Comprehensive inertia real-time estimation method and system containing network construction VSG, storage medium and electronic equipment

Abstract

The invention belongs to the technical field of power system operation control, and relates to a real-time estimation method and system for comprehensive inertia containing a network construction VSG, a storage medium and electronic equipment, and the method comprises the following steps: S1, constructing a system model; s2, inertia constant state conversion is carried out; s3, mutation detection and correction are carried out; aiming at the problem of real-time evaluation when inertia sudden change occurs in a power system, on the basis of an improved unscented Kalman filtering algorithm, conversion from parameter estimation to real-time state estimation and verification and correction of inertia sudden change conditions are realized, the real-time performance and accuracy of an inertia evaluation result are ensured, and the real-time performance of the power system is improved. And the evaluation method of inertia real-time estimation is enriched.

Description

TECHNICAL FIELD

[0001] The present application belongs to the technical field of power system operation control, and particularly relates to a comprehensive inertia real-time estimation method, system, storage medium and electronic device containing network VSG. BACKGROUND

[0002] With the promotion of the construction of the "double-high" power system, high-proportion new energy is connected to the grid to replace synchronous units, resulting in a decrease in system inertia and an increase in the risk of frequency instability. New energy is connected to the grid through power electronic devices, which lack physical inertia, exacerbating the problem of system inertia shortage. In order to maintain system frequency stability, virtual synchronous machine (VSG) control technology has emerged, which can simulate inertia to support the system. However, the volatility and faults of new energy and DC sources will cause the simulated inertia provided by them to change dynamically or even suddenly, threatening system safety. Therefore, it is urgent to study a real-time evaluation method of comprehensive inertia.

[0003] Existing methods such as offline evaluation methods based on disturbance power rely on accurate models and data, and have poor real-time performance; small disturbance methods based on wide-area measurement have weak anti-interference ability, and large disturbance methods have requirements for system stability; online methods have slow response and low accuracy in dynamic processes, for example, parameter identification methods such as extended Kalman filter have significant errors under large disturbances due to linearization processing. This is an insufficient aspect of the prior art. SUMMARY

[0004] The present application aims to solve the above technical problems by providing a comprehensive inertia real-time estimation method, system, storage medium and electronic device containing network VSG to address the deficiencies of the prior art.

[0005] To achieve the above-mentioned purpose, the present application provides the following technical solution: a comprehensive inertia real-time estimation method containing network VSG, comprising the following steps: constructing a frequency response model of multi-source inertia coupling under the inertia response time scale; based on the dynamic characteristics of the power system, modeling the comprehensive inertia constant as a state variable with time-varying characteristics and taking it as a system state variable to construct a nonlinear state space equation after completing the system model construction; then, using one-step prediction and state update of unscented Kalman filter to generate real-time estimation values of the comprehensive inertia of the power system; to deal with the problem that the real inertia of the system may suddenly change due to the switching of new energy and virtual inertia, after obtaining the real-time estimation values, an inertia sudden change detection mechanism and filter reset mechanism are introduced; by defining the modulus of the innovation sequence and taking a set threshold as a criterion, it is determined in real time whether the comprehensive inertia of the system has suddenly changed; once a sudden change is detected, the reset mechanism is triggered immediately to correct the initial state of the filter at the next time, thereby obtaining the real-time estimation curve of the comprehensive inertia of the power system.

[0006] As preferred, under the inertia response time scale, a multi-source inertia coupling frequency response model is constructed, specifically including: based on the dynamic equation of the power system, the influence of the new energy virtual inertia control strategy change and the regional power grid power transmission on the system comprehensive inertia is analyzed, including: the power system inertia represents the ability of the power system to suppress the change of frequency, which is related to the dynamic characteristics of the power system, and the specific equation expression is as follows:

[0007] Wherein, is the power angle of the first generator; is the inertia time constant of the first generator; is the angular velocity change of the first generator corresponding to; is the mechanical power change of the first generator; is the electromagnetic power change of the first generator; is the damping coefficient of the first generator; represents the derivative with respect to time . According to the multi-time scale theory, the key factors affecting the inertia response of the power system are analyzed, including: based on the multi-time scale theory, the frequency response of the power system can be divided into inertia response, primary frequency modulation and secondary frequency modulation stages. Inertia evaluation mainly occurs in the initial time after disturbance, that is, the inertia response stage. Due to the existence of frequency modulation dead zone, primary frequency modulation has not yet acted in this time period. Therefore, it can be considered that during this period, the mechanical power set value of the synchronous generator remains unchanged, that is, the mechanical power change is zero, that is, . Combined with the influence of new energy equivalent inertia and regional power grid interconnection, a frequency response model of multi-source inertia coupling power system is constructed. New energy adopts virtual inertia control mode to provide inertia support for the power system. Considering that the new energy inertia control strategy may be affected by system scheduling, unit access or large-capacity DC line fault, control strategy change and other factors, resulting in a sharp change in the system comprehensive inertia at a certain moment, which is expressed in the form of step change in numerical value, therefore, the equivalent inertia expression of new energy output fluctuation is written as:

[0008] Wherein, represents the equivalent inertia of the first new energy unit; and represent the equivalent inertia constant under different control strategies respectively; represents the time of control strategy change. The comprehensive inertia of the new energy power system and equivalent damping coefficient It can be represented as:

[0009]

[0010] in, , These represent the sets of synchronous machines and new energy generators using virtual inertia control, respectively. Indicates the first The inertial time constant of the synchronous generator; Indicates the first The virtual damping coefficient of the new energy generator unit using virtual inertia control; Indicates the first The damping coefficient of a synchronous generator. In the time scale of inertial response, the generator terminal voltage and inter-generator impedance are often considered constant. Therefore, the generator's electromagnetic power has a sinusoidal relationship with the inter-generator power angle difference, specifically:

[0011] in, The coefficient representing the change in electromagnetic power of each equivalent generator is a constant in the inertial response time scale; For the first The power angle of the generator; For the first The power angle of the generator. Therefore, the frequency response model expression considering the changes in the virtual inertia control strategy of new energy sources and the impact of regional power system interconnection is as follows:

[0012] in, , These represent the combined inertia constant and equivalent damping coefficient under the synchronous machine and new energy virtual inertia control methods, respectively, within this region; This represents the sum of the mechanical power of all generators within the region; The coefficient representing the change in electromagnetic power of each equivalent generator is a constant in the inertial response time scale; This represents the frequency change of the equivalent synchronous machine in that region. Based on the principle of the frequency response model, it can be expressed as... express ; For the first The power angle of the generator; For the first The power angle of the generator.

