Power grid transient process deduction method based on digital twinning, electronic device and medium

Abstract

The application discloses a power grid transient process deduction method based on digital twinning, an electronic device and a medium, and the method comprises the following steps: constructing a digital twinning simulation model based on a new energy single-machine simulation model and a three-dimensional field station virtual model, and outputting a virtual value from the digital twinning simulation model; collecting core parameters in a monitoring period; calculating a deviation rate of a physical value and the virtual value of a physical field station based on the core parameters; and adjusting a transient process control strategy based on an adjustment object and the deviation rate, wherein the adjustment object comprises the virtual value, the core parameters, scene parameters and a model control mode. Through closed-loop correction of the deviation rate of the physical value and the virtual value, the digital twinning model can be calibrated in real time, the simulation of the power grid transient process can be ensured to be highly consistent with the physical actuality, and the accuracy and reliability of the deduction result are improved.

Description

Technical Field

[0001] This invention relates to the field of power grid transient analysis technology, specifically to a method, electronic equipment, and medium for extrapolating power grid transient processes based on digital twins. Background Technology

[0002] Currently, the proportion of installed capacity from new energy sources is rapidly increasing. The power output of renewable energy sources such as photovoltaics and wind power is greatly affected by weather conditions: short-term cloud cover can cause photovoltaic power plant output to drop by more than 30% within seconds; wind power may experience voltage flicker and frequency oscillations due to gusts or turbulence. These millisecond-second transient processes directly threaten the voltage / frequency stability of the power grid, necessitating high-precision simulations to assess risks in advance and formulate transient process control strategies.

[0003] Among related technologies, digital twin technology is considered a key means to solve the above problems due to its "real-time mapping-closed-loop optimization" capability. Its mainstream implementation scheme is as follows: first, an electromagnetic transient model is constructed based on the primary topology and factory parameters of the site; then, through a limited number of field tests or offline identification of historical data, control and protection strategies, inverter / wind turbine characteristics, etc., are abstracted into fixed parameters and written into the twin; finally, real-time calculations are performed with a fixed step size to complete transient simulation and fault prediction.

[0004] However, long-term field operation has revealed that this "fixed parameter" twin architecture has the following inherent defects: equipment aging leads to parameter drift; and the environment and operating conditions are constantly changing. Therefore, traditional digital twin technology exhibits a vicious cycle of "rapid model inaccuracy—accumulation of extrapolation errors—failure of transient process control strategies" in high-proportion renewable energy power grids, becoming a bottleneck restricting the accuracy of transient stability analysis and fault handling. Summary of the Invention

[0005] The purpose of this invention is to improve the accuracy of transient stability analysis and fault handling by adjusting the transient process control strategy by calculating the deviation rate between physical and virtual values. This is a digital twin-based method for power grid transient process simulation.

[0006] To achieve the above objectives, this invention proposes a method for extrapolating power grid transient processes based on digital twins, comprising: constructing a digital twin simulation model based on a new energy single-machine simulation model and a three-dimensional power station virtual model, wherein the digital twin simulation model outputs virtual values; collecting core parameters during the monitoring period, wherein the core parameters include 35kV bus voltage, grid-connected frequency, total active power, reactive power, inverter output current, transformer temperature, SVG reactive power output, illuminance, wind speed, ambient temperature, simulated voltage, simulated power, and simulated reactive power output; calculating the deviation rate between the physical values ​​of the physical power station and the virtual values ​​based on the core parameters; and adjusting the transient process control strategy based on the adjustment object and the deviation rate, wherein the adjustment object includes the virtual values, the core parameters, scenario parameters, and model control mode.

[0007] In one optional implementation, a digital twin simulation model is constructed based on a new energy single-unit simulation model and a three-dimensional station virtual model. Specifically, this includes: constructing a single-unit twin model based on the new energy single-unit simulation model; correcting the multiplication factor based on the single-unit twin model to obtain a real-time correction factor; feeding back the real-time correction factor to the single-unit twin model to obtain a corrected single-unit twin model; constructing a three-dimensional station virtual model at a 1:1 physical topology; and constructing the digital twin simulation model based on the three-dimensional station virtual model and the corrected single-unit twin model.

[0008] In one optional implementation, the transient process control strategy is adjusted based on the adjustment object and the deviation rate, specifically including: dividing the adjustment range based on the deviation rate, the adjustment range including a normal range, a warning range and an abnormal range; adjusting the corresponding triggering conditions based on the adjustment range; and performing corresponding operation steps on the adjustment object based on the constraint conditions and the triggering conditions, the operation steps including correcting the virtual value, correcting the core parameters, correcting the scene parameters and switching the model control mode.

[0009] In one optional implementation, the correspondence between the deviation rate and the adjustment range satisfies the following: δ≤3% is the normal range; 3%<δ≤5% is the warning range; δ>5% is the abnormal range; where δ represents the deviation rate.

[0010] In one optional implementation, the physical values ​​of the physical station and the virtual values ​​are time-stamped to obtain standard data, wherein the synchronization accuracy of the standard data is ≤ ±1ms.

[0011] In one optional implementation, the three-dimensional station virtual model is a three-dimensional virtual model constructed with differentiated precision based on the core area, related areas, and peripheral areas.

[0012] In an optional implementation, the digital twin-based power grid transient process simulation method further includes: constructing a dynamic reproduction and prediction model of the transient process based on time and space dimensions; mapping dynamic data to the dynamic reproduction and prediction model of the transient process to generate a dynamic simulation screen, wherein the dynamic data includes time window control, electrical quantity rendering, trajectory generation and transient process control strategy.

[0013] In an optional implementation, the digital twin-based power grid transient process simulation method further includes: based on the interactive control unit and the three-dimensional station virtual model, controlling the dynamic reproduction and prediction model of the transient process to generate a simulation effect and comparison report.

