A method, system, and medium for identifying VSG virtual inertia based on static power adjustment difference equivalent damping identification.

Abstract

This invention relates to the field of smart grid technology, and discloses a method, system, and medium for identifying the virtual inertia of a VSG based on the identification of equivalent damping of static power adjustment difference. The method includes: constructing a dynamic relationship model between the virtual inertia of the VSG and the equivalent damping coefficient, and the virtual mechanical power and virtual electromagnetic power; establishing a closed-loop transfer function of VSG command power to output power based on the deviation of VSG output active power; setting the VSG command frequency to deviate from the grid's rated angular frequency to obtain the VSG steady-state power deviation; identifying the VSG equivalent damping coefficient based on the steady-state power deviation; inputting the identified VSG equivalent damping coefficient into the closed-loop transfer function, and after discretization, using the improved recursive auxiliary variable algorithm AFFRIVM with an adaptive variable forgetting factor to iteratively identify the virtual inertial time constant of the VSG.

Description

Technical Field

[0001] This invention relates to the field of smart grid technology, and in particular to the inertial parameter identification technology of virtual synchronous machines. Specifically, it relates to a method, system and medium for identifying the virtual inertia of a VSG based on the equivalent damping identification of static power adjustment difference. Background Technology

[0002] Currently, the penetration rate of renewable energy sources such as wind power in the power system continues to increase. However, the output of renewable energy is characterized by strong intermittency, volatility, and randomness, posing a serious challenge to the stable operation of the power system. Because this type of energy is typically connected to the grid via power electronic converters, it cannot provide the natural inertia and damping support found in traditional synchronous generators. This leads to a significant decrease in the overall inertia level of the power system, making frequency stability and frequency support capabilities increasingly prominent issues. The system exhibits low-inertia or even zero-inertia operating characteristics, posing a potential threat to grid security.

[0003] To address the issue of insufficient inertia in new power systems and improve system frequency regulation capabilities, Virtual Synchronous Generation (VSG) technology has attracted widespread attention. VSG simulates the rotor motion characteristics of a synchronous generator through control algorithms, transforming a static power electronic device without rotating parts into a rotating motor. By simulating the primary frequency and voltage regulation of a synchronous generator, it enables the system to dampen rapid voltage and frequency fluctuations, automatically distribute power, and operate synchronously with the grid, providing virtual inertia support, suppressing frequency fluctuations, and enhancing grid stability. The VSG topology includes a DC power supply, a power electronic converter, and an output LC filter. By embedding synchronous generator equations into the converter control system, VSG enables power exchange between the DC power supply and the system according to the characteristics of a synchronous generator. This technology provides a new approach for the stable operation of high-proportion renewable energy grids. Precise adjustment of the virtual inertia of the VSG can not only improve the system's equivalent inertia level but also enhance the frequency recovery capability after disturbances.

[0004] However, in practical applications, VSG devices from different manufacturers are often considered "black boxes," as their internal control parameters are not publicly available, making it impossible to directly obtain virtual inertia. How to accurately identify the virtual inertia of a VSG using externally measurable data has become a key issue in evaluating the overall system inertia level and optimizing scheduling strategies. Currently, accurate identification of VSG virtual inertia remains a challenge and a current research hotspot and difficulty. Summary of the Invention

[0005] Given the problems that VSG virtual inertia parameters cannot be directly observed and existing identification methods cannot be directly transferred in the prior art, the purpose of this invention is to provide a VSG virtual inertia identification method based on static power adjustment difference equivalent damping identification. This method generates a static angular frequency adjustment difference by instructing the VSG and obtains the static power adjustment difference to achieve equivalent damping identification. Furthermore, it expands the application by introducing the Adaptive Forgetting Factor Recursive Instrumental Variable Method (AFFRIVM) algorithm with an adaptive variable forgetting factor λ(k). Through iteration, it achieves fast and accurate identification of VSG virtual inertia, providing a fast and accurate identification scheme for real-time online monitoring and operation control of VSG.

[0006] According to a first aspect of the present invention, a method for identifying VSG virtual inertia based on static power adjustment difference equivalent damping identification is proposed, comprising the following steps:

[0007] Step 1: Construct a dynamic relationship model between the virtual inertia and equivalent damping coefficient of the VSG and the virtual mechanical power and virtual electromagnetic power, and establish a closed-loop transfer function of the VSG command power to the VSG output power based on the deviation of the VSG output active power.

[0008] Step 2: Set the VSG command frequency to deviate from the grid's rated angular frequency and obtain the VSG steady-state power deviation;

[0009] Step 3: Identify the VSG equivalent damping coefficient based on the steady-state power deviation;

[0010] Step 4: Input the identified VSG equivalent damping coefficient into the closed-loop transfer function. After discretization, use an improved recursive auxiliary variable algorithm with an adaptive variable forgetting factor λ(k) to iteratively identify the virtual inertial time constant of the VSG.

[0011] According to a second aspect of the present invention, a computer system is provided, comprising:

[0012] One or more processors;

[0013] The memory stores operable instructions that, when executed by the one or more processors, cause the one or more processors to perform operations, including the process of performing the VSG virtual inertia identification method based on static power adjustment difference equivalent damping identification as described in the foregoing embodiments.

[0014] In a third aspect of the present invention, a computer-readable storage medium is provided for storing a computer program comprising instructions / instruction set executable by one or more processors, wherein the instructions / instruction set, when executed by the one or more processors, implements the process of the VSG virtual inertia identification method based on static power adjustment difference equivalent damping identification of the foregoing embodiments.