[0013] Preferably, based on the completed system model construction, and considering the dynamic characteristics of the power system, the comprehensive inertia constant is modeled as a state variable with time-varying characteristics, and used as the system state variable to construct a nonlinear state-space equation. Then, using one-step prediction and state update with unscented Kalman filtering, a real-time estimate of the power system's comprehensive inertia is generated. Specifically, during normal operation, the ideal comprehensive inertia of the power system is a constant. Using its differential equation, the comprehensive inertia constant is modeled as a state variable with time-varying characteristics, as follows:

[0014] in, This represents the overall inertia of the system. The state-space equations and observation equations in the unscented Kalman filter algorithm model are constructed as follows:

[0015] in, and These are state variables, input variables, and observation variables; and These are the nonlinear state equation and the observation equation, respectively. Based on the system dynamic equation, and taking the system's power angle deviation and frequency deviation as state variables, the system's overall inertia is modeled as a time-varying state variable, thus constructing the nonlinear state-space equation in the improved unscented Kalman filter algorithm model. The specific expression is as follows:

[0016] in, , This represents the system state vector to be estimated; and These represent the previous time and the current time, respectively. Indicates the time step; , and These represent Gaussian noise for the system state variables; , They represent the first generator time, The angle of attack at any moment; They represent the first generator time, The change in angular velocity at time t; Representing the system time, The combined inertia at any given moment; Indicates the first Mechanical power of the generator; Represents the inertial time constant; Indicates the first generator Active power at any given moment.

[0017] Based on the observables and state equations of the power system, the frequency changes and electromagnetic power changes of each generator are selected as observables. Considering the influence of system observation noise, the system observation equations are constructed, and the specific expressions are as follows:

[0018] in, , This represents the observation vector in the unscented Kalman filter algorithm; frequency change. electromagnetic power change As observations, initial values ​​for state variables and the estimated error covariance matrix are set using power system measurement data and past experience. Utilizing the first... State estimate at time 1 By using a proportionally symmetrical sampling method, the total... Sigma points are used as sampling points, denoted as . ,in , The dimension representing the system state variables. Utilizing... Calculation of the state quantity sigma point set at time 1 The sigma point set for the predicted state variables at time t is calculated as follows:

[0019] After setting the initial values ​​for the state variables and the estimation error covariance matrix, the one-step predicted values ​​of the state variables are obtained through unscented transformation and weighted averaging. The specific expression is as follows:

[0020] in, and They represent the first The one-step prediction of the state variables and the state covariance matrix at time 1; This represents the set of sigma sampling points for the predicted state value after unscented transformation; This represents the deviation between the predicted value at one step and the set of sampled points of the state variables; and These are the weights of the sampling points in the mean and covariance matrices, respectively. This represents the covariance matrix of the process noise. According to the... The improved unscented Kalman filter gain is calculated from the actual observations at time 1, resulting in the inertia estimate after state update. The specific expression is as follows:

[0021] in, Indicates the first The predicted observations of the sigma point set at time t; Indicates the first The predicted average of observations at any given time; This indicates the deviation between the predicted observation mean and the predicted observation value at the sigma point; This indicates the deviation between the actual observed value and the predicted observed mean; Indicates the first Actual observed value at any given time.

[0022] At the same time, according to the first The improved unscented Kalman filter gain is calculated from the actual observations at time 1, yielding the estimated state covariance after state update. The specific expression is as follows:

[0023] in, Indicates the first Kalman gain at time step; The covariance matrix representing the observation noise; This represents the updated system state estimate; This represents the updated state covariance matrix.

[0024] Preferably, after obtaining the real-time estimate, an inertia mutation detection mechanism and a filter reset mechanism are introduced. By defining the innovation sequence modulus and using a set threshold as a criterion, it is determined in real time whether a comprehensive inertia mutation occurs in the system. Once a mutation is detected, the reset mechanism is immediately triggered to correct the initial state of the filter at the next moment, thereby obtaining the real-time estimation curve of the power system's comprehensive inertia. Specifically, this includes defining the innovation sequence modulus based on the modulus of the difference between the actual value and the estimated value of the observations in the improved unscented Kalman filter model. The specific calculation method is as follows:

[0025] in, Indicates the modulus of the innovation sequence; Indicates in The actual observed value of the observation at a given time; express The expected value of the predicted value of the time-time observation.

[0026] Set an appropriate threshold for the new information sequence modulus. When the modulus of the new information sequence exceeds the set threshold, i.e. At that time, it is determined that a sudden change in inertia has occurred in the system, triggering the filter reset mechanism to correct the state covariance matrix for the next time step, specifically as follows:

[0027] in, This represents the initial value of the state error covariance. After passing through the state variable abrupt change detection mechanism and the filter reset mechanism, the comprehensive inertia is estimated for the next time step, ultimately yielding the real-time estimation curve of the power system's comprehensive inertia. The maximum inertia estimation error is introduced. And mean inertia estimation error As an error analysis index for system inertia estimation, the calculation method is as follows:

[0028] in, This represents the maximum value among all inertia estimates within the estimated time period; This represents the true value of inertia; This indicates the number of all inertia estimates within the estimated time period; Indicates the first The estimated value of inertia at time t; This represents the average inertia over the estimated time period.