[0014] The present invention also proposes an electronic device, comprising: at least one processor; and a memory communicatively connected to the at least one processor; wherein the memory stores instructions executable by the at least one processor, the instructions being executed by the at least one processor to enable the at least one processor to execute any of the digital twin-based power grid transient process simulation methods described herein.

[0015] The present invention also proposes a medium storing a computer program, which, when executed by a processor, implements any of the methods for power grid transient process deduction based on digital twins.

[0016] The beneficial effects of this invention are as follows: through closed-loop correction of the deviation rate between physical and virtual values, the digital twin model can be calibrated in real time, which can ensure that the simulation of the transient process of the power grid is highly consistent with the physical reality, thereby improving the accuracy and reliability of the inference results. Attached Figure Description

[0017] Figure 1 A flowchart illustrating a power grid transient process extrapolation method based on digital twins, provided in an embodiment of the present invention;

[0018] Figure 2 A flowchart of a power grid transient process simulation method based on digital twins provided in another embodiment of the present invention. Detailed Implementation

[0019] Accurate transient process simulation is crucial for fault handling. Current transient simulation technology suffers from the following shortcomings: 1. Disconnect between physical and virtual data: Physically measured data from stations, synchronized phasor measurement devices (millisecond-level) and SCADA systems (second-level) have significant timescale discrepancies with virtual simulation data, resulting in insufficient fusion accuracy and persistently high simulation errors; 2. Lack of visualization: Transient processes are only displayed through two-dimensional curves, failing to intuitively present the spatial distribution and temporal evolution of "equipment-area-station," making fault location reliant on manual experience and time-consuming. Therefore, this invention proposes a transient simulation method with virtual-physical collaboration, adaptive modeling, and panoramic visualization, meeting the core requirements of new energy operation and maintenance.

[0020] This invention aims to provide a panoramic simulation method for transient processes of new energy based on digital twins, which solves the problems of asynchronous virtual and real data, poor model adaptability, insufficient visualization and weak adaptability in the existing technology, and achieves the technical goal of "high-precision simulation, dynamic adaptation and intuitive presentation" of transient processes.

[0021] The present invention will be further described in detail below with reference to the accompanying drawings and specific embodiments.

[0022] like Figure 1 and Figure 2 As shown in the embodiments of the present invention, in one aspect, a method for extrapolating power grid transient processes based on digital twins is provided, comprising the following steps:

[0023] Step S101: Construct a digital twin simulation model based on the new energy single-machine simulation model and the three-dimensional station virtual model, and output virtual values ​​from the digital twin simulation model.

[0024] Step S103: Collect core parameters during the monitoring period. Core parameters include 35kV bus voltage, grid-connected frequency, total active power, reactive power, inverter output current, transformer temperature, SVG reactive output, illuminance, wind speed, ambient temperature, simulated voltage, simulated power, and simulated reactive output.

[0025] Step S105: Calculate the deviation rate between the physical and virtual values ​​of the physical station based on the core parameters.

[0026] Step S107: Adjust the transient process control strategy based on the adjustment object and the deviation rate. The adjustment objects include virtual values, core parameters, scenario parameters and model control modes.

[0027] In this embodiment, a digital twin simulation model is constructed using the physical entity of a new energy power generation station as a prototype. This model integrates a single-unit simulation model of the new energy unit and a three-dimensional virtual model of the station. The single-unit simulation model includes detailed models of equipment such as photovoltaic inverters and wind turbine converters, encompassing parameters such as equipment topology, control logic, and electrical characteristics. The three-dimensional virtual model covers system-level models such as the overall station layout, power collection lines, substations, and reactive power compensation devices. A digital twin simulation model is built using data mapping and modeling techniques. This digital twin simulation model can simulate the real-time operating state of the physical station, outputting virtual values ​​corresponding to the physical entity, such as virtual bus voltage and virtual power output. It can also digitally mirror and simulate transient processes in the power grid, such as voltage fluctuations, frequency spikes, and fault transients.

[0028] Set a fixed monitoring cycle, such as milliseconds or seconds, and dynamically adjust it according to the response speed of the transient process. Collect core physical parameters in real time through sensors, smart terminals and data acquisition systems deployed in physical sites. At the same time, collect corresponding simulation parameters, such as virtual values ​​of simulation voltage, simulation power and simulation reactive power output, from the digital twin simulation model.

[0029] The data acquisition process must ensure that the timestamps of the data are aligned to guarantee the spatiotemporal consistency between physical and virtual values.

[0030] Furthermore, the method for extrapolating power grid transient processes based on digital twins also includes the following steps: time-stamping the physical and virtual values ​​of physical stations to obtain standard data, such that the synchronization accuracy of the standard data is ≤ ±1ms.

[0031] The timescale calibration uses the GPS timescale of the synchronized phasor measurement device as the benchmark, calculates the timescale deviation Δt between the physical data and the virtual data, and corrects it through dynamic interpolation: corrected timescale = original timescale - Δt × ω, where ω is the confidence weight, synchronized phasor measurement device = 0.95, SCADA = 0.8, virtual simulation = 0.85. After calibration, the synchronization accuracy is ≤ ±1ms, which can meet the requirements of transient analysis.

[0032] The physical measured data (physical values) collected and the virtual simulation data output by the digital twin simulation model, i.e., virtual values, are coupled and stored. The coupling storage method is to store them in the format of "timestamp + physical value + virtual value + deviation rate", such as "2025-09-0110:00:00.000,220V,218V,0.9%", which can be divided into "5s before transient / during transient / 5s after transient" and supports querying by time range.

