A grid-connection compatibility joint simulation test system for a grid-connection type energy storage power station

Abstract

The application discloses a kind of network type energy storage power station grid-connected compatibility joint simulation test systems, belong to energy storage power station grid-connected test technical field.The system includes: data acquisition module, obtains target power grid real-time operating parameter and network type energy storage power station control parameter to be tested;Mirror modeling module, construct the high-fidelity grid-connected mirror model between target power grid and energy storage power station;Simulation deduction module, based on multi-time scale dynamic state estimation executes backbone simulation deduction, obtains multiple time-domain simulation results;Disturbance injection module, actively injects preset power grid disturbance sequence at each time scale switching point and collects dynamic response data;Index calculation module, coupling comparison between dynamic response data and time-domain simulation results, calculate multi-dimensional compatibility index sequence.The system realizes the accurate quantitative evaluation of network type energy storage power station grid-connected compatibility by multi-time scale joint simulation and switching point disturbance injection.

Description

Technical Field

[0001] This invention relates to the field of grid-connected testing technology for energy storage power stations, specifically to a joint simulation testing system for grid-connected compatibility of grid-connected energy storage power stations. Background Technology

[0002] Grid-based energy storage power stations provide active support to the power grid by simulating the external characteristics of synchronous generators. Their grid compatibility directly affects the grid's stability recovery capability under fault disturbances. Existing grid-based testing methods mainly rely on offline simulation or field short-circuit tests on a single time scale. Offline simulations typically analyze only one time scale of electromagnetic or electromechanical transients independently, failing to reflect the true dynamic behavior of grid-based energy storage in the cross-time-scale coupling processes of electromagnetic response, electromechanical inertia response, and steady-state power flow recovery. While field short-circuit tests can obtain response data under actual disturbances, the test conditions are limited, making it difficult to systematically cover all types of grid disturbances, from instantaneous voltage drops to long-term frequency fluctuations. Furthermore, the test process carries the risk of irreversible damage to equipment. Existing technical solutions have significant drawbacks: single-time-scale simulations artificially separate electromagnetic and electromechanical processes, ignoring the complex interaction between the microsecond-level switching behavior and millisecond-level inertia support of grid-based energy storage converters. This leads to assessments of voltage transient support capability and frequency inertial response capability deviating from the actual operating trajectory after grid connection. Meanwhile, the timing of disturbance testing injection is not aligned with the dynamic state switching points of the simulation model, making it impossible to apply excitation at the boundary moments when grid-connected energy storage systems are most prone to instability. The resulting dynamic response data lacks ergonomic coverage of operating conditions, making it difficult to support the construction of a refined multi-dimensional compatibility quantification index system. To address these issues, it is necessary to solve how to achieve continuous multi-timescale simulation of grid-connected energy storage power stations from electromagnetic transients to steady-state power flow within the same simulation framework, and how to accurately inject grid disturbances at the boundaries of different timescales to obtain the dynamic response trajectory under all operating conditions. Furthermore, it is also necessary to solve how to perform multi-dimensional waveform comparison between the disturbance response data and the benchmark simulation trajectory to achieve synchronous quantification of voltage transient support deviation and frequency inertial response deviation, thereby forming an accurate evaluation of grid connection point compatibility. Summary of the Invention

[0003] This invention provides a joint simulation test system for grid-connected compatibility of grid-connected energy storage power stations. The purpose is to solve the problems of continuous simulation of dynamic behavior across time scales and single dimension of compatibility quantitative evaluation in existing grid-connected tests by realizing continuous deduction of dynamic state estimation and time-domain disturbance injection under the joint simulation architecture, and performing multi-dimensional waveform coupling comparison of disturbance response.

[0004] To achieve the above objectives, this invention provides the following technical solution: This invention provides a joint simulation test system for grid-connected compatibility of a grid-connected energy storage power station, including a data acquisition module, a mirror modeling module, a simulation deduction module, a disturbance injection module, and an index calculation module. The data acquisition module acquires real-time operating parameters of the target power grid and the control parameters to be tested from the grid-connected energy storage power station; the real-time operating parameters include voltage phase angle, fundamental component of active power, and harmonic distortion rate synchronously acquired from multiple measurement points; the control parameters to be tested include virtual synchronous machine inertia time constant, primary frequency regulation droop coefficient, and reactive power voltage support curve slope. The data acquisition module performs outlier removal and linear interpolation completion on the voltage phase angle, fundamental component of active power, and harmonic distortion rate, and performs parameter normalization processing on the virtual synchronous machine inertia time constant, primary frequency regulation droop coefficient, and reactive power voltage support curve slope. The obtained standardized grid parameters and standardized energy storage parameters are associated and stored in the same data record row of the joint simulation database according to a globally unified timestamp, ensuring high consistency and real-time performance of multi-source data.

[0005] The mirror modeling module extracts the voltage phase angle and the fundamental component of active power from the co-simulation database based on the real-time operating parameters and the control parameters to be tested, and establishes the electrical dynamic equations of the target power grid under multiple typical operating scenarios. It substitutes the virtual synchronous machine inertia time constant, the primary frequency regulation droop coefficient, and the slope of the reactive power voltage support curve into the electromechanical transient model of the grid-connected energy storage power station to obtain the power station's dynamic response equation. A high-fidelity grid-connected mirror model is generated by iteratively solving the coupled boundary conditions of the electrical dynamic equations and the power station's dynamic response equations. Preferably, during the iterative solution process, the Newton-Raphson method combined with an adaptive relaxation factor is used to accelerate the convergence of the boundary conditions, and the relative error between the grid connection point voltage and power is checked after each iteration to ensure it is below a preset threshold, thereby significantly improving the construction accuracy and convergence speed of the grid-connected mirror model.

[0006] The simulation module, based on the high-fidelity grid-connected mirror model, performs backbone simulation based on multi-timescale dynamic state estimation. This module decomposes the high-fidelity grid-connected mirror model into an electromagnetic transient submodule, an electromechanical transient submodule, and a steady-state power flow submodule. The electromagnetic transient submodule is driven to perform the first round of simulation with microsecond-level steps, generating the electromagnetic transient simulation trajectory. Using the endpoint of the electromagnetic transient simulation trajectory as the initial state of the electromechanical transient submodule, a second round of simulation is performed with millisecond-level steps, generating the electromechanical transient simulation trajectory. The endpoint of the electromechanical transient simulation trajectory is used as the initial state of the steady-state power flow submodule, and a third round of simulation is performed with second-level steps, generating the steady-state power flow simulation trajectory. The electromagnetic transient submodule, electromechanical transient submodule, and steady-state power flow submodule communicate with each other via a shared memory bus through zero-copy transfer of boundary state variables. Synchronization verification of state variables is automatically performed at each timescale switching point, eliminating numerical jumps at cross-scale interfaces and ensuring the consistency and continuity of the simulation trajectory. When splicing the electromagnetic transient simulation trajectory, electromechanical transient simulation trajectory, and steady-state power flow simulation trajectory on the time axis, the trajectory segments corresponding to the last 100 simulation steps of the electromagnetic transient simulation trajectory are selected as the first transition trajectory segment, and the trajectory segments corresponding to the first 100 simulation steps of the electromechanical transient simulation trajectory are selected as the second transition trajectory segment. The first and second transition trajectory segments are smoothly connected using a weighted average method to eliminate trajectory abrupt change points. Then, they are time-series concatenated with the steady-state power flow simulation trajectory to obtain multiple sets of smooth and continuous time-domain simulation results, providing a high-precision simulation benchmark for subsequent compatibility evaluation.