[0029] Furthermore, this invention also provides a real-time estimation system for the integrated inertia of a structured power grid (VSG), comprising: a system model construction module configured to construct a frequency response model of multi-source inertia coupling at the inertia response time scale; an inertia constant state transformation module configured to, based on the completed system model construction and the dynamic characteristics of the power system, model the integrated inertia constant as a time-varying state variable, and use this as the system state variable to construct a nonlinear state-space equation. Then, using one-step prediction and state update of an unscented Kalman filter, a real-time estimate of the integrated inertia of the power system is generated; and a mutation detection and correction module configured to, to address the potential mutation in the system's real inertia caused by the switching of new energy sources and virtual inertia, introduce an inertia mutation detection mechanism and a filter reset mechanism after obtaining the real-time estimate. By defining the innovation sequence modulus and using a set threshold as a criterion, it is determined in real time whether a mutation in the integrated inertia of the system has occurred. Once a mutation is detected, the reset mechanism is immediately triggered to correct the initial state of the filter at the next moment, thereby obtaining the real-time estimation curve of the integrated inertia of the power system.

[0030] Preferably, in the system model construction module, a multi-source inertia coupled frequency response model is constructed at the inertia response time scale. Specifically, this includes: based on the power system dynamic equations, analyzing the impact of changes in new energy virtual inertia control strategies and regional power grid power transmission on the overall system inertia, including: power system inertia represents the power system's ability to suppress frequency changes, and it is related to the dynamic characteristics of the power system, with the specific relationship expression as follows:

[0031] in, For the first The power angle of the generator; For the first The inertial time constant of the generator; This represents the change in angular velocity corresponding to the generator node; For the first The change in mechanical power of the generator; For the first The change in electromagnetic power of the generator; For the first Damping coefficient of the generator; Indicates time The derivative of . Based on the multi-timescale theory, the key factors affecting the inertial response of a power system are analyzed, including: Based on the multi-timescale theory, the frequency response of a power system can be divided into stages such as inertial response, primary frequency regulation, and secondary frequency regulation. Inertial assessment mainly occurs in the initial moment after a disturbance, i.e., the inertial response stage. Due to the existence of the frequency regulation dead zone, primary frequency regulation has not yet taken effect during this period. Therefore, it can be assumed that during this period, the mechanical power setpoint of the synchronous generator remains unchanged, i.e., the change in mechanical power is zero. .

[0032] Combining the effects of the equivalent inertia of new energy sources and regional grid interconnection, a frequency response model of a multi-source inertia-coupled power system is constructed. New energy sources employ virtual inertia control to provide inertia support for the power system. Considering that the new energy inertia control strategy may be affected by factors such as system scheduling, unit access, faults in large-capacity DC lines, and changes in control strategies, leading to a sudden and dramatic change in the system's overall inertia at a certain moment, numerically exhibiting a step change, the equivalent inertia expression for the power output fluctuation of new energy sources is written as:

[0033] in, Indicates the first The equivalent inertia of the new energy unit; and These represent the equivalent inertia constants under different control strategies; Indicates the time when the control strategy change occurs. Overall inertia of the new energy power system. and equivalent damping coefficient It can be represented as:

[0034]

[0035] in, , These represent the sets of synchronous machines and new energy generators using virtual inertia control, respectively. Indicates the first The inertial time constant of the synchronous generator; Indicates the first The virtual damping coefficient of the new energy generator unit using virtual inertia control; Indicates the first Damping coefficient of the synchronous generator.

[0036] In the time scale of inertial response, the generator terminal voltage and inter-generator impedance are often considered constant. Therefore, the generator's electromagnetic power has a sinusoidal relationship with the inter-generator power angle difference, specifically:

[0037] in, The coefficient representing the change in electromagnetic power of each equivalent generator is a constant in the inertial response time scale; For the first The power angle of the generator; For the first The power angle of the generator.

[0038] Therefore, the frequency response model expression considering the changes in the virtual inertia control strategy of new energy sources and the impact of regional power system interconnection is as follows:

[0039] in, , These represent the combined inertia constant and equivalent damping coefficient under the synchronous machine and new energy virtual inertia control methods, respectively, within this region; This represents the sum of the mechanical power of all generators within the region; The coefficient representing the change in electromagnetic power of each equivalent generator is a constant in the inertial response time scale; This represents the frequency change of the equivalent synchronous machine in that region. Based on the principle of the frequency response model, it can be expressed as... express ; For the first The power angle of the generator; For the first The power angle of the generator.

[0040] Preferably, in the inertia constant state transformation module, based on the completion of the system model construction, the comprehensive inertia constant is modeled as a state variable with time-varying characteristics based on the dynamic characteristics of the power system, and used as the system state variable to construct a nonlinear state-space equation; then, using one-step prediction and state update of unscented Kalman filtering, a real-time estimate of the comprehensive inertia of the power system is generated, specifically including: during normal operation, the ideal comprehensive inertia of the power system is a constant, and using its differential equation, the comprehensive inertia constant is modeled as a state variable with time-varying characteristics, as follows:

[0041] in, This represents the overall inertia of the system.

[0042] The state-space equations and observation equations in the unscented Kalman filter algorithm model are constructed as follows:

[0043] in, and These are state variables, input variables, and observation variables; and These are the nonlinear state equation and the observation equation, respectively.

[0044] Based on the system dynamic equations, and taking the system's power angle deviation and frequency deviation as state variables, the system's overall inertia is modeled as a time-varying state variable. This leads to the construction of the nonlinear state-space equations in the improved unscented Kalman filter algorithm model, with the specific expression as follows:

[0045] in, , This represents the system state vector to be estimated; and These represent the previous time and the current time, respectively. Indicates the time step; , and These represent Gaussian noise for the system state variables; , They represent the first generator time, The angle of attack at any moment; They represent the first generator time, The change in angular velocity at time t; Representing the system time, The combined inertia at any given moment; Indicates the first Mechanical power of the generator; Represents the inertial time constant; Indicates the first generator Active power at any given moment.

[0046] Based on the observables and state equations of the power system, the frequency changes and electromagnetic power changes of each generator are selected as observables. Considering the influence of system observation noise, the system observation equations are constructed, and the specific expressions are as follows:

[0047] in, , This represents the observation vector in the unscented Kalman filter algorithm; frequency change. electromagnetic power change As an observation.