[0033] Among them, a time-scale calibration unit can be constructed to perform time-scale calibration on physical and virtual values; and a coupled storage unit can be constructed to perform coupled storage on physical and virtual values.

[0034] Based on the core parameters collected, physical values ​​(such as the actual 35kV bus voltage) and virtual values ​​(such as the simulated 35kV bus voltage) of corresponding dimensions are selected, and the deviation rate between the two is calculated. The formula for calculating the deviation rate is: δ = (physical value - virtual value) / physical value × 100%. The deviation rate reflects the degree of consistency between the digital twin model and the physical entity. If the deviation rate exceeds the threshold (such as 5%), it indicates that the model needs to be corrected to improve the simulation accuracy, providing a quantitative basis for subsequent control strategy adjustments.

[0035] Through closed-loop correction of the deviation rate between physical and virtual values, the digital twin model can be calibrated in real time, ensuring that the simulation of power grid transient processes (such as fault transients and new energy fluctuation transients) is highly consistent with the physical reality, thus providing a reliability guarantee for the simulation results.

[0036] Dynamic optimization transient process control strategies based on deviation rate can include the following:

[0037] If the virtual value deviates significantly from the physical value, the virtual value output of the digital twin model is directly corrected to make it closer to the physical entity.

[0038] For data showing abnormal fluctuations in core parameters (such as sudden voltage drops), reverse calibration of the acquisition system or additional sensor deployment should be performed to ensure data reliability.

[0039] Modify scene parameters (such as the illumination / wind speed model under extreme weather conditions, and the short-circuit current parameters in fault scenarios) to improve the model's adaptability to complex scenarios;

[0040] Adjust the model control mode (such as switching from "normal control" to "transient stable control"), optimize the control logic of devices such as inverters and SVG, and simulate the dynamic response during transient processes.

[0041] Based on a dynamically adjusted control strategy, potential risks such as voltage instability and frequency limit violations during transient processes can be simulated in advance, providing predictive control schemes (such as rapid reactive power compensation and smooth power regulation) for physical power plants, reducing the probability of transient fault escalation. Large-scale transient tests within the physical power plant are unnecessary; various extreme transient scenarios (such as short-circuit faults and sudden load changes) can be simulated through a digital twin model, reducing equipment losses and safety risks while shortening the testing cycle. Since the model can be continuously updated according to the operating status of the physical power plant (such as equipment aging and capacity expansion), and through parameter iteration and control strategy optimization, this invention can adapt to transient process changes throughout the entire lifecycle of the power grid, improving the long-term reliability of the system.

[0042] In addition, by simulating the impact of environmental factors such as sunlight and wind speed on transient processes, the impact of renewable energy output fluctuations on the power grid can be accurately assessed, which can help formulate more reasonable grid connection strategies and thus improve the efficiency of renewable energy consumption.

[0043] Further, step S101 involves constructing a digital twin simulation model based on the new energy single-unit simulation model and the three-dimensional station virtual model, specifically including the following steps:

[0044] Step S1011: Construct a single-machine twin model based on the new energy single-machine simulation model.

[0045] Step S1013: Based on the single-machine twin model, correct the multiplication factor to obtain the real-time correction factor.

[0046] Step S1015: Feed back the real-time correction coefficient to the single-machine twin model to obtain the corrected single-machine twin model.

[0047] Step S1017: Construct a 3D virtual model of the site according to the physical topology at a 1:1 scale.

[0048] Step S1019: Construct a digital twin simulation model based on the 3D virtual station model and the corrected single-machine twin model.

[0049] In this embodiment, the core parameters include 13 sub-indicators, which can be divided into three categories: a. Key electrical quantities: 35kV bus voltage, grid-connected frequency, inverter output current, etc., which are directly related to the core indicators of grid voltage / frequency stability during transient processes (e.g., when photovoltaic output drops by more than 30% due to cloud cover, bus voltage fluctuations will appear in milliseconds, requiring accurate data collection to capture transient inflection points); b. Environmental impact quantities: solar intensity, wind speed, ambient temperature, designed for the "strong weather dependence" of new energy output (e.g., gusts can cause wind power voltage flicker, requiring real-time wind speed data collection to correct the power output characteristics of the virtual model); c. Simulation-related quantities: simulation voltage, simulation power, simulation reactive power output, achieving direct benchmarking between physical and virtual values ​​(conventional data collection only focuses on physical data, while this application collects "physical-virtual" parameters simultaneously, providing a data basis for deviation rate calculation).

[0050] By collecting three different types of core parameters, and considering the unique characteristics of the "millisecond-second transient process" in the new energy power grid, a three-in-one parameter system of "electrical quantity, environmental quantity, and simulation quantity" is formed through multi-dimensional technical screening. When these core parameters participate in the calculation of the deviation rate between the physical and virtual values ​​of physical stations, the calculation results will be more accurate. This is a prerequisite for effectively avoiding the shortcomings of traditional digital twin technology in new energy power grids, which is mentioned in the background technology, in terms of rapid model inaccuracy.

[0051] By analyzing the hardware topology (such as the IGBT switching logic of the inverter and the pitch control mechanism of the wind turbine), operating mechanism (such as the maximum power point tracking (MPPT) strategy and low voltage ride-through (LVRT) capability) and fault characteristics (such as overcurrent protection action curves and temperature threshold response) of individual equipment, physical parameters are transformed into mathematical equations and logic modules in the digital model. This constructs a twin model corresponding one-to-one with the individual equipment, thus obtaining a single-unit twin model. This single-unit twin model can accurately simulate the output characteristics of the individual unit under different operating conditions (such as sudden changes in light intensity and wind speed fluctuations), providing refined equipment-level support for subsequent site-level simulations.