[0007] At each timescale switching point in the main simulation, the disturbance injection module actively injects a preset power grid disturbance sequence and collects dynamic response data of the high-fidelity grid-connected mirror model under the power grid disturbance sequence. The data acquisition and control system monitors the current simulation progress in real time. When it is determined that a timescale switching point has been reached, it automatically triggers a disturbance generator that matches the fault type at the switching point. The disturbance generator generates a power grid disturbance sequence in the order of injecting small-signal disturbances first and then large-signal disturbances. At the first switching point between the electromagnetic transient submodule and the electromechanical transient submodule, a voltage sag fault sequence is injected into the grid connection point of the high-fidelity grid-connected mirror model. This voltage sag fault sequence includes single-phase grounding faults, two-phase short-circuit faults, and three-phase short-circuit faults, injected sequentially in ascending order of fault severity. At the second switching point between the electromechanical transient submodule and the steady-state power flow submodule, a frequency fluctuation disturbance sequence is injected into the active power control loop of the high-fidelity grid-connected mirror model. The voltage waveform change trajectory at the grid connection point and the output frequency response curve of the active power control loop are recorded simultaneously, and these are combined into dynamic response data. This hierarchical and orderly injection of disturbances at cross-scale switching points can more comprehensively stimulate the dynamic support characteristics of the grid-connected energy storage power station during different transient processes, enabling the collected dynamic response data to fully reflect the ultimate performance of the equipment.

[0008] The index calculation module couples and compares the dynamic response data with multiple sets of time-domain simulation results to calculate a multi-dimensional compatibility index sequence. Specifically, it compares the voltage waveform change trajectory in the dynamic response data with the electromagnetic transient simulation trajectory of the corresponding time period in multiple sets of time-domain simulation results to calculate the voltage transient support deviation value; it compares the output frequency response curve in the dynamic response data with the electromechanical transient simulation trajectory of the corresponding time period in multiple sets of time-domain simulation results to calculate the frequency inertial response deviation value; it inputs the voltage transient support deviation value and the frequency inertial response deviation value into a pre-trained fuzzy membership function to obtain the compatibility membership degree, and then generates a grid connection point compatibility score sequence through weighted summation, thereby achieving a precise quantitative assessment of the grid connection compatibility of grid-connected energy storage power stations.

[0009] Preferably, based on the multidimensional compatibility index sequence, the system automatically marks the risk parameter combinations in the control parameters to be tested of the grid-connected energy storage power station that exceed the preset compatibility boundary, and outputs the risk parameter combinations and their corresponding grid connection failure simulation scenario snapshots through the data acquisition and control system; the grid connection failure simulation scenario snapshots are written back to the high-fidelity grid connection mirror model, and the backbone simulation deduction based on multi-timescale dynamic state estimation is repeatedly executed until all multidimensional compatibility index sequences are within the preset compatibility boundary, thereby forming a parameter optimization closed loop, which automatically finds the control parameter combinations that meet the grid connection compatibility requirements while ensuring simulation accuracy, significantly improving test efficiency and effectiveness.

[0010] The technical effects and advantages provided by the present invention in the above technical solution are as follows:

[0011] By decomposing the high-fidelity grid-connected mirror model into electromagnetic transient sub-modules, electromechanical transient sub-modules, and steady-state power flow sub-modules, and driving the data acquisition and control system to perform three rounds of simulations in microsecond, millisecond, and second increments, electromagnetic transient simulation trajectories, electromechanical transient simulation trajectories, and steady-state power flow simulation trajectories are generated. These three trajectories are then stitched together along the time axis to form multiple sets of time-domain simulation results, achieving continuous evolution simulation of the grid-connected dynamic behavior of grid-connected energy storage across electromagnetic, electromechanical, and steady-state time scales. A voltage drop fault sequence is injected at the first switching point between the electromagnetic and electromechanical transient sub-modules, and a frequency fluctuation disturbance sequence is injected at the second switching point between the electromechanical and steady-state power flow sub-modules. This ensures that grid disturbances precisely act on the most sensitive boundary moments of the dynamic state transition of the grid-connected energy storage system. By collecting the voltage waveform change trajectory and the output frequency response curve of the active power control loop at the grid connection point, dynamic response data reflecting the entire process from transient voltage forced support to smooth transition of inertial frequency is obtained, avoiding the loss of dynamic behavior caused by the disconnect between the timing of disturbance injection and model state switching in traditional offline simulation. The voltage waveform change trajectory in the dynamic response data is compared with the electromagnetic transient simulation trajectory of the corresponding time period in multiple sets of time-domain simulation results to calculate the voltage transient support deviation value; the output frequency response curve is compared with the electromechanical transient simulation trajectory of the corresponding time period to calculate the frequency inertial response deviation value, thereby quantifying the degree of deviation between the grid-connected dynamic performance of the grid-connected energy storage power station and the benchmark simulation behavior in two independent dimensions: voltage transient and frequency inertial. Based on the voltage transient support deviation value and the frequency inertial response deviation value, a grid connection point compatibility scoring sequence is generated. This changes the traditional grid connection test method that relies on a single threshold for qualification judgment. The continuous scoring sequence characterizes the compatibility boundary of grid-type energy storage under various disturbance conditions, providing multi-dimensional quantitative basis for grid-type optimization of control parameters and grid adaptability verification. Attached Figure Description

[0012] To more clearly illustrate the technical solutions in the embodiments of this application or the prior art, the drawings used in the embodiments will be briefly introduced below. Obviously, the drawings described below are only some embodiments recorded in this invention. For those skilled in the art, other drawings can be obtained based on these drawings.