[0048] Initial values ​​for state variables and the estimated error covariance matrix are set using power system measurement data and past experience. Then, the first... State estimate at time 1 By using a proportionally symmetrical sampling method, the total... Sigma points are used as sampling points, denoted as . ,in , The dimension representing the system state variables. Utilizing... Calculation of the state quantity sigma point set at time 1 The sigma point set for the predicted state variables at time t is calculated as follows:

[0049] After setting the initial values ​​for the state variables and the estimation error covariance matrix, the one-step predicted values ​​of the state variables are obtained through unscented transformation and weighted averaging. The specific expression is as follows:

[0050] in, and They represent the first The one-step prediction of the state variables and the state covariance matrix at time 1; This represents the set of sigma sampling points for the predicted state value after unscented transformation; This represents the deviation between the predicted value at one step and the set of sampled points of the state variables; and These are the weights of the sampling points in the mean and covariance matrices, respectively. The covariance matrix represents the process noise.

[0051] According to the The improved unscented Kalman filter gain is calculated from the actual observations at time 1, resulting in the inertia estimate after state update. The specific expression is as follows:

[0052] in, Indicates the first The predicted observations of the sigma point set at time t; Indicates the first The predicted average of observations at any given time; This indicates the deviation between the predicted observation mean and the predicted observation value at the sigma point; This indicates the deviation between the actual observed value and the predicted observed mean; Indicates the first Actual observed value at any given time.

[0053] At the same time, according to the first The improved unscented Kalman filter gain is calculated from the actual observations at time 1, yielding the estimated state covariance after state update. The specific expression is as follows:

[0054] in, Indicates the first Kalman gain at time step; The covariance matrix representing the observation noise; This represents the updated system state estimate; This represents the updated state covariance matrix.

[0055] Preferably, in the mutation detection and correction module, after obtaining the real-time estimate, an inertia mutation detection mechanism and a filter reset mechanism are introduced; by defining the innovation sequence modulus and using a set threshold as a criterion, it is determined in real time whether a comprehensive inertia mutation has occurred in the system; once a mutation is detected, the reset mechanism is immediately triggered to correct the initial state of the filter at the next moment, thereby obtaining the real-time estimation curve of the comprehensive inertia of the power system, specifically including: Based on the modulus of the difference between the true and estimated values ​​of the observations in the improved unscented Kalman filter model, the modulus of the innovation sequence is defined as follows:

[0056] in, Indicates the modulus of the innovation sequence; Indicates in The actual observed value of the observation at a given time; express The expected value of the predicted value of the time-time observation.

[0057] Set an appropriate threshold for the new information sequence modulus. When the modulus of the new information sequence exceeds the set threshold, i.e. At that time, it is determined that a sudden change in inertia has occurred in the system, triggering the filter reset mechanism to correct the state covariance matrix for the next time step, specifically as follows:

[0058] in, This represents the initial value of the state error covariance.

[0059] After employing a state variable abrupt change detection mechanism and a filter reset mechanism, the comprehensive inertia is estimated for the next time step, ultimately yielding a real-time estimation curve of the power system's comprehensive inertia. A maximum inertia estimation error is introduced. And mean inertia estimation error As an error analysis index for system inertia estimation, the calculation method is as follows:

[0060] in, This represents the maximum value among all inertia estimates within the estimated time period; This represents the true value of inertia; This indicates the number of all inertia estimates within the estimated time period; Indicates the first The estimated value of inertia at time t; This represents the average inertia over the estimated time period.

[0061] The present invention also provides a computer-readable storage medium having a program stored thereon, which, when executed by a processor, implements the steps in the above-described method for real-time estimation of the integrated inertia of a structured network VSG.

[0062] The present invention also provides an electronic device, including a memory, a processor, and a program stored in the memory and executable on the processor, wherein the processor executes the program to implement the steps in the above-described method for real-time estimation of integrated inertia of a structured network VSG.

[0063] The beneficial effects of this invention are as follows: considering the influence of the virtual inertia of new energy sources and the power transmission of the regional power grid, starting from the time scale of the system's inertia response, and ignoring the influence of the primary and secondary frequency regulation of the system, a simplified power system frequency response model is constructed while ensuring the accuracy of the system's dynamic response. The dynamic equation of the system frequency is determined, and the main parameters affecting the system's inertia response are analyzed. Compared with the full-state response analysis method, this effectively reduces the analysis difficulty of complex systems and improves the calculation speed. Utilizing the dynamic characteristics of the system's comprehensive inertia, a differential equation for the system's comprehensive inertia is constructed. The comprehensive inertia is introduced as a state variable into the improved unscented Kalman filter algorithm, realizing the conversion from parameter estimation to trajectory estimation. The improved unscented Kalman filter algorithm used does not rely on a long time window and utilizes a symmetric sampling method and a weighted averaging method to achieve... Providing real-time estimates of state variables in each prediction-filtering loop ensures the real-time nature and accuracy of inertia assessment results. This invention addresses the problem of real-time estimation of the overall inertia when system inertia mutations occur. By introducing a state variable mutation detection mechanism and a filter reset mechanism on top of the traditional unscented Kalman filter algorithm, it achieves verification and correction of inertia mutations, enriching the evaluation methods for real-time inertia estimation. Compared with existing methods, this invention achieves online real-time estimation without relying on long time windows and accurate offline models, meeting the needs of real-time power grid scheduling and control. It can quickly and accurately track slow changes and instantaneous mutations in the system's overall inertia, significantly improving estimation accuracy and response speed in complex dynamic processes. It solves the problems of large errors and slow response of traditional methods (such as extended Kalman filtering) under large disturbances or parameter mutations.

[0064] Therefore, it is evident that the present invention has outstanding substantive features and significant progress compared with the prior art, and the beneficial effects of its implementation are also obvious. Attached Figure Description

[0065] To more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are only embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on the provided drawings without creative effort.