[0015] The VSG virtual inertia identification method based on equivalent damping identification of static power adjustment difference, as described in the above embodiments of the present invention, determines the static power adjustment difference of the VSG by setting its command angular frequency to deviate from the rated angular frequency of the power grid and collecting data at the equivalent port. This allows for the identification of the VSG's equivalent damping coefficient, fundamentally solving the problems of the "black box" nature of VSGs from different manufacturers and the unavailability of detailed models and parameters. This achieves high-precision estimation of the VSG's equivalent damping. Furthermore, the Adaptive Forgetting Factor Recursive Auxiliary Variable Method (AFFRIVM) is extended to the virtual inertia identification of VSGs. By replacing the unobservable speed equation with the active power-command power transfer function, the method unifies the synchronous generator inertia identification and VSG virtual inertia identification methods.

[0016] Compared with the prior art, the VSG virtual inertia identification method based on static power adjustment difference equivalent damping identification of the present invention has the following significant advantages:

[0017] 1) To address the problem that key parameters such as the equivalent damping coefficient are difficult to obtain directly due to the lack of transparency (black box characteristics) of the internal control logic and parameters of VSGs from different manufacturers, this invention proposes an equivalent damping identification strategy based on static power adjustment difference. By instructing the VSG command frequency to deviate from the grid rated frequency, the equivalent damping coefficient is identified by using the difference between the output active power under steady state and the commanded active power. The identification has high accuracy and reliability, does not rely on the internal model provided by the manufacturer, and combines the black box limitation of VSG;

[0018] 2) Based on the second-order transient model of the synchronous generator, the closed-loop transfer function is reconstructed, transforming the unobservable speed-power angle relationship into an observable output power-command power relationship. Using the directly observable command power and output power data, the virtual inertia H is inversely calculated using the AFFRIVM algorithm. VSG It eliminates the need to measure the unobservable virtual rotor angular frequency or power angle, enabling direct and non-invasive identification of virtual inertia.

[0019] 3) Addressing the issue of dynamic response to historical data interference in VSG identification, direct application leads to slow convergence and large errors. This invention introduces an adaptive variable forgetting factor, dynamically adjusting it based on identification errors. It dynamically updates the results based on tests that reduce the influence of old data when errors are large and stabilize the results when errors are small, achieving rapid convergence and low fluctuations. Simultaneously, the identification process does not require altering the VSG's internal control logic or interrupting grid-connected operation. It uses only small frequency offsets (e.g., ±0.1Hz) and sudden load increases (e.g., 20kW) as excitations (meeting the grid frequency fluctuation range requirements specified in GB / T 12325-2022, without affecting grid voltage and frequency stability), achieving non-intrusive identification and online monitoring.

[0020] It should be understood that all combinations of the foregoing concepts and the additional concepts described in more detail below may be considered part of the inventive subject matter of this disclosure, provided that such concepts do not contradict each other. Furthermore, all combinations of the claimed subject matter are considered part of the inventive subject matter of this disclosure.

[0021] The foregoing and other aspects, embodiments, and features of the teachings of the present invention will be more fully understood from the following description in conjunction with the accompanying drawings. Other additional aspects of the invention, such as features and / or beneficial effects of exemplary embodiments, will become apparent from the following description or may be learned through practice of specific embodiments according to the teachings of the present invention. Attached Figure Description

[0022] The accompanying drawings are not intended to be drawn to scale. In the drawings, each identical or nearly identical component shown in the various figures may be denoted by the same reference numeral. For clarity, not every component is labeled in each figure. Embodiments of various aspects of the invention will now be described by way of example and with reference to the accompanying drawings.

[0023] Figure 1 This is a flowchart illustrating the VSG virtual inertia identification method based on static power adjustment difference equivalent damping identification according to an embodiment of the present invention.

[0024] Figure 2 This is a structural diagram of a VSG grid-connected system according to an embodiment of the present invention.

[0025] Figure 3 According to the embodiment f of the present invention ref =49.9Hz active power response.

[0026] Figure 4 According to the embodiment f of the present invention ref =50.1Hz active power response.

[0027] Figure 5 According to the embodiment f of the present inventionref =49.9Hz equivalent damping coefficient identification results.

[0028] Figure 6 According to the embodiment f of the present invention ref =50.1Hz equivalent damping coefficient identification results.

[0029] Figure 7 This is a schematic diagram of the frequency response results according to an embodiment of the present invention.

[0030] Figure 8 This is a schematic diagram of the command power and output power response according to an embodiment of the present invention.

[0031] Figure 9 H according to an embodiment of the present invention VSG The recursive update curve of the estimated value.

[0032] Figure 10 This is a schematic diagram comparing the outputs of VSG and the auxiliary model according to an embodiment of the present invention. Detailed Implementation

[0033] To better understand the technical content of the present invention, specific embodiments are described below in conjunction with the accompanying drawings.

[0034] Various aspects of the invention are described in this disclosure with reference to the accompanying drawings, which illustrate numerous illustrative embodiments. The embodiments of this disclosure are not necessarily intended to encompass all aspects of the invention. It should be understood that the various concepts and embodiments described above, as well as those described in more detail below, can be implemented in any of many ways, because the concepts and embodiments disclosed herein are not limited to any particular implementation. Furthermore, some aspects of the invention disclosed may be used alone or in any suitable combination with other aspects of the invention disclosed.

[0035] Combination Figure 1 As shown, the VSG virtual inertia identification method based on static power adjustment difference equivalent damping identification according to an embodiment of the present invention includes the following steps:

[0036] Step 1: Construct a dynamic relationship model between the virtual inertia and equivalent damping coefficient of the VSG and the virtual mechanical power and virtual electromagnetic power, and establish a closed-loop transfer function of the VSG command power to the VSG output power based on the deviation of the VSG output active power.