[0013] Figure 1 This is a flowchart of the joint simulation test system for grid-connected compatibility of grid-connected energy storage power stations; Figure 2 It is a flowchart of the standardization and associated storage of power grid parameters and energy storage parameters; Figure 3 This is a flowchart of the high-fidelity grid-connected mirror model generation process; Figure 4 It is a flowchart of multi-timescale simulation and smooth transition connection; Figure 5 This is a flowchart of the multi-timescale hybrid simulation disturbance injection control. Detailed Implementation

[0014] To make the objectives, technical solutions, and advantages of the embodiments of the present invention clearer, 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 some embodiments of the present invention, not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0015] See Figure 1 This invention provides a joint simulation test system for grid-connected compatibility of a grid-connected energy storage power station, comprising a data acquisition module, a mirror modeling module, a simulation deduction module, a disturbance injection module, and an index calculation module. The data acquisition module acquires real-time operating parameters of the target power grid and the control parameters to be tested from the grid-connected energy storage power station. The mirror modeling module constructs a high-fidelity grid-connected mirror model between the target power grid and the grid-connected energy storage power station based on the real-time operating parameters and the control parameters to be tested. The simulation deduction module performs a backbone simulation deduction based on multi-timescale dynamic state estimation based on the high-fidelity grid-connected mirror model, obtaining multiple sets of time-domain simulation results. The disturbance injection module actively injects a preset power grid disturbance sequence at each time-scale switching point of the backbone simulation deduction and collects the dynamic response data of the high-fidelity grid-connected mirror model under the power grid disturbance sequence. The index calculation module couples and compares the dynamic response data with the multiple sets of time-domain simulation results to calculate a multi-dimensional compatibility index sequence.

[0016] Example 1: In specific implementation, please refer to Figure 2The data acquisition module, through a data acquisition and control system deployed at multiple measurement points in the target power grid, synchronously acquires voltage phase angle, fundamental active power component, and harmonic distortion rate. The data acquisition and control system sends a synchronization trigger signal with a GPS time stamp to each measurement point. Upon receiving the trigger signal, the measurement points perform data acquisition at a uniform sampling frequency of 4800 Hz, ensuring 80 cycles are acquired per power frequency cycle, meeting the requirements for accurate capture of voltage phase angle, fundamental active power component, and harmonic distortion rate. The voltage phase angle is acquired by calculating the angular offset of the voltage vector at the measurement point relative to the GPS synchronization reference phasor in real time using a phase-locked loop algorithm. The fundamental active power component is extracted by performing a discrete Fourier transform on the instantaneous voltage and current to calculate the fundamental active power. The harmonic distortion rate is calculated using a fast Fourier transform, representing the percentage ratio of the root of the square of each harmonic content to the fundamental content. The data acquisition and control system simultaneously receives the control parameters to be tested from the grid-type energy storage power station. These parameters include the virtual synchronous machine inertia time constant, the primary frequency regulation droop coefficient, and the slope of the reactive power voltage support curve. The control system of the grid-type energy storage power station uploads the virtual synchronous machine inertia time constant, primary frequency regulation droop coefficient, and reactive power voltage support curve slope to the data acquisition and control system via a communication protocol. The communication protocol adopts the IEC61850 standard, and the data frame contains parameter identifiers, parameter values, and a timestamp indicating when the parameters take effect.

[0017] After synchronous acquisition is completed, outlier removal and linear interpolation are performed on the voltage phase angle, the fundamental component of active power, and the harmonic distortion rate. Outlier removal uses the Laida criterion; for the voltage phase angle sequence, the arithmetic mean of the voltage phase angle sequence is calculated. and standard deviation If the voltage phase angle collected at a certain moment satisfy Then determine the voltage phase angle. If the value is an outlier, the voltage phase angle data point is removed from the sequence. The same method is applied to the active power fundamental component sequence and the harmonic distortion rate sequence, respectively, to calculate the mean of the active power fundamental component sequence. with standard deviation and the mean of the harmonic distortion rate sequence with standard deviation Data points that do not meet the Raida criterion are removed as outliers. Linear interpolation completion is performed at the gaps created by outlier removal, assuming the voltage phase angle sequence is in the _th_ position. The data points at time 1 were removed, and the data points at time 2 were removed. Voltage phase angle at time With the Voltage phase angle at time If the data is valid, then pad the value. From the linear interpolation formula The calculated linear interpolation method for the fundamental component of active power and harmonic distortion rate is the same. The voltage phase angle sequence, the fundamental component sequence of active power, and the harmonic distortion rate sequence after outlier removal and linear interpolation constitute the standardized power grid parameters.

[0018] The virtual synchronous machine inertia time constant, the primary frequency regulation droop coefficient, and the slope of the reactive power voltage support curve are subjected to parameter normalization processing to obtain standardized energy storage parameters. The parameter normalization processing employs range transformation, mapping all parameters to the interval [0,1]. The transformation formula is as follows: .

[0019] in, This represents any one of the original parameter values ​​among the virtual synchronous machine inertia time constant, primary frequency regulation droop coefficient, or reactive power voltage support curve slope. This represents the minimum value within the corresponding parameter range. This represents the maximum value within the corresponding parameter range. This represents the standardized energy storage parameter values ​​after conversion. The range of the virtual synchronous machine inertia time constant is determined based on the physical characteristics of the inverters used in grid-type energy storage power stations. The value is 2 seconds. The value is set to 20 seconds. The range of the primary frequency regulation droop coefficient is determined based on the standard for the active power regulation zone of the power plant participating in primary frequency regulation. The value is 20. The value is set to 50. The range of the slope of the reactive voltage support curve is set according to the linear region slope of the power plant's reactive voltage control strategy. The value is 1. The value is set to 10. Each control parameter to be tested is normalized according to the above formula to obtain the normalized virtual synchronous machine inertia time constant, the normalized primary frequency regulation droop coefficient, and the normalized reactive power voltage support curve slope, which are then combined to form standardized energy storage parameters.

[0020] The data acquisition module associates and stores standardized grid parameters and standardized energy storage parameters in the same data record row of the co-simulation database using a globally unified timestamp. The globally unified timestamp, derived from GPS timing with microsecond accuracy, is used by the data acquisition and control system to add a timestamp to the sampling time or parameter effective time for each data point of voltage phase angle, active power fundamental component, harmonic distortion rate, and each record of control parameter under test when acquiring raw data. During storage, the co-simulation database establishes a time-series data table using the timestamp as the primary key. Each row in the time-series data table corresponds to a timestamp, and the fields within the row store the standardized voltage phase angle, standardized active power fundamental component, standardized harmonic distortion rate, normalized virtual synchronous machine inertia time constant, normalized primary frequency regulation droop coefficient, and normalized reactive power voltage support curve slope, respectively. If a timestamp contains only grid parameters or only energy storage parameters, the corresponding field in that row remains empty and is filled by the mirror modeling module using linear interpolation based on adjacent timestamp data during subsequent simulation data loading.

[0021] Example 2: In specific implementation, please refer to Figure 3 The mirror modeling module extracts the standardized voltage phase angle and the standardized fundamental active power component from the co-simulation database. The standardized grid parameters stored in the co-simulation database are aligned according to timestamps. The mirror modeling module reads the voltage phase angle sequence and the fundamental active power component sequence within a continuous time window through the database query interface. The length of the time window is determined according to the typical operating scenario switching cycle of the target grid and is set to 3600 seconds.