[0052] A "dynamic coefficient module" has been added to the single-machine twin model to dynamically calibrate the matching degree between the single-machine model and the station-level operating status. The specific logic is as follows:

[0053] Calculate the initial multiplication factor based on the rated total capacity of the new energy power station (e.g., a 100MW photovoltaic power station) and the rated capacity of a single device (e.g., a single inverter with a capacity of 500kW). = Rated capacity of the power station / Capacity of a single unit (e.g., 100MW / 500kW=200). The physical meaning is: under ideal conditions, the output of the single-machine model needs to be amplified according to this coefficient to match the overall capacity scale of the station, providing a basic coefficient for the power coordination between the single-machine model and the three-dimensional station virtual model.

[0054] The dynamic coefficient module collects the actual total active power of the physical power station (physical power, such as the current actual output of 80MW) and the virtual power of the single-unit twin model (such as the simulated output of 420kW for a single unit), and corrects the coefficient K in real time according to the formula K= The value of K is corrected by multiplying (physical power / virtual power) (e.g., K = 200 × (80MW / 420kW) ≈ 200 × 190.48 ≈ 38096). The corrected K value can be fed back to the single-unit twin model in real time to adjust the output gain of the model, so that the virtual value of the single unit keeps dynamically matched with the actual operating status of the plant (such as total power and load distribution).

[0055] The real-time correction coefficient K is applied in real time to the calculation of core parameters such as power output and current response of the single-machine twin model: when the physical power is higher than the virtual power, the value of K increases, which increases the output weight of the single-machine model and avoids underestimation of the virtual value; when the physical power is lower than the virtual power, the value of K decreases, which suppresses the output deviation of the single-machine model and ensures that the total power of the three-dimensional site virtual model is consistent with that of the physical site.

[0056] By real-time correction of the correction coefficient K, the issue of the scale difference between the rated capacity of the single-unit model and the total capacity of the power station is resolved. This avoids distortion of the power station-level virtual values ​​(such as errors in total power calculation) caused by the accumulation of single-unit simulation deviations, making the overall output of the digital twin model closer to physical reality. When power grid transients occur, the dynamic coefficient can quickly respond to sudden changes in physical power and adjust the output characteristics of the single-unit model in real time. This ensures that the three-dimensional power station virtual model more accurately simulates dynamic behaviors such as power fluctuations and voltage support during transient processes, providing a reliable basis for adjusting transient control strategies.

[0057] In particular, when the actual capacity of the power station changes due to equipment commissioning / decommissioning, maintenance, or environmental changes (such as shading of some photovoltaic modules), the dynamic coefficient K can be automatically corrected through real-time comparison of physical power and virtual power, eliminating the need for manual adjustment of model parameters and improving the adaptive capability of the digital twin model. Real-time correction of coefficient K enables power matching between individual units and the power station, avoiding complex coupling logic in multi-unit model collaborative calculations, reducing the model's computational load, and improving the real-time performance of transient process simulations.

[0058] Based on the actual physical layout of new energy power plants, and following the principles of "consistent spatial topology and equivalent equipment association," a 1:1 replica of the plant's system-level structure is constructed to create a 3D virtual model of the plant, which includes the following:

[0059] Reconstruct the core facilities of the power station, such as the power collection line network (e.g., cable type, length, impedance parameters), the substation topology (e.g., 35kV / 110kV transformer connection method, busbar structure), and reactive power compensation devices (e.g., SVG, installation location and capacity of capacitor banks).

[0060] Map the control links of the physical site (such as the communication protocols and data interaction logic between the central control system and each device) to ensure that the collaborative relationships of the devices in the virtual model are consistent with those of the physical site.

[0061] Embedding environmental boundary conditions of the site (such as geographic coordinates and the impact of terrain occlusion on lighting / wind speed) improves the model's adaptability to real-world scenarios.

[0062] A digital twin simulation model is obtained by connecting the 3D site virtual model with the modified stand-alone twin model through virtual lines. That is, the stand-alone twin model and the 3D site virtual model are connected at the electrical and logical levels through preset virtual lines (digital links constructed based on the cable and overhead line parameters of the physical site, including electrical characteristics such as impedance, capacitance, and transmission delay).

[0063] The output of the single-machine twin model (such as the AC side of the photovoltaic inverter) is connected to the collection line of the virtual power station through a virtual line, and then connected to the power grid model through a virtual booster station to form a complete power flow path and complete the electrical connection.

[0064] Establish a real-time communication interface between models to enable the real-time uploading of the operating status data (such as output and temperature) of the single-machine twin model to the three-dimensional station virtual model. At the same time, the control commands (such as power scheduling signals) of the virtual station can be sent to the single-machine twin model, realizing dynamic collaboration between the equipment level and the system level and enabling data interaction.

[0065] Through the above connections, a digital twin simulation model covering the entire level of "single unit-station-grid interface" is finally formed.

[0066] Step S107, adjusting the transient process control strategy based on the adjustment object and deviation rate, specifically includes the following steps:

[0067] Step S1071: Divide the adjustment range based on the deviation rate. The adjustment range includes the normal range, the warning range and the abnormal range.

[0068] Step S1073: Adjust the corresponding triggering conditions based on the adjustment range;

[0069] Step S1075: Based on the constraints and triggering conditions, perform the corresponding operation steps on the adjustment object. The operation steps include correcting virtual values, correcting core parameters, correcting scene parameters, and switching model control modes.

[0070] The relationship between the deviation rate and the adjustment range satisfies:

[0071] δ≤3% is within the normal range;

[0072] The warning range is defined as 3% < δ ≤ 5%.

[0073] δ > 5% is considered an abnormal range;

[0074] Where δ represents the deviation rate.