[0037] Step 2: Set the VSG command frequency to deviate from the grid's rated angular frequency and obtain the VSG steady-state power deviation;

[0038] Step 3: Identify the VSG equivalent damping coefficient based on the steady-state power deviation;

[0039] Step 4: Input the identified VSG equivalent damping coefficient into the closed-loop transfer function. After discretization, use the improved recursive auxiliary variable algorithm AFFRIVM, which introduces an adaptive variable forgetting factor λ(k), to iteratively identify the virtual inertial time constant of the VSG.

[0040] By combining the implementation of the above VSG virtual inertia identification method with the equivalent damping identification based on static power adjustment difference and the extended application of AFFRIVM, high-precision estimation of VSG equivalent damping and rapid and accurate identification of VSG virtual inertia are achieved, providing an accurate and feasible identification scheme for real-time online monitoring and operation control of VSG.

[0041] As an optional implementation, in step 1, a dynamic relationship model is constructed between the virtual inertia and equivalent damping coefficient of the VSG and the virtual mechanical power and virtual electromagnetic power. A closed-loop transfer function of the VSG command power to the VSG output power is established based on the deviation of the VSG output active power, including:

[0042] Step 1.1: During the system's dynamic process, based on the fact that the VSG virtual rotor angular frequency is approximately equal to the standard unit, a dynamic relationship model is established between the VSG's virtual inertia and equivalent damping coefficient and the virtual mechanical power and virtual electromagnetic power, expressed in incremental form as follows:

[0043] 2H VSG (dΔω VSG / dt)=ΔP m.VSG – ΔP e.VSG - D VSG Δω VSG ;

[0044] dΔδ VSG / dt = Δω VSG ω n ;

[0045] Among them, H VSG Δω represents the virtual inertial time constant of the VSG. VSG ΔP represents the change in the angular frequency of the VSG virtual rotor. m.VSG ΔP represents the power change of the VSG command. e.VSG Δδ represents the change in VSG output power. VSG ω is the change in the virtual power angle of the VSG. n D represents the reference value for the angular frequency of the VSG virtual rotor. VSG This represents the VSG equivalent damping coefficient;

[0046] Step 1.2: Use the output of the equivalent port after the grid-connected inverter to represent the equivalent VSG output, based on the VSG equivalent integrated power factor S. E.VSG Obtain the change in VSG output power ΔPe.VSG :

[0047] ΔP e.VSG = S E.VSG Δδ VSG ;

[0048] Step 1.3: Establish the closed-loop transfer function of VSG command power to VSG output power:

[0049] ΔP e.VSG (s) / ΔP m.VSG (s) = μ VSG / (s 2 +γ VSG s +μ VSG );

[0050] Where, γ VSG =D VSG / 2H VSG μ VSG =ω n S E.VS / 2H VSG In the formula, 's' represents the Laplace operator.

[0051] As an optional implementation, in step 2, the VSG command frequency is set to deviate from the grid's rated angular frequency to obtain the VSG steady-state power deviation, including:

[0052] When the power grid is operating in steady state, the time derivative of the virtual rotor angular frequency is 0, i.e., dΔω. VSG When / dt=0, the relationship between the VSG virtual rotor angular frequency, VSG command power, and VSG output power in steady state is obtained:

[0053] ΔP static = P e.VSG.N –P m.VSG.N = D VSG.N ω VSG.N (ω ref - ω VSG.N );

[0054] Where the subscript N represents the named value of the parameter, P e.VSG With P m.VSG These represent the VSG command power and VSG output power, respectively, ω ref This represents the VSG command angular frequency, which is the rated angular frequency of the power grid, ω. VSG Indicates the VSG virtual rotor angular frequency;

[0055] Based on the relationship between the VSG virtual rotor angular frequency, the VSG commanded power, and the VSG output power, a static angular frequency adjustment difference is generated by instructing the VSG, causing the VSG commanded angular frequency to deviate from the grid's rated angular frequency, i.e.: ω ref - ω VSG.N ≠0;

[0056] Determine the static power adjustment difference P based on the static angular frequency adjustment difference. e.VSG.N –P m.VSG.N The static power adjustment difference under steady state is taken as the steady-state power deviation of VSG.

[0057] As an optional implementation, in step 3, identifying the VSG equivalent damping coefficient based on the steady-state power deviation includes:

[0058] When the power grid is in steady-state operation, the actual operating angular frequency of the power grid is equal to its rated angular frequency. When the VSG is connected to the grid, it generates a static angular frequency adjustment difference through the aforementioned commands. By collecting data from the equivalent port of the grid-connected inverter after the VSG is connected to the grid, the static power adjustment difference is obtained as the steady-state power deviation of the VSG. The equivalent damping coefficient of the VSG is then identified according to the following formula:

[0059] D VSG.N = (P e.VSG.N –P m.VSG.N ) / [ω VSG.N (ω ref - ω VSG.N )];

[0060] Where, ω ref This indicates the VSG command angular frequency, which is the rated angular frequency of the power grid.

[0061] As an optional implementation, in step 4, the closed-loop transfer function is discretized based on the zero-order preserved discretization method, transforming the continuous-domain transfer function into a discrete-domain model:

[0062] ΔP e.VSG,f (k)= α 1.VSG ΔP e.VSG,f (k-1) + α 2.VSG ΔP e.VSG,f (k-2) + b0. VSG ,f u f (k);

[0063] Among them, T s Indicates the sampling period; α 1.VSG α 2.VSG b 0.VSGThese are the identification parameters; k represents the discrete time step, k=1,2,3,…,N, where N is the number of sampling points; the subscript f indicates that the parameter corresponds to the data after filtering preprocessing; u f (k) represents the input item, expressed as:

[0064] u f (k) = ΔP m.VSG,f (k) + 2ΔP m.VSG,f (k-1) + ΔP m.VSG,f (k-2);

[0065] Among them, the three identification parameters α 1.VSG α 2.VSG b 0.VSG The relationship with the continuous domain parameters is as follows:

[0066] ;

[0067] ;

[0068] .