[0066] Figure 1 This is a flowchart of a real-time estimation method for the integrated inertia of a structured network VSG provided by the present invention.

[0067] Figure 2 This is a block diagram of the model for changing the virtual inertia control strategy of the system provided by the present invention.

[0068] Figure 3 This is a schematic diagram of the three-region AC synchronous interconnected power system structure provided by the present invention.

[0069] Figure 4 This is a simplified frequency response model for a single region provided by the present invention.

[0070] Figure 5 This invention provides a power system model that considers the virtual inertia of new energy sources and the interconnection of regional power systems.

[0071] Figure 6 This is the novel sequence modulus change curve provided by the present invention for a traditional unscented Kalman filter.

[0072] Figure 7 This is the curve showing the change in the modulus of the innovation sequence of the improved unscented Kalman filter provided by this invention.

[0073] Figure 8 This is the comprehensive inertia estimation result for each region provided by the present invention.

[0074] Figure 9 This is a comparison chart of the system integrated inertia estimation results provided by the present invention.

[0075] Figure 10 This is a block diagram of a real-time inertia estimation system for a structured network VSG provided by the present invention.

[0076] Among them, 1-system model construction module, 2-inertia constant state transformation module, 3-mutation detection and correction module. Detailed Implementation

[0077] The present invention will now be described in detail with reference to the accompanying drawings and specific embodiments. The following embodiments are explanations of the present invention, but the present invention is not limited to the following implementation methods.

[0078] Example 1: like Figure 1As shown in this embodiment, a real-time estimation method for the integrated inertia of a structured power grid (VSG) includes the following steps: Constructing a multi-source inertia-coupled frequency response model to characterize the dynamic relationship between system state variables and observations at the inertia response time scale; modeling the system's integrated inertia constant as a time-varying state variable, and using this time-varying state variable as the system state variable; constructing the state-space equation in the improved unscented Kalman filter algorithm model based on the multi-source inertia-coupled frequency response model; then using unscented transformation and weighted averaging to obtain a one-step prediction value of the system state variables, and updating it based on real observations to generate a real-time estimate of the power system's integrated inertia; after obtaining the real-time estimate of the power system's integrated inertia, introducing a state variable mutation detection mechanism: determining whether a mutation in the system's integrated inertia occurs by defining the innovation sequence modulus and a set threshold; once a mutation is detected, immediately triggering a filter reset mechanism to correct the initial value of the filter at the next moment, thereby obtaining the real-time estimation curve of the power system's integrated inertia.

[0079] The multi-source inertial coupling frequency response model represents the change in active power output from the inertial response of all new energy units and synchronous generators within the region. The specific expression is as follows:

[0080] in, , These represent the combined inertia constant and equivalent damping coefficient under the synchronous machine and new energy virtual inertia control methods, respectively, within this region; This represents the sum of the mechanical power of all generators within the region; The coefficient representing the change in electromagnetic power of each equivalent generator is a constant in the inertial response time scale; This represents the frequency change of the equivalent synchronous machine in that region. Based on the principle of the frequency response model, it can be expressed as... express ; For the first The power angle of the generator; For the first The power angle of the generator; The frequency response model of the multi-source inertia coupling is constructed based on the frequency response characteristics of a single generator. The specific equation expression of the frequency response characteristics of a single generator is as follows:

[0081] in,

[0082] Based on the frequency response characteristic expression of a single generator, a multi-source inertia coupled frequency response model is constructed: all new energy units and generators in the region are equivalent to a single generator, and its equivalent power angle can be expressed as: The sum of the active power changes output by the inertial response of all new energy units and generators within the region is... Among them, the comprehensive inertia constant of the new energy power system Due to the generator's own inertia and new energy units The equivalent inertia is composed of the following expression:

[0083] For the electromagnetic active power change portion of the inertial response output of all new energy units and generators within the region, under the time scale of the inertial response, it is generally assumed that the generator terminal voltage and inter-generator impedance are constants. The electromagnetic active power change of the generator has a sinusoidal relationship with the inter-generator power angle difference, specifically:

[0084] in, The coefficient representing the change in electromagnetic power of each equivalent generator is a constant in the inertial response time scale; For the first The power angle of the generator; For the first The power angle of the generator; n represents the number of generators in the system; therefore, the change in electromagnetic active power output of all new energy units and generators in the region can be equivalent to a single power generation unit, i.e. .

[0085] The change in mechanical active power output of all new energy generating units and generators within the region is the sum of the changes in mechanical active power output of multiple new energy generating units and generators. The active power variation affected by damping of all new energy units and generators within the region, and the equivalent damping coefficient of the new energy power system. Damped by synchronous generator Equivalent damping provided by new energy units (through control) Composition, expressed as follows:

[0086] The state-space equations and observation equations in the improved unscented Kalman filter algorithm model are as follows: The state-space equations are expressed as follows:

[0087] in, , This represents the system state variables to be estimated. and These represent the previous time and the current time, respectively. Indicates the time step; , and These represent Gaussian noise for the system state variables; , , express Estimates of the power angle, frequency deviation, and combined inertia constant at each moment; , They represent Estimates of the power angle, frequency deviation, and combined inertia constant at each moment; Indicates the first Mechanical power of the generator; Represents the inertial time constant; Indicates the first generator Active power at any given moment; The observation equation, specifically expressed as follows:

[0088] in, , This represents the observation vector in the unscented Kalman filter algorithm; frequency change. electromagnetic power change As an observation; The aforementioned state-space equations and observation equations are improvements upon the state-space equations and observation equations in the unscented Kalman filter algorithm model. The general forms of the state-space equations and observation equations in the unscented Kalman filter algorithm model are specifically expressed as follows:

[0089] in, and These are state variables, input variables, and observation variables; and These are the state-space equation and the observation equation, respectively. The state-space equation of the unscented Kalman filter is improved by modeling the system's integrated inertia as a state variable with time-varying characteristics. Based on the frequency response model, a state-space equation with the system's power angle deviation, frequency deviation, and integrated inertia as state variables is constructed and discretized to obtain the improved state-space equation of the unscented Kalman filter model. The specific expression is as follows:

[0090] in, , This represents the system state variables to be estimated. and These represent the previous time and the current time, respectively. Indicates the time step; , and These represent Gaussian noise for the system state variables; express Estimates of the power angle, frequency deviation, and combined inertia constant at each moment; They represent Estimates of the power angle, frequency deviation, and combined inertia constant at each moment; Indicates the first Mechanical power of the generator; Represents the inertial time constant; Indicates the first generator Active power at any given moment; Based on the state-space equations of the power system, the frequency changes and electromagnetic power changes of each generator are selected as observations. Considering the influence of system observation noise, the observation equations of the system are constructed, and the specific expressions are as follows:

[0091] in, , This represents the observation in the unscented Kalman filter algorithm; frequency change. electromagnetic power change As an observation; The method of recursively solving the equations in the unscented Kalman filter algorithm model using unscented transformation and weighted averaging yields one-step predictions and state updates of state variables, and generates real-time estimates of the integrated inertia of the power system. This includes: setting initial values ​​for state variables and the estimation error covariance matrix based on power system measurement data and past experience; and using the first... State estimate at time 1 The state variables at time k are sampled using a proportionally symmetric sampling method to obtain the total... Sigma points are used as sampling points, denoted as . ,in , Representing the dimensions of system state variables; reuse Calculation of the state quantity sigma point set at time 1 The sigma point set for the predicted state variables at time t is calculated as follows:

[0092] in, The state-space equations are as follows; The sigma point set is used as the initial value of the state variable, and is set to 0 based on past experience. A constant within a certain range is used as the initial value of the estimation error covariance matrix. Then, through unscented transformation and weighted averaging, the one-step predicted value of the state variables is obtained. The specific expression is as follows:

[0093] in, and They represent the first The one-step prediction of the state variables and the state covariance matrix at time 1; This represents the set of sigma sampling points for the predicted state value after unscented transformation; This represents the deviation between the predicted value at one step and the set of sampled points of the state variables; and These are the weights of the sampling points in the mean and covariance matrices, respectively. The covariance matrix represents the process noise; According to the The improved unscented Kalman filter gain is calculated from the actual observations at time 1, resulting in the inertia estimate after state update. The specific expression is as follows:

[0094] in, Indicates the first The predicted observations of the sigma point set at time t; Indicates the first The predicted average of observations at any given time; This indicates the deviation between the predicted observation mean and the predicted observation value at the sigma point; This indicates the deviation between the actual observed value and the predicted observed mean; Indicates the first Actual observed value at any given time; At the same time, according to the first The improved unscented Kalman filter gain is calculated from the actual observations at time 1, yielding the estimated state covariance after state update. The specific expression is as follows:

[0095] in, Indicates the first Kalman gain at time step; The covariance matrix representing the observation noise; This represents the updated system state estimate; This represents the updated state covariance matrix.

[0096] The modulus of the innovation sequence is:

[0097] in, Indicates the modulus of the innovation sequence; Indicates in The actual observed value of the observation at a given time; express The expected value of the predicted value of the observation at time step; the information sequence modulus is defined by standard based on the modulus of the difference between the true value and the estimated value of the observation in the improved unscented Kalman filter model; After defining the information sequence modulus, set an appropriate threshold for the information sequence modulus. When the modulus of the new information sequence exceeds the set threshold, i.e. At that time, it is determined that a sudden change in inertia has occurred in the system, triggering the filter reset mechanism to correct the state covariance matrix for the next time step, specifically as follows:

[0098] in, This represents the initial value of the state error covariance; After using the state variable mutation detection mechanism and the filter reset mechanism, the comprehensive inertia of the next moment is estimated, and finally the real-time estimation curve of the comprehensive inertia of the power system is obtained. Introducing the maximum inertia estimation error And mean inertia estimation error As an error analysis index for system inertia estimation, the calculation method is as follows:

[0099] in, This represents the maximum value among all inertia estimates within the estimated time period; This represents the true value of inertia; This indicates the number of all inertia estimates within the estimated time period; Indicates the first The estimated value of inertia at time t; This represents the average inertia over the estimated time period.

[0100] Example 2: like Figure 10 As shown in the figure, this embodiment provides a real-time estimation system for integrated inertia of a structured network VSG, comprising: System model building module 1 is configured to: build a multi-source inertial coupling frequency response model to characterize the dynamic relationship between system state variables and observations of the power system at the inertial response time scale; The inertia constant state transformation module 2 is configured to: model the system's comprehensive inertia constant as a state variable with time-varying characteristics, and use the modeled state variable comprehensive inertia constant as a system state variable; construct the state space equation in the improved unscented Kalman filter algorithm model based on the frequency response model of the multi-source inertia coupling; then use the unscented transformation and weighted average method to obtain the one-step prediction value of the system state variable, and update it based on the real observations, thereby generating a real-time estimate of the comprehensive inertia of the power system. The mutation detection and correction module 3 is configured to: after obtaining the real-time estimate of the power system's overall inertia, introduce a state variable mutation detection mechanism: determine whether a mutation in the system's overall inertia has occurred by defining the innovation sequence modulus and a set threshold; once a mutation is detected, immediately trigger the filter reset mechanism to correct the initial value of the filter at the next moment in order to obtain the real-time estimate curve of the power system's overall inertia.