[0075] When the deviation rate is within the normal range, if the deviation rate is ≤ the preset threshold A (e.g., 3%), it indicates that the virtual value of the digital twin model is highly consistent with the physical value, the transient process simulation results are reliable, and there is no need to actively adjust the strategy.

[0076] When the deviation rate is within the warning range, if threshold A < deviation rate ≤ threshold B (e.g., 3% < deviation rate ≤ 5%), it indicates a slight deviation between the model and the physical entity (which may be caused by environmental fluctuations, slight equipment aging, etc.), requiring continued attention and preparation for adjustment.

[0077] When the deviation rate is in the abnormal range, if the deviation rate is greater than the threshold B (e.g., greater than 5%), it indicates that the model deviates significantly from the physical entity (which may be caused by sensor failure, model parameter mismatch, transient failure, etc.), and the adjustment mechanism needs to be triggered immediately.

[0078] Among them, thresholds A and B can be dynamically adapted according to the sensitivity of the transient process (e.g., the threshold for voltage transients is stricter than that for frequency transients).

[0079] The monitoring period can be, for example, 10ms. Differentiated trigger conditions can be set for different adjustment ranges, and continuous periodic monitoring can avoid misjudgment of single deviations and improve the reliability of the adjustment strategy.

[0080] Triggering conditions include: being in the warning range for N consecutive monitoring periods, being in the abnormal range for M consecutive monitoring periods, or being in the abnormal range for P consecutive monitoring periods after an abnormal adjustment, where N, M, and P may be the same or different. For example, N can be 2, M can be 1, and P can be 3.

[0081] If the physical value and the virtual value are in the warning range for N consecutive monitoring cycles (e.g., N=2 cycles, each cycle 10ms), it indicates that the deviation is accumulating, and then the "mild adjustment level" is triggered.

[0082] If the data is in the abnormal range for M consecutive monitoring periods (e.g., M=1 period), it indicates that the deviation has affected the accuracy of the transient inference and triggers the "moderate adjustment level".

[0083] If the abnormal interval remains in the abnormal interval for P consecutive monitoring periods (e.g., P=3 periods) after the initial adjustment, it indicates that the initial adjustment did not solve the problem and triggers the "deep adjustment level" (e.g., model parameter reconstruction).

[0084] The values ​​of N, M, and P are set according to the response speed of the transient process (smaller values ​​for fault transients and larger values ​​for steady-state fluctuations), and can be iteratively optimized through historical data.

[0085] The following are the targeted operations performed on the adjusted object based on the triggering conditions:

[0086] Correcting virtual values ​​(minor adjustment level): When the warning condition is triggered, the virtual values ​​of the digital twin model are directly calibrated based on the physical values ​​and the deviation rate (such as correcting the simulated voltage to "physical voltage - deviation amount") without changing the core parameters of the model, which can quickly reduce the deviation.

[0087] Correcting core parameters (moderate adjustment level): When an emergency condition is triggered, reverse-calibrate the sensor data of the physical site (such as eliminating abnormal fluctuations in inverter current) or update the equipment parameters in the model (such as correcting the resistance temperature coefficient of the transformer) to reduce deviation from the data source.

[0088] Correcting scenario parameters (depth adjustment level): If a depth adjustment is triggered, parameters are optimized for environmental or fault scenarios (such as updating the wind turbine power curve under extreme wind speeds and correcting the transition resistance model for short-circuit faults) to improve the model's adaptability to complex scenarios.

[0089] Switching model control mode: When the transient process involves a fault (such as a voltage drop), the model is automatically switched from "normal operation mode" to "transient stable mode", activating the inverter's low voltage ride-through control logic, SVG's fast reactive power support algorithm, etc., so that the virtual simulation is more in line with the actual response of the physical equipment.

[0090] By dividing the data into intervals and using multi-dimensional triggering conditions, a "one-size-fits-all" approach to adjustment is avoided. Minor deviations only require simple corrections, while serious deviations are processed in depth, reducing the interference of invalid operations on transient inference.

[0091] By combining the operation steps with the power grid constraints, it can be ensured that the adjusted model can not only fit the physical entity, but also does not violate the power grid's safe operation rules (such as avoiding the corrected voltage from exceeding the allowable range), thus improving the reliability of transient control.

[0092] Through continuous periodic monitoring and multi-level triggering mechanisms, the cumulative trend and persistence of deviations can be quickly identified, and adjustments can be initiated in the early stages of transient faults (such as short circuits), avoiding the expansion of deviations that could lead to inaccurate projections and buying time for fault handling.

[0093] For scenarios that remain abnormal after adjustment, in-depth adjustments are triggered, forcing iterative optimization of model and scenario parameters. Long-term use can enhance the adaptability of digital twin models to complex working conditions (such as equipment aging and sudden environmental changes) and extend the effective life cycle of the model.

[0094] Switching control modes (such as transient stability mode) allows virtual simulations to be directly linked to the control logic of physical devices, enabling a closed-loop "simulation-control" system and providing a practical strategy for predicting and handling transient processes in the power grid.

[0095] Specifically, taking the 35kV bus voltage as an example, the corresponding operation steps are performed on the adjustment object based on the constraint conditions and trigger conditions, as shown in Table 1.

[0096] Table 1

[0097]

[0098] Adjustment effect verification: The deviation rate is recalculated within one monitoring cycle after adjustment. If it does not meet the standard (mild adjustment δ>3%, moderate / deep adjustment δ>4%), the adjustment is repeated (maximum 2 times); if it still does not meet the standard, the adjustment level is upgraded (e.g., mild → moderate).

