acquired by the FPGA is parsed into a switching function m, which is substituted into the controlled source equation to obtain the instantaneous parameters; based on the backward Euler method, the inductor current is solved iteratively with an equivalent small step size. I abc capacitor voltage V p , V n The state variables are mapped to controlled source instantaneous parameters, thus fully reproducing the transient characteristics of the switch.

[0034] During the main step size calculation stage, the controlled source parameters within the small step size are weighted and averaged to obtain the equivalent injection currents Ip and In at the main step size level, and the equivalent controlled voltage source. V con By integrating these equivalent currents with the admittances of the AC-side inductance and DC-side capacitance, a global node voltage equation [G]×[U]=[I] is constructed, enabling seamless interconnection and solution between the converter model and the external power grid.

[0035] See Figure 5 The small-step simulation architecture of this invention has been fully verified in engineering applications through three typical test scenarios, demonstrating its core advantages in accuracy, efficiency and scalability.

[0036] Figure 5 (a) is an engineering test example of a two-resistor converter. The converter is simulated using a traditional two-resistor average model. The modulation wave frequency is selected as 10kHz (corresponding to a switching period of approximately 100μs), and the engineering simulation step size is 20μs (corresponding to Δ). T CPU This model uses on-resistance. R on Turn-off resistor R off While accurately describing the switching action, the simulation step size is relatively large compared to the current switching cycle, resulting in the loss of transient details of high-frequency switching. This leads to obvious glitches and distortions in the simulated current waveform, making it impossible to accurately capture the rapid electrical dynamics inside the converter and failing to meet the accuracy requirements of high switching frequency scenarios.

[0037] Figure 5 (b) is an example of engineering testing for a small-step converter, using the small-step simulation architecture of this invention. The engineering simulation step size (corresponding to Δ) T CPU The time interval is 20 μs, with an equivalent small step size of 1 μs (corresponding to Δ). T FPGA The converter is subjected to fine calculations. The step size can accurately capture the nanosecond-level transient process of the switching device. The internal state variables of the converter are solved recursively and iteratively in the CPU with a small step size. The simulated current waveform is smooth and distortion-free, thus accurately reproducing the switching transient characteristics of the converter while maintaining high computational efficiency. This verifies the feasibility and accuracy advantages of small step size calculation.

[0038] The above three types of engineering test scenarios, from improving the accuracy of a single converter to verifying the scalability of large-scale systems, fully demonstrate the core value of this invention in engineering applications: it not only ensures the simulation accuracy of nanosecond-level switching transients, but also realizes efficient real-time simulation of large-scale power electronic systems, providing reliable technical support for engineering testing in fields such as new energy grid connection and DC power transmission.

Claims

1. A small step real-time simulation method for power electronic converters, characterized in that, include: Construct a standardized multi-port equivalent model for power electronic converters; A CPU-based method for solving electrical networks with nested main step sizes is adopted. A small-step hardware-in-the-loop simulation architecture and process based on CPU computing and FPGA communication is adopted. The method is adapted to the high-precision simulation requirements of large-scale, high-switching-frequency power electronic systems.

2. The method of claim 1, wherein the method further comprises: The construction of the model specifically includes: Accurate equivalence of nonlinear switching behavior can be achieved through controlled source networks driven by switching functions or based on improved nodal methods. The IGBT switching network is equivalent to one controlled voltage source and two controlled current sources. V con The output is driven by the switching function m, satisfying... V con =m· V p +(1-m)· V n Where m is the upper bridge arm switching function (0 / 1 corresponds to turn-off / turn-on), directly replicating the modulation effect of the switching action on the AC side voltage; controlled current source I p , I n The output is determined by the AC side inductor current. I abc Mapped to the switching function m, it satisfies I p =m· I abc , I n =(1-m)· I abc , respectively, represent the equivalent current flowing through the positive and negative DC busbars; The controlled source network is combined with the linear passive network of the converter (including filter inductor, support capacitor, and line impedance) to form a continuous linear time-varying model; standardized AC and DC ports are defined for the model to achieve seamless interconnection with the external power grid model.