[0022] The mirror modeling module establishes dynamic equations for electrical quantities of the target power grid under several typical operating scenarios. These typical operating scenarios include peak load, valley load, and fault recovery scenarios. For the peak load scenario, the voltage phase angle variation trend during peak load periods is extracted from the voltage phase angle sequence, and the corresponding active power fundamental component variation trend is extracted from the active power fundamental component sequence. A swing equation is established between the node voltage phase angle and the injected active power fundamental component, in the form of a second-order differential equation, describing the synchronous generator rotor motion state and the electromagnetic power exchange relationship. For the valley load scenario, data from low load periods is extracted to establish dynamic equations for electrical quantities under light load conditions. The equation parameters reflect the slow drift characteristics of the voltage phase angle at low active power fundamental component levels. For the fault recovery scenario, voltage phase angle and active power fundamental component data are extracted during the grid recovery phase after fault clearance. Dynamic equations for electrical quantities, including a damped recovery term, are established. An exponential decay term is added to the equations to describe the convergence process of the voltage phase angle towards the steady-state value after the fault.

[0023] The mirror modeling module substitutes the normalized virtual synchronous machine inertia time constant, the normalized primary frequency regulation droop coefficient, and the normalized reactive power voltage support curve slope into the electromechanical transient model of the grid-type energy storage power station. The electromechanical transient model of the grid-type energy storage power station includes the virtual synchronous machine rotor motion equation, the primary frequency regulation control equation, and the reactive power voltage control equation. In the virtual synchronous machine rotor motion equation, the normalized virtual synchronous machine inertia time constant is mapped back from the normalized interval [0,1] to the original physical value interval [2 seconds, 20 seconds] and then input into the rotor inertia element. The rotor motion equation describes the dynamic relationship between the virtual rotor angular velocity and the active power deviation. In the primary frequency regulation control equation, the normalized primary frequency regulation droop coefficient is mapped back to the original interval [20,50] and then input into the droop characteristic curve. The primary frequency regulation control equation outputs the active power adjustment amount. In the reactive power voltage control equation, the slope of the normalized reactive power voltage support curve is mapped back to the original interval [1,10] and then substituted into the linear relationship between reactive power and voltage deviation. The reactive power voltage control equation outputs a reactive power reference value. The above three equations are combined to form the dynamic response equation of the power station. The dynamic response equation of the power station takes the grid connection point voltage and grid connection point frequency as inputs and the active power and reactive power output by the grid-connected energy storage power station as outputs.

[0024] The mirror modeling module generates a high-fidelity grid-connected mirror model by iteratively solving the coupled boundary conditions of the electrical quantity dynamic equation and the power plant dynamic response equation. The coupled boundary conditions are that the output voltage of the electrical quantity dynamic equation at the grid connection point is equal to the input voltage of the power plant dynamic response equation, and the output power of the power plant dynamic response equation is equal to the injected power of the electrical quantity dynamic equation. During the iterative solution process, the Newton-Raphson method combined with an adaptive relaxation factor is used to accelerate the convergence of the boundary conditions. In each iteration step, a Jacobian matrix is ​​constructed for four state variables: grid connection point voltage amplitude, grid connection point voltage phase angle, injected active power, and injected reactive power. The Jacobian matrix has a dimension of 4×4, and each element is determined by the partial derivatives of the electrical quantity dynamic equation and the power plant dynamic response equation with respect to the four state variables at the current iteration point. The adjustment rule for the adaptive relaxation factor is: Let the... The relaxation factor for the next iteration is , The initial value is 0.5; if the first... If the L2 norm of the state variable residual decreases after the next iteration compared to the previous iteration, then... Set as Multiply by a coefficient of 1.2, with an upper limit not exceeding 1.0; if the residual L2 norm increases, then... Set as Multiply by a coefficient of 0.5, with a lower limit not lower than 0.1. After each iteration, verify whether the relative error of the grid connection point voltage and the relative error of the grid connection point active power are lower than preset thresholds. The calculation method for the relative error of the grid connection point voltage is: based on the grid connection point voltage amplitude correction amount... The absolute value divided by the current grid connection point voltage amplitude To obtain the ratio The calculation method for the relative error of active power at the grid connection point is as follows: based on the active power correction amount at the grid connection point. The absolute value divided by the current active power at the grid connection point To obtain the ratio The preset threshold is uniformly set to... When both of the above ratios are lower than When the iteration converges, the combined parameters of the electrical dynamic equation and the power plant dynamic response equation at the convergence point are output to form a high-fidelity grid-connected mirror model.

[0025] Example 3: In specific implementation, please refer to Figure 4 The simulation module decomposes the high-fidelity grid-connected mirror model into three sub-modules: electromagnetic transient, electromechanical transient, and steady-state power flow. The electromagnetic transient sub-module describes the voltage and current transient behavior of power electronic switching devices near the grid connection point on a microsecond-scale timescale, including an insulated-gate bipolar transistor (IGBT) switching model, a DC bus capacitor charging and discharging model, and an LCL filter dynamic model. The electromechanical transient sub-module describes the motion state of the virtual synchronous machine rotor in a grid-connected energy storage power station and the electromechanical oscillation process of the power grid, including the virtual synchronous machine rotor motion equation, the primary frequency regulation droop characteristic equation, and the reactive power voltage control equation; the timescale of the electromechanical transient sub-module is on the millisecond scale. The steady-state power flow sub-module describes the node voltage distribution and active and reactive power flow distribution of the target power grid near the equilibrium point, including the linear power flow equation described by the node admittance matrix and the transformer tap adjustment model; the timescale of the steady-state power flow sub-module is on the second scale.

[0026] The electromagnetic transient submodule, electromechanical transient submodule, and steady-state power flow submodule communicate via a shared memory bus for zero-copy transfer of boundary state variables. The shared memory bus allocates a contiguous address space in the simulation server's physical memory, divided into three regions: the electromagnetic transient submodule output state variable storage area, the electromechanical transient submodule output state variable storage area, and the steady-state power flow submodule output state variable storage area. After completing a full simulation cycle, the electromagnetic transient submodule writes the instantaneous three-phase voltage amplitude, instantaneous three-phase current amplitude, and instantaneous frequency value at the grid connection point into its output state variable storage area. The electromechanical transient submodule reads these state variables through direct memory mapping, without needing to copy data from kernel space to user space. After completing the simulation, the electromechanical transient submodule writes the fundamental voltage amplitude, fundamental active power, fundamental reactive power, and virtual rotor angular velocity at the grid connection point into its output state variable storage area. The steady-state power flow submodule reads these state variables through direct memory mapping. At each timescale switching point, the simulation module automatically performs synchronization verification of the state variables. The synchronization verification method is to compare the timestamp of the state variable written by the previous sub-module with the timestamp of the state variable read by the next sub-module. If the timestamp deviation is less than 50 nanoseconds, the synchronization is considered successful. If the timestamp deviation is greater than or equal to 50 nanoseconds, the next sub-module waits until the timestamps are aligned before reading.