[0069] Furthermore, in step 4, based on the recursive auxiliary variable algorithm that introduces an adaptive variable forgetting factor λ(k), the virtual inertial time constant of VSG is identified iteratively. In this process, by introducing an adaptive variable forgetting factor into the traditional RIVM, the influence of old data on the current identification is weakened, thereby improving the dynamic tracking capability.

[0070] Specifically, when the identification error is large (parameters have not converged), a smaller forgetting factor is used to quickly forget old data. When the identification error is small (parameters tend to stabilize), a larger genetic factor is used to retain valid historical data.

[0071] Below, we will combine the appendix... Figure 1 and appendix Figure 2 The example shown further illustrates the implementation process of the aforementioned method of the present invention.

[0072] Combination Figure 2 The example shown illustrates how a second-order transient model of a synchronous generator is used in the method of this invention to simulate the rotor motion equation of the synchronous generator using a VSG, thereby giving it inertial and damping characteristics.

[0073] As mentioned above, the dynamic relationship between the virtual inertia and equivalent damping coefficient of the virtual synchronous machine and the virtual mechanical power and virtual electromagnetic power can be expressed by the following standardized second-order equation:

[0074] 2H VSG (dω VSG / dt)=P m.VSG / ω VSG– P e.VSG / ω VSG - D VSG (ω VSG -1);

[0075] dδ VSG / dt = ω n (ω VSG -1);

[0076] Among them, H VSG ω represents the virtual inertial time constant of the VSG, in seconds; VSG P represents the VSG virtual rotor angular frequency; m.VSG The power supplied by the VSG command is equivalent to the mechanical power of a synchronous generator; P e.VSG The output power of the VSG is equivalent to the electromagnetic power of a synchronous generator; δ VSG For VSG virtual power angle; ω n The reference value representing the VSG virtual rotor angular frequency, in rad / s; D VSG This represents the VSG equivalent damping coefficient. It should be understood that all parameters in the above formulas are expressed in per-unit values.

[0077] Due to the VSG command power P m.VSG and VSG output power P e.VSG These are parameters that are easily observed; therefore, in this invention, the VSG is analogous to a synchronous generator, and P is established. e.VSG For P m.VSG The closed-loop transfer function, and thus based on P e.VSG For P m.VSG The dynamic response is used to identify the virtual inertia of VSG.

[0078] During the system's dynamic process, the virtual rotor angular frequency of the VSG does not change significantly, i.e. The above second-order equation can be transformed into incremental form as follows:

[0079] 2H VSG (dΔω VSG / dt)=ΔP m.VSG – ΔP e.VSG - D VSG Δω VSG ;

[0080] dΔδ VSG / dt = Δω VSG ω n ;

[0081] Among them, H VSG Δω represents the virtual inertial time constant of the VSG. VSGΔP represents the change in the angular frequency of the VSG virtual rotor. m.VSG ΔP represents the power change of the VSG command. e.VSG Δδ represents the change in VSG output power. VSG ω is the change in the virtual power angle of the VSG. n D represents the reference value for the angular frequency of the VSG virtual rotor. VSG This represents the VSG equivalent damping coefficient.

[0082] Combination Figure 2 In the example shown, the output characteristics of the VSG are equivalently replaced by the output of the equivalent port after the grid-connected inverter. Therefore, the data at the equivalent output port of the VSG is further used to identify the parameters.

[0083] Based on the actual impedance characteristics of medium- and high-voltage distribution networks, assuming the grid reactance is much greater than the resistance, and the VSG virtual reactance is much greater than the virtual resistance, then R g With R VSG Negligible and analogous to the analysis method for synchronous generators, the output active power at the equivalent port of the VSG can be expressed as:

[0084] P e.VSG = [E VSG U g / (X VSG +X g )]sinδ VSG

[0085] Therefore, given a fixed VSG virtual impedance and mains impedance, P e.VSG With VSG electromotive force amplitude E VSG Grid voltage amplitude U g and VSG virtual power angle δ VSG Related, at the same time U g Approximately unchanged, E VSG If it remains approximately constant, then the active power deviation of the VSG output can be further obtained:

[0086] ΔP e.VSG = Δδ VSG [∂P e.VSG / ∂δ VSG ]= S E.VSG Δδ VSG ;

[0087] In the formula, S E.VSG This represents the VSG equivalent step power factor.

[0088] Furthermore, a closed-loop transfer function is established for VSG command power on VSG output power, i.e., the relationship between VSG command power and output power:

[0089] ΔPe.VSG (s) / ΔP m.VSG (s) = μ VSG / (s 2 +γ VSG s +μ VSG );

[0090] in:

[0091] γ VSG =D VSG / 2H VSG ;

[0092] μ VSG =ω n S E.VS / 2H VSG ;

[0093] s represents the Laplace operator.

[0094] Furthermore, after the VSG reaches a steady state, the time derivative of the virtual rotor angular frequency is 0, i.e.: dΔω VSG If / dt=0, then the second-order equation is further transformed into:

[0095] P m.VSG / ω VSG - P e.VSG / ω VSG - D VSG (ω VSG -1)=0;

[0096] The above formula represents the commanded active power P of the VSG when it is in steady-state operation. m.VSG Output active power P e.VSG Equivalent damping coefficient D VSG and virtual rotational speed ω VSG To facilitate further understanding and analysis, the relationship between them is expressed in a named value form, where the named value is denoted by the subscript N:

[0097] ΔP static = P e.VSG.N –P m.VSG.N = D VSG.N ω VSG.N (ω ref - ω VSG.N );

[0098] In the formula, the subscript N represents the named value of the parameter, and P e.VSG With P m.VSG These represent the VSG command power and VSG output power, respectively, ω ref This represents the VSG command angular frequency, which is the rated angular frequency of the power grid, ω. VSGThis represents the angular frequency of the VSG virtual rotor.