[0099] The 3D virtual model of the site is a three-dimensional virtual model constructed with differentiated precision based on the core area, related areas, and peripheral areas. This realizes the idea of ​​constructing a virtual model hierarchically according to the actual data of the physical site, according to the "core area, related area, and peripheral area." First, data such as GIS maps and equipment CAD drawings are collected and standardized. Then, differentiated precision modeling is used for equipment in different areas—a high-precision model with 1000-2000 polygons is used for the core area (such as grid connection points and SVG devices), a medium-precision model with 300-1000 polygons is used for related areas (such as combiner boxes), and a simplified model with <300 polygons is used for peripheral areas (such as photovoltaic panels). Finally, the model is assembled through topological association and the positional accuracy is verified (core area ≤ ±0.5m), achieving a 1:1 accurate mapping from the physical site to the virtual space.

[0100] The precision of the core area, related areas, and peripheral areas decreases sequentially. The core area maintains high precision to meet critical needs (such as accurate mapping of grid connection points), while the reduced precision in the peripheral areas simplifies modeling, reduces data volume, and lowers computational and time costs while ensuring core functionality, thus improving overall construction efficiency. The simplified model in the peripheral areas meets the needs of overall planning and macro-level visualization, enhancing the model's usability. This differentiated structure makes model maintenance more flexible; updates to the core area can be handled specifically, while the simplified design of the peripheral areas makes updates more efficient, reducing long-term maintenance costs.

[0101] Furthermore, such as Figure 2 As shown, the power grid transient process extrapolation method based on digital twins also includes the following steps:

[0102] Step S201: Construct a dynamic reproduction and prediction model for transient processes based on time and space dimensions.

[0103] Step S203: Map the dynamic data to the transient process dynamic reproduction and prediction model to generate a dynamic simulation screen. The dynamic data includes time window control, electrical quantity rendering, trajectory generation and transient process control strategy.

[0104] In this embodiment, a dynamic reproduction and prediction model for transient processes is constructed through coordinated control of time and space dimensions, enabling dynamic reproduction and prediction of transient processes. The time dimension provides three time windows: detail window (0.1s, step size 0.01s), regular window (10s, step size 0.1s), and trend window (10min, step size 10s), supporting playback, pause, and other controls. The spatial dimension divides the station into a 1m×1m grid, calculates the electrical quantity distribution through interpolation, and renders it with a three-color gradient. Simultaneously, the transient starting point is located based on the electrical topology, and a dynamic arrow trajectory is generated according to the propagation path (length corresponds to the change, and flashing frequency reflects the rate of change), thus intuitively presenting the entire process of transient events from occurrence to evolution.

[0105] Furthermore, continue to combine Figure 2 As shown, the power grid transient process simulation method based on digital twins also includes the following steps: based on the interactive control unit and the three-dimensional station virtual model, the dynamic reproduction and prediction model of the control transient process generates a simulation effect and comparison report.

[0106] The interactive control unit provides users with a convenient virtual-real interaction interface, supporting parameter querying, virtual intervention, and view control. Parameter querying allows users to retrieve parameter values, historical curves, and regional statistical results in real time by clicking on the device or selecting a region. Virtual intervention allows users to input operation commands (such as adjusting SVG reactive power), and the system verifies the input, previews the effect in the next 10 seconds, and generates a comparison report. View control supports preset viewpoint switching, zooming, rotation, and transient tracking, ensuring that users can focus on key areas. This interactive control unit realizes real-time linkage between the virtual scene and user operations, improving the interactivity and decision-making efficiency of transient analysis.

[0107] This refers to the system's ability to quickly simulate and preview the real-time effects of a user's virtual intervention command (such as adjusting the reactive power output of an SVG) within the next 10 seconds, based on current grid conditions and equipment parameters. These real-time effects include, but are not limited to, trends in key grid parameters (such as voltage, current, and power), dynamic responses to equipment operation, and any abnormal fluctuations, allowing users to intuitively see the immediate impact of the operation.

[0108] After a 10-second preview, the system automatically generates a report. The core of this report is a comparison between the "state 10 seconds after the operation" and the "current state or baseline state before the operation was performed." The comparison includes, but is not limited to, differences in key parameters, changes in equipment operating efficiency, and differences in stability indicators. This helps users clearly assess the rationality, effectiveness, and potential risks of the operation commands, providing data support for decision-making. The "10-second effect" is a real-time simulation of the operation, while the "comparison report" is a quantitative analysis based on the simulation. The combination of these two allows users to evaluate the effects and optimize decisions before actually performing the operation, improving the accuracy and efficiency of interactive control.

[0109] The system transforms modeling results into integrated spatiotemporal visualization content, comprising a 3D modeling unit, a spatiotemporal simulation unit, and an interactive control unit. The high-precision virtual scene constructed by the 3D modeling unit provides the foundation for spatiotemporal simulation; its equipment model and topological relationships directly determine the spatial display accuracy of transient trajectories. Based on the modeling results, the spatiotemporal simulation unit maps dynamic data such as time window control, electrical quantity rendering, and trajectory generation to the 3D model in real time, transforming the static scene into a dynamic simulation view. The interactive control unit acts as a bridge between the user and the system. Upon receiving user queries or intervention commands, it calls the equipment parameter association function of the 3D modeling unit to obtain basic data, and simultaneously drives the spatiotemporal simulation unit to update the simulation view or generate intervention pre-simulation results. Ultimately, through view control, it achieves focused display of the dynamic scene, forming a closed-loop collaborative mechanism where "modeling supports simulation, simulation depends on interaction, and interaction reacts to modeling and simulation."