[0027] The simulation module uses a data acquisition and control system to drive the electromagnetic transient submodule to perform the first round of simulation in microsecond-level steps. The microsecond-level step size is set to 2 microseconds. Within each 2-microsecond step, the insulated-gate bipolar transistor (IGBT) switch model uses the trapezoidal integral method to solve the voltage and current differential equations during the switch state transition. The LCL filter dynamic model uses the nodal voltage method to solve for the filter capacitor voltage and filter inductor current. The simulation duration for the first round is 100 milliseconds, and the simulation time window covers the transient process of the IGBT switch and the DC bus voltage recovery process, generating an electromagnetic transient simulation trajectory. This trajectory includes the instantaneous values ​​of the three-phase voltage and current at the grid connection point for each 2-microsecond step.

[0028] The initial state of the electromechanical transient submodule is taken as the endpoint of the electromagnetic transient simulation trajectory. The simulation derivation module performs a second round of derivation with a millisecond-level step size. The millisecond-level step size is set to 1 millisecond, and 500 electromagnetic transient step sizes are aggregated into one electromechanical transient step size. The electromechanical transient submodule extracts the fundamental voltage amplitude, fundamental active power, and initial values ​​of the virtual rotor angular velocity at the grid connection point from the endpoint of the electromagnetic transient simulation trajectory. These values ​​are substituted into the virtual synchronous machine rotor motion equation, and numerical integration is performed using the improved Euler method within each 1-millisecond step size to calculate the increments of the virtual rotor angular velocity, active power output, and reactive power output. The simulation duration of the second round of derivation is 10 seconds, generating the electromechanical transient simulation trajectory. The electromechanical transient simulation trajectory includes the grid connection point fundamental voltage amplitude sequence, grid connection point fundamental active power sequence, grid connection point fundamental reactive power sequence, and virtual rotor angular velocity sequence for each 1-millisecond step size.

[0029] The endpoint of the electromechanical transient simulation trajectory is used as the initial state of the steady-state power flow submodule. The simulation derivation module performs the third round of derivation with a second-level step size. The second-level step size is set to 1 second, and 1000 electromechanical transient step sizes are aggregated into one steady-state power flow step size. The steady-state power flow submodule extracts the fundamental voltage amplitude and fundamental active power of the grid connection point from the endpoint of the electromechanical transient simulation trajectory, substitutes them into the linear power flow equation described by the node admittance matrix, and uses the PQ decomposition method to solve for the voltage amplitude and voltage phase angle of all nodes in the target power grid. Within each 1-second step size, the balance relationship between node injected power and node voltage is iteratively calculated. The simulation duration of the third round of derivation is 3600 seconds, generating a steady-state power flow simulation trajectory. The steady-state power flow simulation trajectory contains the voltage amplitude sequence and the voltage phase angle sequence of all nodes in the target power grid at 1-second step sizes.

[0030] The simulation module extracts the trajectory segments corresponding to the last 100 simulation steps of the electromagnetic transient simulation trajectory as the first transition trajectory segment. The total step size of the electromagnetic transient simulation trajectory is 50,000 steps, and the last 100 simulation steps correspond to the instantaneous values ​​of the three-phase voltage and current at the grid connection point from steps 49,901 to 50,000. The module also extracts the trajectory segments corresponding to the first 100 simulation steps of the electromechanical transient simulation trajectory as the second transition trajectory segment. The total step size of the electromechanical transient simulation trajectory is 10,000 steps, and the first 100 simulation steps correspond to the fundamental voltage amplitude and fundamental active power sequences at the grid connection point from steps 1 to 100.

[0031] On the time axis, the simulation module uses a weighted average method to smoothly connect the first and second transition trajectory segments. The first transition trajectory segment is mapped to the electromechanical transient time scale. The mapping method involves taking the root mean square value of the instantaneous three-phase voltage at the grid connection point for every 500 electromagnetic transient steps, obtaining an electromagnetic transient equivalent voltage sequence corresponding one-to-one with the electromechanical transient step. Let the first transition trajectory segment be... The electromagnetic transient equivalent voltage value corresponding to each mapping point is , The value range is from 1 to 100, and the second transition trajectory segment is the first The amplitude of the electromechanical transient fundamental voltage corresponding to each point is , The value range is from 1 to 100, and the transition voltage sequence after smooth connection is... Calculated using the following weighted average formula: .

[0032] in, Represents the smooth connection of the transition segment. Voltage value at each point Represents the first transition trajectory segment Electromagnetic transient equivalent voltage values ​​at each mapping point Represents the second transition trajectory segment The amplitude of the electromechanical transient fundamental voltage at each point Represents the sampling point number of the transition section. The weighting coefficient increases from 1 to 100. Follow The linear increase results in a high proportion of electromagnetic transient equivalent voltage at the connection initiation and a high proportion of electromechanical transient fundamental voltage at the connection termination. Applying the same weighted average to the active power sequence eliminates abrupt changes in the trajectory, thus smoothing the transition voltage sequence after the connection. The electromagnetic transient simulation trajectory, the first 49,900 steps of the electromagnetic transient simulation trajectory, the smooth transition section after connection, and all 3,600 steps of the steady-state power flow simulation trajectory are sequentially connected end to end to obtain multiple sets of time-domain simulation results.

[0033] Example 4: In specific implementation, please refer to Figure 5The disturbance injection module monitors the current simulation progress of the main simulation in real time through the data acquisition and control system. The data acquisition and control system allocates a progress flag storage area in the shared memory bus of the simulation module. This storage area contains three integer variables, recording the number of simulation steps completed by the electromagnetic transient submodule, the electromechanical transient submodule, and the steady-state power flow submodule, respectively. The data acquisition and control system polls at a 100-microsecond interval, cyclically reading the number of simulation steps completed by the electromagnetic transient submodule and the electromechanical transient submodule, and comparing it with the preset step thresholds corresponding to the time scale switching points. The step threshold corresponding to the first switching point between the electromagnetic transient submodule and the electromechanical transient submodule is 50,000 steps, which is the preset total number of steps for the electromagnetic transient submodule; the step threshold corresponding to the second switching point between the electromechanical transient submodule and the steady-state power flow submodule is 10,000 steps, which is the preset total number of steps for the electromechanical transient submodule.

[0034] When the data acquisition and control system determines that the current simulation progress of the backbone simulation has reached a time scale switching point, it automatically triggers a disturbance generator that matches the fault type at the time scale switching point. The disturbance generator is a software module that runs on the graphics processing unit of the simulation server. Internally, the disturbance generator maintains two disturbance sequence buffers: a small-signal disturbance sequence buffer and a large-signal disturbance sequence buffer. When the triggering condition is met, the disturbance generator generates a preset grid disturbance sequence in the order of injecting small-signal disturbances first, followed by large-signal disturbances. The small-signal disturbance includes a 100-millisecond voltage harmonic injection with an amplitude of 5% of the rated voltage at the grid connection point. The harmonic frequencies are set to a superposition of the 3rd, 5th, and 7th harmonics, with superposition weights of 0.6, 0.3, and 0.1, respectively. The large-signal disturbance is a voltage drop event or a frequency shift event. The disturbance generator first retrieves voltage harmonic injection data from the small-signal disturbance sequence buffer, writes it into the corresponding injection point of the high-fidelity grid-connected mirror model, and after a delay of 100 milliseconds, retrieves large-signal disturbance data from the large-signal disturbance sequence buffer to perform injection.