3. The method for solving electrical networks using CPU-based main step size nested small step size calculations as described in claim 1, characterized in that, The method is executed independently by a simulation engine on the CPU, progressing cyclically according to the main step size cycle, and specifically includes: The simulation engine is triggered by a high-precision real-time clock, starting the i-th major step (Δ). T CPU The calculation cycle is adjusted to ensure that the simulation timing is strictly synchronized with the hardware clock. The CPU receives and parses the compressed pulse sequence bitstream from the FPGA via the PCIe high-speed bus. This data packet contains all PWM channels within the current main step size, and is acquired by the FPGA using a step size (Δ). T FPGA () for a precise history of on / off events at resolution; Entering the small step size calculation stage, initialize the small step size index j=1, and prepare based on Δ T FPGA Perform recursive fine calculations; For the current j-th small step, extract all the switching state combinations of the converters within the ΔTFPGA time slice from the parsed pulse sequence, clarify the on / off state of the switching transistors, and obtain the basis for the switching function; Based on the converter port voltage of the previous main step, the port voltage boundary conditions of the converter equivalent model under the current small step are recursively derived by the linear extrapolation algorithm. In the current fixed switch state, with Δ T FPGA Using the step size, the instantaneous values ​​of state variables such as inductor current and capacitor voltage inside the converter are solved by recursive iteration using the backward Euler method. Based on the current small-step switching state function, the iteratively obtained internal state variables are mapped to the instantaneous parameters of the controlled voltage source / controlled current source in the converter equivalent circuit; Determine whether all N small step lengths within this main step length have been traversed (N = Δ T CPU / Δ T FPGA ), if j < N, update j and return to the switch state parsing step, if all have been traversed, end the small step subnet calculation; The equivalent parameters of the main step stage converter are obtained by averaging the instantaneous parameters of the controlled source in N small steps within a main step. The equivalent admittance of all converters and linear network elements is integrated to construct a global constant admittance matrix [G]. Combined with the equivalent injected current source [I] of the converter, the equation [G]×[U]=[I] is solved by the nodal voltage method to update the nodal voltage of the entire network. The grid state variables and converter port variables are converted into hardware-in-the-loop interface compatible data types and sent to the FPGA via Aurora high-speed serial communication. Determine if the simulation task has ended. If not, set i = i + 1 and return to the main step size calculation start step. If completed, end the simulation process and output the result data file.

4. The small-step hardware-in-the-loop simulation architecture and process based on CPU computing and FPGA communication as described in claim 1, characterized in that, The architecture consists of a closed loop comprised of a high-performance multi-core CPU, an FPGA high-speed communication board, and physical control devices. The communication process and step size are strictly matched, specifically including: The CPU operates with a fixed microsecond-level main step size Δ T CPU The simulation task is run periodically, and the FPGA high-speed communication board acquires data in nanosecond-level steps Δ. T FPGA The PWM signal is continuously sampled, and the two satisfy the step size coordination relationship N=Δ T CPU / Δ T FPGA (N is a positive integer); In each Δ T CPU During the cycle, the FPGA generates a pulse sequence bit stream of length N for each PWM channel. At the beginning of the CPU main step cycle, the bit stream data packet is transmitted to the CPU in one go through the PCIe high-speed bus to ensure that the data transmission is synchronized with the CPU main step timing. The FPGA is only responsible for high-speed PWM pulse signal acquisition, pulse sequence compression, bit stream transmission and simulation result feedback. It does not participate in any power system-related simulation calculations. All simulation tasks are completed independently by the CPU. After the CPU converts the processed grid state variables and converter port variables into hardware-in-the-loop interface compatible data types, it sends them to the FPGA via Aurora high-speed serial communication. After receiving data, the FPGA feeds back the signal to the physical control device through its analog / digital interface. The physical control device adjusts the output high-speed PWM pulse signal according to the feedback signal, forming a complete hardware-in-the-loop simulation closed loop.

5. A real-time simulation system for small-step power electronic converters that implements the method of any one of claims 1-4, characterized in that, The system is a dedicated hardware and software integration adapted for high-precision simulation of large-scale, high-switching-frequency power electronic systems, specifically including: Hardware components: Composed of a high-performance multi-core CPU, an FPGA high-speed communication board, and a physical control device; the high-performance multi-core CPU serves as the system's sole core computing engine, while the FPGA high-speed communication board connects to the CPU via a PCIe high-speed bus and only performs high-speed PWM pulse signal acquisition, pulse sequence compression, bit stream transmission, and simulation result feedback functions, without participating in any power system-related simulation calculations; the physical control device is an inverter controller. Software component: It is a real-time simulation software that runs on a high-performance multi-core CPU. Its architecture has been deeply optimized to make full use of the CPU's multi-core parallel capabilities and the characteristics of the real-time operating system. It can handle concurrent tasks such as solving large-scale network equations and calculating a large number of small steps of converters in parallel. In this system, only compressed pulse sequence bit streams and simulation result data are transmitted between the CPU and the FPGA, resulting in extremely low communication overhead. This enables real-time simulation of power electronic systems on a single CPU platform that far exceeds the scale of traditional FPGA solutions.