[0110] In summary, this invention achieves precise fusion of physical and virtual data by constructing a virtual-real data fusion module, performing time-scale calibration based on a specific benchmark, and storing data in a quadruplet format partitioned by transient stage. It also achieves dynamic model adaptation to different states by constructing an adaptive twin modeling module, based on a hierarchical adjustment mechanism of single-machine dynamic coefficient correction and core parameter deviation monitoring. Furthermore, it provides an intuitive presentation of transient processes through a panoramic simulation visualization module, employing hierarchical 3D modeling, dynamic mapping of spatiotemporal dimensions, and an interactive control closed-loop mechanism. Finally, by combining closed-loop collaboration and dynamic compensation strategies, real-time deviation adjustment and periodic parameter updates ensure the adaptability and accuracy of the simulation, forming a comprehensive panoramic simulation scheme for new energy transient processes covering virtual-real fusion, adaptive modeling, panoramic visualization, and dynamic optimization.

[0111] Taking the transient process of cloud shading at a photovoltaic power station as an example, the present invention presents the following method for extrapolating the transient process of the power grid based on digital twins.

[0112] I. Application Background

[0113] A 100MW photovoltaic power station experienced a sudden drop in output from 80MW to 30MW due to cloud cover in the afternoon, causing voltage fluctuations at the grid connection point (the lowest measured value was 0.92pu). The panoramic simulation technology of this invention is needed to reconstruct the transient process, locate the scope of impact, and optimize subsequent response strategies.

[0114] II. Implementation Steps

[0115] 1. Data Acquisition and Virtual-Real Fusion

[0116] Data input: Collect physical site data (voltage and power data of synchronous phasor measurement device, inverter status data of SCADA) and virtual simulation data (transient simulation results based on site topology).

[0117] Time scale calibration: Using the GPS time scale of the synchronous phasor measurement device as a reference, the time scale deviation of the virtual simulation data Δt=2ms is calculated. Through dynamic interpolation correction (the confidence weight ω is taken as the standard value of the corresponding device), the synchronization accuracy of virtual and real data is controlled within ±1ms.

[0118] Coupled storage: Stored in the format of "timestamp + physical value + virtual value + deviation rate" (e.g., "2024-08-10 14:30:00.000,0.92pu,0.91pu,1.1%), and archived by partitioning into "5s before occlusion / during occlusion / 5s after occlusion".

[0119] 2. Adaptive Twin Modeling

[0120] Basic model construction:

[0121] Standalone model: Add a "dynamic coefficient module" to the photovoltaic inverter, initial coefficients = Rated capacity of the power station / single unit capacity, with the real-time correction factor K dynamically adjusted based on the ratio of physical power to virtual power.

[0122] Station Model: A virtual station (including inverters, combiner boxes, 35kV busbars, and other equipment) is constructed at a 1:1 scale based on the physical topology and connected via virtual lines (resistance and reactance are consistent with the actual ones).

[0123] Deviation monitoring: Core parameters (grid connection point voltage, inverter output current, illuminance, etc.) are periodically collected, and the deviation rate δ is calculated. When shading occurs, the deviation between the virtual simulation voltage and the physically measured voltage is δ=1.1% (normal range), and no immediate adjustment is required.

[0124] Dynamic adjustment: 3 seconds after the occlusion, the deviation rate δ rises to 4.2% (warning range) because the cloud layer moves faster than the simulation preset value, triggering a slight adjustment - correcting the light intensity attenuation coefficient, so that the deviation rate drops back to 2.8%.

[0125] 3. Panoramic simulation visualization

[0126] 3D modeling: The virtual scene is constructed in a hierarchical manner according to the "core area (grid connection point, 35kV busbar), associated area (combiner box), and peripheral area (photovoltaic array)". The location accuracy of the core area is consistent with that of the physical site.

[0127] Spatiotemporal simulation: In the time dimension, select "detail window" (0.1s duration, 0.01s step) to clearly show the instantaneous changes in voltage fluctuations; in the spatial dimension, render the voltage distribution with a 1m×1m grid (red indicates low voltage area), and present the propagation path of voltage fluctuations from the photovoltaic array to the grid connection point through dynamic arrow trajectories (arrow length corresponds to fluctuation amplitude).

[0128] Interactive control: By clicking on the 35kV bus in the virtual scene, the voltage change curve can be retrieved in real time; by inputting the command "increase SVG reactive power output", the voltage recovery effect in the next 10 seconds can be previewed (the simulated display shows the voltage recovering to 0.96pu).

[0129] 4. Closed-loop optimization feedback

[0130] Real-time adjustment: Based on the simulation results, the on-site control SVG device increased the reactive power output, and the measured voltage rebounded to 0.95 pu within 10 seconds, consistent with the simulation effect.

[0131] Model update: Based on the data from this occlusion event, the correlation model parameters between cloud occlusion speed and power attenuation were updated, resulting in a 30% reduction in the simulation deviation rate for similar scenarios in the future.

[0132] The beneficial effects of the technical solutions provided by the embodiments of the invention include at least the following:

[0133] 1. Improved accuracy of virtual-real fusion: Through dynamic time-stamping calibration and data coupling mechanism, the synchronization deviation between physical and virtual data can be significantly reduced, which can greatly improve the fusion accuracy compared with traditional solutions, while ensuring that data integrity is at a high level.

[0134] 2. Optimization of inference accuracy: The hierarchical adjustment mechanism effectively reduces the prediction error of transient parameters, and the accuracy is significantly improved compared with the fixed model;

[0135] 3. Improved decision-making efficiency: 3D visualization technology can accelerate fault location, and virtual intervention can reduce the cost of physical trial and error, thereby improving overall decision-making efficiency.

[0136] 4. Enhanced scenario adaptability: It is compatible with multiple voltage levels and energy types, significantly shortens the adaptation cycle for new scenarios, and has strong scenario adaptability.