[0035] At the first switching point between the electromagnetic transient submodule and the electromechanical transient submodule, the disturbance injection module injects a voltage sag fault sequence into the grid connection point of the high-fidelity grid-connected mirror model. The voltage sag fault sequence includes single-phase grounding faults, two-phase short-circuit faults, and three-phase short-circuit faults, injected sequentially in ascending order of fault severity. For a single-phase grounding fault, the voltage sag is set to 30% of the rated voltage amplitude at the grid connection point, with a duration of 150 milliseconds. The faulty phase is phase A, and the phase-to-ground impedance of phase A drops to 0.1 ohms during the fault. For a two-phase short-circuit fault, the voltage sag is set to 60% of the rated voltage amplitude at the grid connection point, with a duration of 120 milliseconds. The faulty phases are phases B and C, which are shorted together by a 0.05-ohm impedance. For a three-phase short-circuit fault, the voltage sag is set to 90% of the rated voltage amplitude at the grid connection point, with a duration of 100 milliseconds. Phases A, B, and C are simultaneously shorted to the neutral point by a 0.02-ohm impedance. The interval between two adjacent fault injections is 500 milliseconds. During this interval, the high-fidelity grid-connected mirror model is restored to the steady-state operating state before the fault. The specific operation of voltage drop fault sequence injection is as follows: the disturbance generator superimposes the fault voltage vector onto the instantaneous values ​​of the three-phase voltages at the grid connection point. During a single-phase ground fault, the instantaneous value of phase A voltage is multiplied by the drop factor. , The value is 0.7, calculated from the voltage drop amplitude of 30% and the rated voltage amplitude at the grid connection point. During a two-phase short-circuit fault, the instantaneous voltage values ​​of phase B and phase C are multiplied by the voltage drop factor, respectively. , The value is 0.4, calculated from the voltage drop amplitude of 60% and the rated voltage amplitude at the grid connection point. During a three-phase short-circuit fault, the instantaneous values ​​of the three-phase voltages are all multiplied by a voltage drop factor. , The value is 0.1, calculated from the 90% voltage drop and the rated voltage amplitude at the grid connection point. .

[0036] At the second switching point between the electromechanical transient submodule and the steady-state power flow submodule, the disturbance injection module injects a frequency fluctuation disturbance sequence into the active power control loop of the high-fidelity grid-connected mirror model. The frequency fluctuation disturbance sequence consists of three consecutive sinusoidal frequency fluctuations with frequencies of 0.5 Hz, 1.0 Hz, and 2.0 Hz, respectively, and an amplitude of 0.2 Hz for each fluctuation. Each fluctuation lasts for 5 seconds, with a 3-second interval between adjacent fluctuations. The specific operation of injecting the frequency fluctuation disturbance sequence into the active power control loop is as follows: the disturbance generator superimposes a frequency offset onto the virtual rotor angular velocity reference value of the high-fidelity grid-connected mirror model. According to the variation of the sine function with time: .

[0037] in, This represents the frequency offset of the injected active power control loop, in Hertz. This represents the amplitude of frequency fluctuation, which is fixed at 0.2 Hz. The frequencies representing frequency fluctuations are taken as 0.5 Hz, 1.0 Hz, and 2.0 Hz, respectively. It represents the relative time from the start of each frequency fluctuation, in seconds; Representing the initial phase, it is fixed at 0 radians, ensuring that each frequency fluctuation starts from a zero offset. Frequency offset. After being superimposed onto the virtual rotor angular velocity reference value, it is converted into an active power regulation quantity through a primary frequency modulation droop characteristic equation, and then converted into a reactive power regulation quantity through a reactive voltage control equation, driving the high-fidelity grid-connected mirror model to generate a frequency response.

[0038] The disturbance injection module synchronously records the voltage waveform change trajectory and the output frequency response curve of the active power control loop at the grid-connected point through the data acquisition and control system. During voltage dip fault sequence injection, the data acquisition and control system records the instantaneous three-phase voltage values ​​at the grid-connected point with a sampling period of 100 microseconds, corresponding to a sampling frequency of 10 kHz. The recording duration is the duration of each fault plus a steady-state reference segment of 200 milliseconds before and after the fault. The voltage waveform change trajectory is formed by connecting consecutive sampling points. During frequency fluctuation disturbance sequence injection, the data acquisition and control system records the frequency response value output by the active power control loop with a sampling period of 10 milliseconds. The recording duration is the duration of each frequency fluctuation plus a steady-state reference segment of 3 seconds before and after the fluctuation. The output frequency response curve is formed by the frequency offset. The data acquisition and control system combines the voltage waveform change trajectory and the output frequency response curve according to the timestamps to form dynamic response data, which is then written into the dynamic response data storage table of the co-simulation database.

[0039] Example 5: In practical implementation, the index calculation module performs waveform overlap comparison between the voltage waveform change trajectory in the dynamic response data and the electromagnetic transient simulation trajectory for the corresponding time period in multiple sets of time-domain simulation results. The voltage waveform change trajectory in the dynamic response data consists of the sequence of instantaneous three-phase voltage values ​​at the grid connection point recorded by the data acquisition and control system at a sampling frequency of 10 kHz during the voltage drop fault sequence injection. The electromagnetic transient simulation trajectory is the sequence of instantaneous three-phase voltage values ​​at the grid connection point generated by the simulation deduction module in the first round of deduction in the electromagnetic transient submodule, with a time step of 2 microseconds. The index calculation module extracts the timestamp range of the voltage waveform change trajectory from the dynamic response data storage table in the joint simulation database, and extracts the voltage instantaneous value sequence for the same time period from the electromagnetic transient simulation trajectories of multiple sets of time-domain simulation results based on the timestamp range. The voltage transient support deviation value is calculated as follows: the voltage amplitude is subtracted from the voltage waveform change trajectory and the electromagnetic transient simulation trajectory at each sampling point on the same time axis, the absolute value of the difference is taken, and then the average value is calculated for the absolute values ​​of the differences of all sampling points within the fault duration. Let the voltage waveform change trajectory at the sampling point... The instantaneous amplitude of phase A voltage at point A is The electromagnetic transient simulation trajectory at the corresponding sampling point The instantaneous amplitude of phase A voltage at point A is The total number of sampling points is Then the voltage transient support deviation value Given by the following formula: .

[0040] in, This represents the voltage transient support deviation value, in volts. Represents the first in the dynamic response data The voltage waveform change trajectory at each sampling point; the instantaneous amplitude of phase A voltage; The first one in the electromagnetic transient simulation trajectory The instantaneous amplitude of phase A voltage at each sampling point; The total number of sampling points representing the voltage waveform change trajectory and the electromagnetic transient simulation trajectory within the fault duration period, after aligning them with the time axis. It is obtained by multiplying the fault duration by a sampling frequency of 10 kHz. For example, for a single-phase ground fault lasting 150 milliseconds, ; Represents the sampling point number, increasing from 1 to... The voltage transient support deviation values ​​for phases B and C were calculated using the same method, and the arithmetic mean of the three-phase voltage transient support deviation values ​​was taken as the final voltage transient support deviation value.