[0099] This formula shows the named value ω of the VSG virtual angular frequency under steady state. VSG.N When the VSG output power is not equal to its command angular frequency, there is a difference P between the steady-state VSG output power and the commanded active power. e.VSG.N –P m.VSG.N , is defined as the static power adjustment difference.

[0100] Therefore, we generate a static angular frequency adjustment difference by instructing the VSG, causing the VSG command angular frequency to deviate from the grid's rated angular frequency, i.e.: ω ref - ω VSG.N ≠0; Determine the static power adjustment difference P based on the static angular frequency adjustment difference. e.VSG.N –P m.VSG.N The static power adjustment difference under steady state is taken as the steady-state power deviation of VSG.

[0101] It should be understood that when the power grid is in steady-state operation, the actual operating angular frequency of the power grid is equal to its rated angular frequency. When the VSG is connected to the grid, it generates a static angular frequency adjustment difference through commands, and obtains the static power adjustment difference as the steady-state power deviation of the VSG by collecting data from the equivalent port of the grid-connected inverter during VSG grid-connected operation. The equivalent damping coefficient of the VSG is then identified according to the following formula, i.e., the VSG equivalent damping coefficient identification based on the static power adjustment difference:

[0102] D VSG.N = (P e.VSG.N –P m.VSG.N ) / [ω VSG.N (ω ref - ω VSG.N )];

[0103] Where, ω ref This indicates the VSG command angular frequency, which is the rated angular frequency of the power grid.

[0104] Therefore, when the active power of the VSG command and the active power of the output are directly observable, in order to identify the value of the equivalent damping coefficient in steady state, a static angular frequency adjustment difference can be generated in the VSG, thereby obtaining the static power adjustment difference and identifying the equivalent damping coefficient of the VSG based on this difference.

[0105] In an embodiment of the present invention, we employ an improved recursive auxiliary variable algorithm (AFFRIVM) for VSG virtual inertia identification.

[0106] As an optional embodiment, in step 4, the closed-loop transfer function is discretized based on the zero-order preserved discretization method, transforming the continuous-domain transfer function into a discrete-domain model:

[0107] ΔP e.VSG,f (k)= α 1.VSG ΔP e.VSG,f (k-1) + α 2.VSG ΔP e.VSG,f (k-2) + b0. VSG ,f u f (k);

[0108] Among them, T s Indicates the sampling period; α 1.VSG α 2.VSG b 0.VSG These are the identification parameters; k represents the discrete time step, k=1,2,3,…,N, where N is the number of sampling points; the subscript f indicates that the parameter corresponds to the filtered preprocessed data to suppress measurement noise; u f (k) represents the input item, expressed as:

[0109] u f (k) = ΔP m.VSG,f (k) + 2ΔP m.VSG,f (k-1) + ΔP m.VSG,f (k-2);

[0110] Among them, the three identification parameters α 1.VSG α 2.VSG b 0.VSG The relationship with the continuous domain parameters is as follows:

[0111] ;

[0112] ;

[0113] .

[0114] Based on discretization, an adaptive variable forgetting factor λ(k) is introduced into the traditional recursive auxiliary variable algorithm to weaken the influence of historical data on the current identification process, thus forming the AFFRIVM algorithm.

[0115] As an optional implementation, an improved recursive auxiliary variable algorithm based on introducing an adaptive variable forgetting factor λ(k) is used to iteratively identify the virtual inertial time constant of the VSG, including the following steps:

[0116] Step 4.1: Based on the adaptive variable forgetting factor λ(k), the iterative formula for the improved recursive auxiliary variable algorithm is designed as follows:

[0117] ;

[0118] Wherein, the regression vector q VSG.f(k) is represented as:

[0119] [−ΔP e.VSG,f (k−1),−ΔP e.VSG,f (k−2),ΔP m.VSG,f (k)+2ΔP m.VSG,f (k−1)+ΔP m.VSG,f [(k−2)];

[0120] The auxiliary variable z(k) is expressed as:

[0121] [−ΔP e.VSG,f (k−1),−ΔP e.VSG,f (k−2),ΔP m.VSG,f (k)+2ΔP m.VSG,f (k−1)+ΔP m.VSG,f [(k−2)];

[0122] The identification parameters are represented in vector form as follows:

[0123] θ VSG (k)=[α 1.VSG (k), α 2.VSG (k), b 0.VSG (k)] T ;

[0124] Where P(k) is the covariance matrix, and its initial value is set to (10). 4 ~10 5 I is the initial iteration stability; K(k) is the gain matrix used to control the iteration convergence speed; and I is the identity matrix.

[0125] Step 4.2, according to the sampling period T s (For example, a value of 0.01 seconds) Collect data from the equivalent output port of the VSG to obtain the VSG command power and VSG output power, and perform filtering to obtain the filtered VSG power data P. m.VSG.f (k) and P e.VSG.f (k);

[0126] The filtered data is incremented to obtain ΔP. m.VSG.f (k) and ΔP e.VSG.f (k):

[0127] ΔP m.VSG.f (k)= P m.VSG.f (k)- P m.VSG.f (k-1);

[0128] ΔP e.VSG.f (k)= P e.VSG.f (k)- P e.VSG.f (k-1);

[0129] In this example, a low-pass filter (such as a first-order RC filter) is used for filtering to obtain filtered power data and eliminate high-frequency noise.