[0101] The multi-source inertial coupling frequency response model represents the change in active power output from the inertial response of all new energy units and synchronous generators within the region. The specific expression is as follows:

[0102] in, , These represent the combined inertia constant and equivalent damping coefficient under the synchronous machine and new energy virtual inertia control methods, respectively, within this region; This represents the sum of the mechanical power of all generators within the region; The coefficient representing the change in electromagnetic power of each equivalent generator is a constant in the inertial response time scale; This represents the frequency change of the equivalent synchronous machine in that region. Based on the principle of the frequency response model, it can be expressed as... express ; For the first The power angle of the generator; For the first The power angle of the generator; The frequency response model of the multi-source inertia coupling is constructed based on the frequency response characteristics of a single generator. The specific equation expression of the frequency response characteristics of a single generator is as follows:

[0103] in, For the first The power angle of the generator; Indicates the first The derivative of the generator's power angle with respect to time; This represents the change in angular velocity corresponding to the generator node; For the first The change in the mechanical active power of the generator; For the first The change in electromagnetic active power of the generator; For the first Damping coefficient of the generator; Based on the frequency response characteristic expression of a single generator, a multi-source inertia coupled frequency response model is constructed: all new energy units and generators in the region are equivalent to a single generator, and its equivalent power angle can be expressed as: ; The sum of the active power changes output by the inertial response of all new energy generating units and generators in the region is Among them, the comprehensive inertia constant of the new energy power system Due to the generator's own inertia and new energy units The equivalent inertia is composed of the following expression:

[0104] For the electromagnetic active power change portion of the inertial response output of all new energy units and generators within the region, under the time scale of the inertial response, it is generally assumed that the generator terminal voltage and inter-generator impedance are constants. The electromagnetic active power change of the generator has a sinusoidal relationship with the inter-generator power angle difference, specifically:

[0105] in, The coefficient representing the change in electromagnetic power of each equivalent generator is a constant in the inertial response time scale; For the first The power angle of the generator; For the first The power angle of the generator; n represents the number of generators in the system; therefore, the change in electromagnetic active power output of all new energy units and generators in the region can be equivalent to a single power generation unit, i.e. ; The change in mechanical active power output of all new energy generating units and generators within the region is the sum of the changes in mechanical active power output of multiple new energy generating units and generators. The active power variation affected by damping of all new energy units and generators within the region, and the equivalent damping coefficient of the new energy power system. Damped by synchronous generator Equivalent damping provided by new energy units (through control) Composition, expressed as follows: .

[0106] The state-space equations and observation equations in the improved unscented Kalman filter algorithm model are as follows: The state-space equations are expressed as follows:

[0107] in, , This represents the system state variables to be estimated. and These represent the previous time and the current time, respectively. Indicates the time step; , and These represent Gaussian noise for the system state variables; express Estimates of the power angle, frequency deviation, and combined inertia constant at each moment; They represent Estimates of the power angle, frequency deviation, and combined inertia constant at each moment; Indicates the first Mechanical power of the generator; Represents the inertial time constant; Indicates the first generator Active power at any given moment; The observation equation, specifically expressed as follows:

[0108] in, , This represents the observation vector in the unscented Kalman filter algorithm; frequency change. electromagnetic power change As an observation; The aforementioned state-space equations and observation equations are improvements upon the state-space equations and observation equations in the unscented Kalman filter algorithm model. The general forms of the state-space equations and observation equations in the unscented Kalman filter algorithm model are specifically expressed as follows:

[0109] in, and These are state variables, input variables, and observation variables; and These are the state-space equation and the observation equation, respectively. The state-space equation of the unscented Kalman filter is improved by modeling the system's integrated inertia as a state variable with time-varying characteristics. Based on the frequency response model, a state-space equation with the system's power angle deviation, frequency deviation, and integrated inertia as state variables is constructed and discretized to obtain the state-space equation in the improved unscented Kalman filter algorithm model. The specific expression is as follows:

[0110] in, , This represents the system state vector to be estimated; and These represent the previous time and the current time, respectively. Indicates the time step; , and These represent Gaussian noise for the system state variables; They represent Estimates of the power angle, frequency deviation, and combined inertia constant at each moment; They represent Estimates of the power angle, frequency deviation, and combined inertia constant at each moment; Indicates the first Mechanical power of the generator; Represents the inertial time constant; Indicates the first generator Active power at any given moment; Based on the state-space equations of the power system, the frequency changes and electromagnetic power changes of each generator are selected as observations. Considering the influence of system observation noise, the observation equations of the system are constructed, and the specific expressions are as follows:

[0111] in, , This represents the observation vector in the unscented Kalman filter algorithm; frequency change. electromagnetic power change As an observation; The method of recursively solving the equations in the unscented Kalman filter algorithm model using unscented transformation and weighted averaging yields one-step predictions and state updates of state variables, and generates real-time estimates of the integrated inertia of the power system. This includes: setting initial values ​​for state variables and the estimation error covariance matrix based on power system measurement data and past experience; and using the first step... State estimate at time 1 The state variables at time k are sampled using a proportionally symmetric sampling method to obtain the total... Sigma points are used as sampling points, denoted as . ,in , Representing the dimensions of system state variables; reuse Calculation of the state quantity sigma point set at time 1 The sigma point set for the predicted state variables at time t is calculated as follows:

[0112] in, The state-space equations are as follows; The sigma point set is used as the initial value of the state variable, and is set to 0 based on past experience. A constant within a certain range is set as the initial value of the estimation error covariance matrix. Then, through unscented transformation and weighted averaging, the one-step predicted value of the state variable is obtained. The specific expression is as follows:

[0113] in, and They represent the first The one-step prediction of the state variables and the state covariance matrix at time 1; This represents the set of sigma sampling points for the predicted state value after unscented transformation; This represents the deviation between the predicted value at one step and the set of sampled points of the state variables; and These are the weights of the sampling points in the mean and covariance matrices, respectively. The covariance matrix represents the process noise; According to the The improved unscented Kalman filter gain is calculated from the actual observations at time 1, resulting in the inertia estimate after state update. The specific expression is as follows:

[0114] in, Indicates the first The predicted observations of the sigma point set at time t; Indicates the first The predicted average of observations at any given time; This indicates the deviation between the predicted observation mean and the predicted observation value at the sigma point; This indicates the deviation between the actual observed value and the predicted observed mean; Indicates the first Actual observed value at any given time; At the same time, according to the first The improved unscented Kalman filter gain is calculated from the actual observations at time 1, yielding the estimated state covariance after state update. The specific expression is as follows:

[0115] in, Indicates the first Kalman gain at time step; The covariance matrix representing the observation noise; This represents the updated system state estimate; This represents the updated state covariance matrix.