[0041] The index calculation module compares the output frequency response curve in the dynamic response data with the corresponding electromechanical transient simulation trajectories in multiple sets of time-domain simulation results by performing a rate of frequency change comparison. The output frequency response curve consists of a sequence of active power control loop frequency response values ​​recorded by the data acquisition and control system at a sampling period of 10 milliseconds during the injection of frequency fluctuation disturbance sequences. The electromechanical transient simulation trajectory is a sequence of grid-connected point frequency values ​​generated by the simulation derivation module in the second round of derivation in the electromechanical transient submodule, with a time step of 1 millisecond. The index calculation module extracts the timestamp range of the output frequency response curve from the dynamic response data storage table in the co-simulation database, and extracts the frequency value sequence for the same time period from the electromechanical transient simulation trajectories in multiple sets of time-domain simulation results based on the timestamp range. The frequency inertial response deviation is calculated as follows: The frequency change rate sequence is obtained by numerically differentiating the output frequency response curve; the simulated frequency change rate sequence is obtained by numerically differentiating the electromechanical transient simulation trajectory; the absolute values ​​of the differences between the frequency change rate sequence and the simulated frequency change rate sequence are taken at each sampling point on the same time axis; and the average value is then calculated over the absolute values ​​of the differences at all sampling points within the frequency fluctuation duration. The frequency change rate is calculated using the central difference method. Assuming the output frequency response curve is at the sampling point… The frequency value at that location is The sampling interval is Milliseconds, then the rate of change of frequency The rate of change of frequency of the electromechanical transient simulation trajectory The same central difference method is used for calculation. Frequency inertial response deviation value. The calculation method is similar to that of the voltage transient support deviation value, and the unit is Hertz per second.

[0042] The index calculation module inputs the voltage transient support deviation value and the frequency inertial response deviation value into a pre-trained fuzzy membership function to obtain the compatibility membership degree. The fuzzy membership function adopts a triangular membership function form, containing three fuzzy subsets corresponding to "low compatibility," "medium compatibility," and "high compatibility," respectively. The input to the fuzzy membership function is either the voltage transient support deviation value or the frequency inertial response deviation value, and the output is three compatibility membership degree values, representing the degree to which the input deviation value belongs to "low compatibility," "medium compatibility," and "high compatibility," respectively. The pre-training process of the fuzzy membership function is as follows: samples of voltage transient support deviation values ​​and frequency inertial response deviation values ​​recorded in historical grid-connected simulations are collected. Grid operation experts label each sample with a compatibility level, which is divided into three levels: low, medium, and high. The sample space of voltage transient support deviation values ​​is arranged in ascending order. The 10th and 90th percentiles are selected as the left and right vertices of the "high compatibility" triangle, the 40th and 60th percentiles as the left and right vertices of the "medium compatibility" triangle, and the 5th and 95th percentiles as the left and right vertices of the "low compatibility" triangle. The peak value of the membership function of each triangle is set to 1.0, corresponding to the center value of the fuzzy subset. The fuzzy membership function of the frequency inertial response deviation values ​​is trained in the same way to determine the triangle vertex positions. After training, the fuzzy membership functions are fixed as a set of triangle parameters.

[0043] The index calculation module will calculate the voltage transient support deviation value. The fuzzy membership function corresponding to the transient support deviation of the input voltage yields three membership values. , , ; the frequency inertial response deviation value By inputting the fuzzy membership function corresponding to the frequency inertial response deviation value, three membership values ​​are obtained. , , The final compatibility score is calculated by weighted summation, with the weighting coefficients determined based on a fuzzy rule table. The fuzzy rule table contains nine rules, each corresponding to a combination of a fuzzy subset of voltage transient support deviation values ​​and a fuzzy subset of frequency inertial response deviation values. The rule output is the compatibility score level. For a "high compatibility" rule combination, the compatibility score output value is set to 1.0; for a "medium compatibility" rule combination, the compatibility score output value is set to 0.5; and for a "low compatibility" rule combination, the compatibility score output value is set to 0.0. Final Compatibility Score It is calculated by a weighted average of the activation degree and output value of each rule, using the following formula: .

[0044] in, This represents the final compatibility score, ranging from 0.0 to 1.0. Represents the fuzzy rule sequence number. From 1 to 9; Representing the The activation degree of the fuzzy rule is taken as the smaller value between the membership degree of the voltage transient support deviation value and the membership degree of the frequency inertial response deviation value. ; Representing the The output score value corresponding to each fuzzy rule. The value can be 1.0, 0.5, or 0.0. The compatibility score sequence of all rule combinations constitutes the network point compatibility score sequence.

[0045] The index calculation module automatically marks risk parameter combinations that exceed the preset compatibility boundary among the control parameters to be tested in a grid-connected energy storage power station based on a multi-dimensional compatibility index sequence. The multi-dimensional compatibility index sequence includes a compatibility score for each fault type and each frequency fluctuation disturbance, with the preset compatibility boundary set at 0.6. When the compatibility score corresponding to a certain fault type or frequency fluctuation disturbance is lower than 0.6, the index calculation module marks the normalized virtual synchronous machine inertia time constant, normalized primary frequency regulation droop coefficient, and normalized reactive power voltage support curve slope participating in the simulation as risk parameter combinations. The index calculation module outputs the risk parameter combinations and corresponding grid-connection failure simulation scenario snapshots through the data acquisition and control system. The grid-connection failure simulation scenario snapshots include the time point when the compatibility score is lower than 0.6, a screenshot of the voltage waveform change trajectory, a screenshot of the output frequency response curve, and a snapshot of the state variables of the high-fidelity grid-connected mirror model during the fault.

[0046] The index calculation module writes back the snapshot of the grid connection failure simulation scenario to the high-fidelity grid connection mirror model. The write-back method involves using the snapshot of the state variables in the high-fidelity grid connection mirror model from the grid connection failure simulation scenario as the new initial state, loading it into the mirror modeling module. The mirror modeling module then updates the coupling boundary conditions of the electrical quantity dynamic equations and the power plant dynamic response equations based on the new initial state. The simulation derivation module repeatedly executes the backbone simulation derivation based on multi-timescale dynamic state estimation. The disturbance injection module re-injects the preset grid disturbance sequence at timescale switching points, and the index calculation module recalculates the multi-dimensional compatibility index sequence. This process is repeated until all compatibility scores in the multi-dimensional compatibility index sequence are greater than or equal to 0.6, meaning they are all within the preset compatibility boundaries.

[0047] The above description is merely a specific embodiment of this application, but the scope of protection of this application is not limited thereto. Any changes or substitutions that can be easily conceived by those skilled in the art within the scope of the technology disclosed in this application should be included within the scope of protection of this application.