[0130] Step 4.3: Initialize the identification parameters, covariance matrix, and forgetting factor λ(k). Iterate and update the identification parameters and covariance matrix from k=1 to N until the identification parameters converge. During the iterative update process, the forgetting factor λ(k) is adaptively adjusted according to the identification error e(k).

[0131] Step 4.4: Substitute the converged identification parameters into the identification parameter α. 1.VSG α 2.VSG b 0.VSG The formula relating μ to the continuous domain parameter is used to solve for μ. VSG With γ VSG The estimated value; then based on H VSG With μ VSG γ VSG By reverse calculation of the relationship, the virtual inertial time constant of VSG can be obtained.

[0132] As an optional implementation, to ensure observability, an effective stimulus is applied to the VSG, such as a sudden load increase stimulus, and the simulation time is set to 3-10 seconds to trigger P. m.VSG and P e.VSG The dynamic response provides data support for identification. The applied stimulus covers the dynamic range of normal VSG operation, avoiding slow parameter convergence due to insufficient stimulus.

[0133] As an optional implementation, in step 4.3, the identification parameters, covariance matrix, and forgetting factor λ(k) are initialized, including:

[0134] Initialize the identification parameters: θ VSG (0) = [0,0,0] T ;

[0135] Initialize the covariance matrix: P(0) = (10 4 ~10 5 )I;

[0136] Initialize the forgetting factor: λ(0) = 0.95.

[0137] As an optional implementation, in step 4.3, the identification error is calculated in real time and the forgetting factor is adaptively adjusted during each iteration.

[0138] The identification error e(k) is defined as follows:

[0139] e(k) = ΔPe.VSG.f (k)- q VSG.f (k) θ VSG (k);

[0140] If |e(k)|>ε, the forgetting factor λ(k) = 0.8~0.9; |e(k)|≤ε, the forgetting factor λ(k) = 0.95~0.99;

[0141] Where ε represents the error threshold.

[0142] As an optional implementation, the convergence condition for the identification parameters is set as follows:

[0143] The change in the identification parameter is less than or equal to a preset threshold, which is set to 10. -6 ;

[0144] The covariance matrix tends to stabilize (no longer decreases significantly).

[0145] Furthermore, unlike the identification process for synchronous generators, when AFFRIVM is applied to the identification of VSG virtual inertia, it first uses the VSG equivalent damping identification method based on static power adjustment difference to identify the estimated value of equivalent damping, and then substitutes it into the aforementioned closed-loop transfer function to obtain the VSG virtual inertia.

[0146] Combination Figure 2 as well as Figures 3-10 The example shown is based on Figure 2 The diagram shows the structure of the VSG grid-connected system. The model performance simulation analysis is performed on this grid-connected system, and the applicability of the VSG virtual inertia identification scheme based on the equivalent damping identification of static power adjustment difference - AFFRIVM extended application is verified by simulation.

[0147] The simulation time was set to 8.0 s, and the VSG command frequencies were set to 49.9 Hz and 50.1 Hz. The commanded active power jumped from 10 kW to 20 kW at t=2.0 s, and the VSG equivalent damping coefficient was set to 10 p.u.

[0148] Combination Figure 3 The active power response diagram shows that when the command frequency deviates from the system's rated frequency, the VSG exhibits a static power adjustment difference. That is, during steady-state operation, there is always a difference between the commanded active power and the output active power. This phenomenon occurs because the VSG's command frequency is set to a value that is not equal to the rated frequency, resulting in a static power adjustment difference. This forces the VSG's equivalent damping coefficient to generate damping power to suppress the frequency deviation. Combined with... Figure 4As can be seen from the active power response diagram, there is a static power adjustment difference in the output active power of the VSG. At this time, since the command frequency is greater than the rated frequency, in order to prevent the frequency from shifting, the output active power is greater than the command active power.

[0149] Combined with the method in step 3 above, the VSG equivalent damping coefficient is identified. Figure 5 The equivalent damping coefficient identification results at a VSG command frequency of 49.9Hz show an equivalent damping coefficient of 9.9983 pu, which is very close to the actual value of 10 p.u. Furthermore, combined with... Figure 6 The equivalent damping coefficient identification results when the VSG command frequency is 50.1Hz show that the identified equivalent damping coefficient is 9.9876pu, which is very close to the true value of 10p.u. The identification accuracy is stable and reliable.

[0150] Furthermore, combining the method in step 4 above, the VSG virtual inertia is identified using the AFFRIVM method based on the value of the identified equivalent damping coefficient.

[0151] During the simulation, the VSG command frequency was adjusted back to its rated value of 50Hz, the equivalent damping coefficient was set to 10 p.u., and the virtual inertia time constant was set to 7.77s. The power frequency droop coefficient of the synchronous generator was set to 13kW / Hz, and the power frequency droop coefficient of the VSG was also set to 13kW / Hz. During the identification process, the identification method based on static power adjustment difference was updated in real time, and the AFFRIVM algorithm calculation was continuously updated and iterated. The simulation time was set to 3.5s, the sampling step size was 0.01s, and at t=2.0s, the load suddenly increased by 20kW. The identification result of the VSG virtual inertia time constant based on AFFRIVM is as follows. Figure 7 , Figure 8 , Figure 9 as well as Figure 10 As shown.

[0152] AFFRIVM is used to identify VSG virtual inertia by... Figure 8 It can be seen that the identification time window is [2.03s, 2.50s]. All identification parameters converge quickly, with the three parameters converging to -1.9997, 0.9997, and 1.28 × 10⁻⁶ respectively. -4 The convergence of these parameters indicates that the algorithm quickly found the optimal estimate of the system during the identification process. Figure 9 Furthermore, it can be seen that H is obtained based on these three identification parameters. VSG The method converged quickly, and the obtained identification value was 7.823s, which is very close to the set value of 7.770s, with a relative error of 0.68%, which verifies the effectiveness and high accuracy of the method in VSG virtual inertia identification.