[0137] On the other hand, the present invention proposes an electronic device, characterized in that it includes: at least one processor; a memory communicatively connected to the at least one processor; wherein the memory stores instructions executable by the at least one processor, the instructions being executed by the at least one processor to enable the at least one processor to execute any one of the digital twin-based power grid transient process simulation methods.

[0138] On the other hand, the present invention proposes a medium, which is a computer storage medium storing a computer program, wherein the computer program, when executed by a processor, implements any one of the methods for simulating transient processes of a power grid based on digital twins.

[0139] Computer storage media may be simply referred to as media. Those skilled in the art will understand that all or part of the processes in the methods of the above embodiments can be implemented by a computer program instructing related hardware. The program can be stored in a non-volatile computer-readable storage medium, and when executed, it can include the processes of the embodiments of the above methods. Any references to memory, storage, databases, or other media used in the embodiments provided in this application may include non-volatile and / or volatile memory. Non-volatile memory may include read-only memory (ROM), programmable ROM (PROM), electrically programmable ROM (EPROM), electrically erasable programmable ROM (EEPROM), or flash memory. Volatile memory may include random access memory (RAM) or external cache memory. By way of illustration and not limitation, RAM is available in various forms, such as Static RAM (SRAM), Dynamic RAM (DRAM), Synchronous DRAM (SDRAM), Dual Data SDRAM (DDRSDRAM), Enhanced SDRAM (ESDRAM), Synchlink DRAM (SLDRAM), Rambus Direct RAM (RDRAM), Direct Memory Bus Dynamic RAM (DRDRAM), and Memory Bus Dynamic RAM (RDRAM). The various embodiments described in this specification are presented in a progressive manner, with reference allowed to each other for similar or identical parts. Each embodiment focuses on describing the differences from other embodiments. In particular, the embodiments for apparatuses, devices, and non-volatile computer storage media are described simply because they are substantially similar to the method embodiments; relevant details can be found in the descriptions of the method embodiments.

[0140] The above embodiments are merely illustrative examples and are not intended to limit the implementation. Those skilled in the art will recognize that other variations or modifications can be made based on the above description. It is neither necessary nor possible to exhaustively list all possible implementations. However, obvious variations or modifications derived therefrom are still within the scope of protection of this invention.

Claims

1. A method for extrapolating transient processes in a power grid based on digital twins, characterized in that, include: A digital twin simulation model is constructed based on a single-unit simulation model of new energy and a three-dimensional virtual model of the power station, specifically including: Constructing a single-machine twin model based on a new energy single-machine simulation model; Based on the single-machine twin model, the multiplication factor is corrected to obtain the real-time correction factor; The real-time correction coefficients are fed back to the single-machine twin model to obtain the corrected single-machine twin model; Construct a 3D virtual model of the station based on the physical topology at a 1:1 scale; Based on the three-dimensional virtual station model and the modified single-machine twin model, the digital twin simulation model is constructed by connecting the three-dimensional virtual station model and the modified single-machine twin model through virtual lines. The digital twin simulation model outputs virtual values; During the monitoring period, core parameters are collected, including 35kV bus voltage, grid-connected frequency, total active power, reactive power, inverter output current, transformer temperature, SVG reactive output, illuminance, wind speed, ambient temperature, simulated voltage, simulated power, and simulated reactive output. The deviation rate between the physical values ​​and the virtual values ​​of the physical station is calculated based on the core parameters. The transient process control strategy, based on the adjustment object and the deviation rate, specifically includes: Based on the deviation rate, adjustment intervals are divided, including normal intervals, warning intervals, and abnormal intervals; Adjust the corresponding triggering conditions based on the adjustment range; Based on the constraints and the triggering conditions, corresponding operation steps are performed on the adjustment object. The operation steps include correcting the virtual value, correcting the core parameters, correcting the scene parameters, and switching the model control mode. The adjustment objects include the virtual values, the core parameters, the scene parameters, and the model control mode.

2. The method for extrapolating power grid transient processes based on digital twins according to claim 1, characterized in that, The correspondence between the deviation rate and the adjustment range satisfies: δ≤3% is within the normal range; The warning range is defined as 3% < δ ≤ 5%. δ > 5% is considered an abnormal range; Where δ represents the deviation rate.

3. The method for extrapolating power grid transient processes based on digital twins according to claim 1, characterized in that, The physical values ​​of the physical station and the virtual values ​​are time-calibrated to obtain standard data, and the synchronization accuracy of the standard data is ≤ ±1ms.

4. The method for extrapolating power grid transient processes based on digital twins according to claim 1, characterized in that, The three-dimensional station virtual model is a three-dimensional virtual model constructed with differentiated precision based on the core area, related areas and peripheral areas.

5. The method for extrapolating power grid transient processes based on digital twins according to any one of claims 1 to 4, characterized in that, Also includes: A dynamic reproduction and prediction model for transient processes is constructed based on time and space dimensions; Dynamic data is mapped to the transient process dynamic reproduction and prediction model to generate a dynamic simulation screen. The dynamic data includes time window control, electrical quantity rendering, trajectory generation and transient process control strategy.

6. The method for extrapolating power grid transient processes based on digital twins according to claim 5, characterized in that, Also includes: Based on the interactive control unit and the three-dimensional virtual model of the station, the dynamic reproduction and prediction model of the control transient process generates a preview effect and a comparison report.

7. An electronic device, characterized in that, include: At least one processor; A memory communicatively connected to the at least one processor; wherein the memory stores instructions executable by the at least one processor, the instructions being executed by the at least one processor to enable the at least one processor to perform the digital twin-based power grid transient process simulation method as described in any one of claims 1 to 6.

8. A medium, characterized in that, The system contains a computer program that, when executed by a processor, implements the method for extrapolating power grid transient processes based on digital twins as described in any one of claims 1 to 6.

Images