Claims

1. A joint simulation test system for grid-connected compatibility of a grid-connected energy storage power station, characterized in that, include: The data acquisition module acquires the real-time operating parameters of the target power grid and the control parameters to be tested for the grid-type energy storage power station. The mirror modeling module constructs a high-fidelity grid-connected mirror model between the target power grid and the grid-connected energy storage power station based on the real-time operating parameters and the control parameters to be tested. The simulation module, based on the high-fidelity grid-connected mirror model, performs backbone simulation based on multi-time-scale dynamic state estimation to obtain multiple sets of time-domain simulation results. The disturbance injection module actively injects a preset power grid disturbance sequence at each time scale switching point in the main simulation and collects the dynamic response data of the high-fidelity grid-connected mirror model under the power grid disturbance sequence. The index calculation module couples and compares the dynamic response data with the multiple sets of time-domain simulation results to calculate a multi-dimensional compatibility index sequence.

2. The grid-connection compatibility joint simulation test system for a grid-connected energy storage power station according to claim 1, characterized in that, The steps for obtaining the real-time operating parameters of the target power grid and the control parameters to be tested of the grid-type energy storage power station specifically include: Through the data acquisition and control system, voltage phase angle, active power fundamental component and harmonic distortion rate are synchronously acquired from multiple measurement points of the target power grid. The data acquisition and control system receives the control parameters to be tested from the grid-type energy storage power station. The control parameters to be tested include the virtual synchronous machine inertia time constant, the primary frequency regulation droop coefficient, and the slope of the reactive voltage support curve. The voltage phase angle, the fundamental component of active power, the harmonic distortion rate, the virtual synchronous machine inertia time constant, the primary frequency modulation droop coefficient, and the slope of the reactive voltage support curve are aligned according to the timestamp and then stored in the co-simulation database.

3. The grid-connection compatibility joint simulation test system for a grid-connected energy storage power station according to claim 2, characterized in that, The steps for constructing a high-fidelity grid-connected mirror model between the target power grid and the grid-type energy storage power station based on the real-time operating parameters and the control parameters to be tested specifically include: The voltage phase angle and the fundamental component of active power are extracted from the joint simulation database to establish the electrical dynamic equations of the target power grid under multiple typical operating scenarios. Substituting the virtual synchronous machine inertia time constant, the primary frequency regulation droop coefficient, and the slope of the reactive voltage support curve into the electromechanical transient model of the grid-type energy storage power station, the dynamic response equation of the power station is obtained. The high-fidelity grid-connected mirror model is generated by iteratively solving the coupled boundary conditions of the electrical quantity dynamic equation and the power station dynamic response equation.

4. The grid-connection compatibility joint simulation test system for a grid-connected energy storage power station according to claim 3, characterized in that, In the step of iteratively solving the coupled boundary conditions of the electrical quantity dynamic equation and the power plant dynamic response equation, the Newton-Raphson method combined with an adaptive relaxation factor is used to accelerate the convergence of the boundary conditions, and after each iteration, it is verified whether the relative error between the grid connection point voltage and power is lower than a preset threshold.

5. The grid-connection compatibility joint simulation test system for a grid-connected energy storage power station according to claim 3, characterized in that, Based on the high-fidelity grid-connected mirror model, the specific steps for performing backbone simulation deduction based on multi-timescale dynamic state estimation to obtain multiple sets of time-domain simulation results include: The high-fidelity grid-connected mirror model is decomposed into an electromagnetic transient sub-module, an electromechanical transient sub-module, and a steady-state power flow sub-module. The data acquisition and control system is used to drive the electromagnetic transient submodule to perform the first round of simulation in microsecond-level steps to generate an electromagnetic transient simulation trajectory. The endpoint of the electromagnetic transient simulation trajectory is used as the initial state of the electromechanical transient submodule. A second round of deduction is performed with a step size of milliseconds to generate the electromechanical transient simulation trajectory. The endpoint of the electromechanical transient simulation trajectory is used as the initial state of the steady-state power flow submodule. A third round of simulation is performed with a step size of seconds to generate the steady-state power flow simulation trajectory. The electromagnetic transient simulation trajectory, the electromechanical transient simulation trajectory, and the steady-state power flow simulation trajectory are spliced ​​together on the time axis to obtain the multiple sets of time-domain simulation results.

6. The grid-connection compatibility joint simulation test system for a grid-connected energy storage power station according to claim 5, characterized in that, The electromagnetic transient submodule, the electromechanical transient submodule, and the steady-state power flow submodule transmit boundary state quantities through a shared memory bus, and automatically perform state quantity synchronization verification at each time scale switching point.

7. The grid-connection compatibility joint simulation test system for a grid-connected energy storage power station according to claim 5, characterized in that, The steps of actively injecting a preset power grid disturbance sequence at each time scale switching point in the main simulation and collecting dynamic response data of the high-fidelity grid-connected mirror model under the power grid disturbance sequence specifically include: At the first switching point between the electromagnetic transient submodule and the electromechanical transient submodule, the voltage drop fault sequence is injected into the grid connection point of the high-fidelity grid-connected mirror model; At the second switching point between the electromechanical transient submodule and the steady-state power flow submodule, the frequency fluctuation disturbance sequence is injected into the active power control loop of the high-fidelity grid-connected mirror model; The data acquisition and control system synchronously records the voltage waveform change trajectory of the grid connection point and the output frequency response curve of the active power control loop, and combines the voltage waveform change trajectory and the output frequency response curve into the dynamic response data.

8. The grid-connection compatibility joint simulation test system for a grid-connected energy storage power station according to claim 7, characterized in that, In the step of injecting the voltage dip fault sequence into the grid connection point of the high-fidelity grid-connected mirror model, the voltage dip fault sequence includes single-phase grounding fault, two-phase short-circuit fault and three-phase short-circuit fault, and is injected in order of increasing severity of fault type.

9. The grid-connection compatibility joint simulation test system for a grid-connected energy storage power station according to claim 7, characterized in that, The steps of coupling and comparing the dynamic response data with the multiple sets of time-domain simulation results to calculate the multidimensional compatibility index sequence specifically include: The voltage waveform change trajectory in the dynamic response data is compared with the electromagnetic transient simulation trajectory of the corresponding time period in the multiple sets of time-domain simulation results to calculate the voltage transient support deviation value. The frequency change rate of the output frequency response curve in the dynamic response data is compared with the electromechanical transient simulation trajectory of the corresponding time period in the multiple sets of time-domain simulation results, and the frequency inertial response deviation value is calculated. A grid connection point compatibility score sequence is generated based on the voltage transient support deviation value and the frequency inertial response deviation value.

10. The grid-connection compatibility joint simulation test system for a grid-connected energy storage power station according to claim 9, characterized in that, The steps for generating a grid connection point compatibility score sequence based on the voltage transient support deviation value and the frequency inertial response deviation value specifically include: inputting the voltage transient support deviation value and the frequency inertial response deviation value into a pre-trained fuzzy membership function to obtain the compatibility membership degree, and then calculating the final compatibility score by weighted summation.

Images