[0153] Combination Figure 10 The results further compare the actual active power output of the VSG with that of the auxiliary model. The fit was 92.12% over the entire time period, and as high as 99.37% within the identification time window. This shows that AFFRIVM can accurately reflect the dynamic characteristics of the equivalent output port of the VSG during the identification process, and also confirms the reliability of the obtained identification values.

[0154] In conjunction with the VSG virtual inertia identification method based on static power adjustment difference equivalent damping identification in the above embodiments, the present invention also proposes a computer system, including:

[0155] One or more processors;

[0156] Memory stores instructions that can be operated.

[0157] When the instruction is executed by the one or more processors, it causes the one or more processors to perform an operation, which includes the process of executing the VSG virtual inertia identification method based on static power adjustment difference equivalent damping identification in any of the foregoing embodiments.

[0158] In conjunction with the VSG virtual inertia identification method based on static power adjustment difference equivalent damping identification in the above embodiments, the present invention also proposes a computer-readable storage medium for storing a computer program, the computer program including instructions / instruction sets executable by one or more processors.

[0159] When the instructions / instruction set are executed by the one or more processors, they implement the VSG virtual inertia identification method based on static power adjustment difference equivalent damping identification in any of the foregoing embodiments.

[0160] While the present invention has been disclosed above with reference to preferred embodiments, it is not intended to limit the invention. Those skilled in the art can make various modifications and refinements without departing from the spirit and scope of the invention. Therefore, the scope of protection of the present invention shall be determined by the claims.

Claims

1. A method for identifying VSG virtual inertia based on equivalent damping identification of static power adjustment difference, characterized in that, Includes the following steps: Step 1: Construct a dynamic relationship model between the virtual inertia and equivalent damping coefficient of the VSG and the virtual mechanical power and virtual electromagnetic power, and establish a closed-loop transfer function of the VSG command power to the VSG output power based on the deviation of the VSG output active power. Step 2: Set the VSG command frequency to deviate from the grid's rated angular frequency and obtain the VSG steady-state power deviation; Step 3: Identify the VSG equivalent damping coefficient based on the steady-state power deviation; Step 4: Input the identified VSG equivalent damping coefficient into the closed-loop transfer function. After discretization, use the recursive auxiliary variable algorithm with an adaptive variable forgetting factor λ(k) to iteratively identify the virtual inertial time constant of the VSG.

2. The VSG virtual inertia identification method based on static power adjustment difference equivalent damping identification according to claim 1, characterized in that, In step 1, a dynamic relationship model is constructed between the virtual inertia and equivalent damping coefficient of the VSG and the virtual mechanical power and virtual electromagnetic power. Based on the deviation of the VSG's output active power, a closed-loop transfer function of the VSG command power to the VSG output power is established, including: Step 1.1: During the system's dynamic process, based on the fact that the VSG virtual rotor angular frequency is approximately equal to the standard unit, establish a dynamic relationship model between the VSG's virtual inertia and equivalent damping coefficient and its virtual mechanical power and virtual electromagnetic power, and express it in incremental form as follows: 2H VSG (dSee VSG / dt)=ΔP m.VSG – ΔP e.VSG - D VSG Give VSG ? dDd VSG / dt = Do VSG oh n ? Among them, H VSG Δω represents the virtual inertial time constant of the VSG. VSG ΔP represents the change in the angular frequency of the VSG virtual rotor. m.VSG ΔP represents the power change of the VSG command. e.VSG Δδ represents the change in VSG output power. VSG ω is the change in the virtual power angle of the VSG. n D represents the reference value for the angular frequency of the VSG virtual rotor. VSG This represents the VSG equivalent damping coefficient; Step 1.2: Use the output of the equivalent port after the grid-connected inverter to represent the equivalent VSG output, based on the VSG equivalent integrated power factor S. E.VSG Obtain the change in VSG output power ΔP e.VSG : ΔP e.VSG = S E.VSG Dd VSG ; Step 1.3: Establish the closed-loop transfer function of VSG command power to VSG output power: ΔP e.VSG (s) / ΔP m.VSG (s) = μ VSG / (s 2 +g VSG s +μ VSG ); Where s represents the Laplace operator; γ VSG =D VSG / 2H VSG ; μ VSG =ω n S E.VS / 2H VSG 。 3. The VSG virtual inertia identification method based on static power adjustment difference equivalent damping identification according to claim 2, characterized in that, In step 2, the VSG command frequency is set to deviate from the grid's rated angular frequency to obtain the VSG steady-state power deviation, including: When the power grid is operating in steady state, the time derivative of the virtual rotor angular frequency is 0, i.e., dΔω. VSG With / dt=0, the relationship between the VSG virtual rotor angular frequency, VSG command power, and VSG output power in steady state is obtained: ΔP static =P e.VSG.N –P m.VSG.N =D VSG.N oh VSG.N (oh ref - oh VSG.N ); Where the subscript N represents the named value of the parameter, P e.VSG With P m.VSG These represent the VSG command power and VSG output power, respectively, ω ref This represents the VSG command angular frequency, which is the rated angular frequency of the power grid, ω. VSG Indicates the VSG virtual rotor angular frequency; Based on the relationship between the VSG virtual rotor angular frequency, the VSG commanded power, and the VSG output power, a static angular frequency adjustment difference is generated by instructing the VSG, causing the VSG commanded angular frequency to deviate from the grid's rated angular frequency, i.e.: ω ref - ω VSG.N ≠0; Determine the static power adjustment difference P based on the static angular frequency adjustment difference. e.VSG.N –P m.VSG.N The static power adjustment difference under steady state is taken as the steady-state power deviation of VSG.

4. The VSG virtual inertia identification method based on static power adjustment difference equivalent damping identification according to claim 2, characterized in that, In step 3, the equivalent damping coefficient of the VSG is identified based on the steady-state power deviation, including: When the power grid is in steady-state operation, the actual operating angular frequency of the power grid is equal to its rated angular frequency. When the VSG is connected to the grid, it generates a static angular frequency adjustment difference through the aforementioned commands. By collecting data from the equivalent port of the grid-connected inverter after the VSG is connected to the grid, the static power adjustment difference is obtained as the steady-state power deviation of the VSG. The equivalent damping coefficient of the VSG is then identified according to the following formula: D VSG.N = (P e.VSG.N –P m.VSG.N ) / [ω VSG.N (oh ref - oh VSG.N )]; Where, ω ref This indicates the VSG command angular frequency, which is the rated angular frequency of the power grid.

5. The VSG virtual inertia identification method based on static power adjustment difference equivalent damping identification according to claim 1, characterized in that, In step 4, the closed-loop transfer function is discretized using the zero-order preserved discretization method, transforming the continuous-domain transfer function into a discrete-domain model: ΔP e.VSG,f (k)= α 1.VSG ΔP e.VSG,f (k-1) + α 2.VSG ΔP e.VSG,f (k-2) + b0. VSG ,f u f (k); Among them, T s Indicates the sampling period; α 1.VSG α 2.VSG b 0.VSG These are the identification parameters; k represents the discrete time step, k=1,2,3,…,N, where N is the number of sampling points; the subscript f indicates that the parameter corresponds to the data after filtering preprocessing; u f (k) represents the input item, expressed as: u f (k)= ΔP m.VSG,f (k) + 2ΔP m.VSG,f (k-1) + ΔP m.VSG,f (k-2); Among them, the three identification parameters α 1.VSG α 2.VSG b 0.VSG The relationship with the continuous domain parameters is as follows: ； ； 。 6. The VSG virtual inertia identification method based on static power adjustment difference equivalent damping identification according to claim 5, characterized in that, In step 4, based on the recursive auxiliary variable algorithm that introduces an adaptive variable forgetting factor λ(k), the virtual inertial time constant of the VSG is identified iteratively, including the following steps: Step 4.1: Based on the adaptive variable forgetting factor λ(k), the iterative formula for the recursive auxiliary variable algorithm is designed as follows: ； Wherein, the regression vector q VSG.f (k) is represented as: [−ΔP e.VSG,f (k−1),−ΔP e.VSG,f (k−2),ΔP m.VSG,f (k)+2ΔP m.VSG,f (k−1)+ΔP m.VSG,f (k−2)]; The auxiliary variable z(k) is expressed as: [−ΔP e.VSG,f (k−1),−ΔP e.VSG,f (k−2),ΔP m.VSG,f (k)+2ΔP m.VSG,f (k−1)+ΔP m.VSG,f (k−2)]; The identification parameters are represented in vector form as follows: i VSG (k)=[a 1.VSG (k), a 2.VSG (k), b 0.VSG (k)] T ; Where K(k) is the gain matrix; P(k) is the covariance matrix; and I is the identity matrix. Step 4.2, according to the sampling period T s Data from the equivalent output port of the VSG is collected to obtain the VSG command power and VSG output power, and then filtered to obtain the filtered VSG power data P. m.VSG.f (k) and P e.VSG.f (k); The filtered data is incremented to obtain ΔP. m.VSG.f (k) and ΔP e.VSG.f (k): ΔP m.VSG.f (k)= P m.VSG.f (k)- P m.VSG.f (k-1); ΔP e.VSG.f (k)= P e.VSG.f (k)- P e.VSG.f (k-1); Step 4.3: Initialize the identification parameters, covariance matrix, and forgetting factor λ(k). Iterate and update the identification parameters and covariance matrix from k=1 to N until the identification parameters converge. During the iterative update process, the forgetting factor λ(k) is adaptively adjusted according to the identification error e(k). Step 4.4: Substitute the converged identification parameters into the identification parameter α. 1.VSG α 2.VSG b 0.VSG The formula relating μ to the continuous domain parameter is used to solve for μ. VSG With γ VSG The estimated value; then based on H VSG With μ VSG γ VSG By reverse calculation of the relationship, the virtual inertial time constant of VSG can be obtained.

7. The VSG virtual inertia identification method based on static power adjustment difference equivalent damping identification according to claim 6, characterized in that, In step 4.3, the identification parameters, covariance matrix, and forgetting factor λ(k) are initialized, including: Initialize the identification parameters: θ VSG (0) = [0,0,0] T ; Initialize the covariance matrix: P(0) = (10 4 ~10 5 )I; Initialize the forgetting factor: λ(0) = 0.

95.

8. The VSG virtual inertia identification method based on static power adjustment difference equivalent damping identification according to claim 6, characterized in that, In step 4.3, the identification error e(k) is defined as follows: e(k)= ΔP e.VSG.f (k)- q VSG.f (k) θ VSG (k); If |e(k)|>ε, the forgetting factor λ(k) = 0.8~0.9; if |e(k)|≤ε, the forgetting factor λ(k) = 0.95~0.99; Where ε represents the error threshold.

9. A computer system, characterized in that, include: One or more processors; The memory stores operable instructions that, when executed by the one or more processors, cause the one or more processors to perform operations, including the process of performing the VSG virtual inertia identification method based on static power adjustment difference equivalent damping identification as described in any one of claims 1-8.

10. A computer-readable storage medium for storing a computer program, characterized in that, The computer program includes instructions / instruction sets executable by one or more processors, which, when executed by the one or more processors, implement the process of the VSG virtual inertia identification method based on static power adjustment difference equivalent damping identification as described in any one of claims 1-8.

Images