[0116] The novel sequence modulus in the mutation detection and correction module is:

[0117] in, Indicates the modulus of the innovation sequence; Indicates in The actual observed value of the observation at a given time; express The expected value of the predicted value of the observation at time step; the information sequence modulus is defined by standard based on the modulus of the difference between the true value and the estimated value of the observation in the improved unscented Kalman filter model; After defining the information sequence modulus, set an appropriate threshold for the information sequence modulus. When the modulus of the new information sequence exceeds the set threshold, i.e. At that time, it is determined that a sudden change in inertia has occurred in the system, triggering the filter reset mechanism to correct the state covariance matrix for the next time step, specifically as follows:

[0118] in, The initial value of the state error covariance is represented; after passing the state variable abrupt change detection mechanism and filter reset mechanism, the comprehensive inertia at the next moment is estimated, and finally the real-time estimation curve of the comprehensive inertia of the power system is obtained; the maximum inertia estimation error is introduced. And mean inertia estimation error As an error analysis index for system inertia estimation, the calculation method is as follows:

[0119] in, This represents the maximum value among all inertia estimates within the estimated time period; This represents the true value of inertia; This indicates the number of all inertia estimates within the estimated time period; Indicates the first The estimated value of inertia at time t; This represents the average inertia over the estimated time period.

[0120] Example 3: Taking a standard four-machine, two-region power system as an example, Region 1 contains two synchronous generators, and Region 2 contains one synchronous generator and one new energy unit employing a virtual inertia control strategy. The simulation system structure is as follows: Figure 5 As shown. The system rated frequency is set to 50Hz, the data time window length used for inertia assessment is 10s, the inertia time constants of generators G1 and G2 are 5s, and the inertia time constants of generators G3 and G4 are 4s. Among these, the control strategy of the new energy unit G4 changes at t=4s, and its provided virtual inertia changes from 4s to 6s. The element values ​​of the noise covariance matrix Q and the observation noise covariance matrix R of the unscented Kalman filter process are taken as follows: This value can be modified according to the actual noise level of the system. The threshold value for the innovation sequence modulus is set to [value missing]. To demonstrate the accuracy and real-time performance of the improved unscented Kalman filter algorithm in inertia estimation, initial inertia estimates different from the actual values ​​were selected. The initial inertia estimates for the four generators were 5.5, 4.7, 4.2, and 3.6, respectively.

[0121] Figure 3 This indicates that each area contains multiple synchronous generators and new energy units. and Representing regions and Interconnection line reactance and transmission power.

[0122] Figure 4 This represents a simplified frequency response model that takes into account changes in the virtual inertia control strategy for new energy sources and the impact of regional power system interconnection.

[0123] Figure 6 The graph shows the magnitude variation curve of the innovation sequence when the traditional unscented Kalman filter algorithm is applied to inertia estimation. The magnitude increases rapidly at t=4s, reaches its maximum at t=4.3s, and then slowly decreases. Figure 9 The corresponding comprehensive inertia estimation curve also slowly approaches the true value of the system inertia after the sudden change, and cannot estimate the true value of the state variable after the sudden change in real time and accurately.

[0124] Figure 7 This paper demonstrates the change curve of the innovation sequence modulus based on an improved unscented Kalman filter for inertia estimation. By introducing a state variable abrupt change detection mechanism and a filter reset mechanism, this embodiment detects an innovation sequence modulus exceeding a set threshold at 4.06 s and initiates the filter reset mechanism. After the reset, the innovation sequence modulus rapidly decreases below the threshold, and the inertia estimation curve can quickly and in real-time track the actual value after the abrupt change. Figure 8 , Figure 9 As shown.

[0125] pass Figure 8, Figure 9 The maximum inertia estimation error and the mean estimation error of the inertia estimation method based on the improved unscented Kalman filter can be obtained. It is calculated that the maximum estimation error and the mean estimation error of the combined inertia constant of each generator and the whole system are both less than 0.5%, which shows that the real-time inertia estimation value of the present invention has high accuracy and can adapt to the situation where the combined inertia of the system changes abruptly.

[0126] Example 4: The present invention also provides a computer-readable storage medium having a program stored thereon, which, when executed by a processor, implements the steps in the integrated inertia real-time estimation method for a structured network VSG described in Embodiment 1 above.

[0127] Example 5: The present invention also provides an electronic device, including a memory, a processor, and a program stored in the memory and executable on the processor. When the processor executes the program, it implements the steps in the integrated inertia real-time estimation method of a structured network VSG in Embodiment 1 described above.

[0128] The above-disclosed embodiments are merely preferred embodiments of the present invention, but the present invention is not limited thereto. Any non-creative variations that can be conceived by those skilled in the art, as well as any improvements and modifications made without departing from the principles of the present invention, should fall within the protection scope of the present invention.

Claims

1. A real-time estimation method for the integrated inertia of a structured network VSG, characterized in that, Includes the following steps: A multi-source inertial coupled frequency response model is constructed to characterize the dynamic relationship between system state variables and observations in the power system at the inertial response time scale. The system's overall inertia constant is modeled as a state variable with time-varying characteristics, and this modeled state variable is used as the system state variable. Based on the frequency response model of the multi-source inertia coupling, the state space equation in the improved unscented Kalman filter algorithm model is constructed. Then, the unscented transformation and weighted average method are used to obtain the one-step prediction value of the system state variable, and it is updated based on the real observations, thereby generating the real-time estimate of the power system's overall inertia. After obtaining the real-time estimate of the overall inertia of the power system, a state variable mutation detection mechanism is introduced: by defining the innovation sequence modulus and a set threshold, it is determined whether a mutation in the overall inertia of the system has occurred; once a mutation is detected, the filter reset mechanism is immediately triggered to correct the initial value of the filter at the next moment, so as to obtain the real-time estimate curve of the overall inertia of the power system.

2. The method for real-time estimation of integrated inertia of a VSG containing a structured network according to claim 1, characterized in that, The frequency response model of multi-source inertia coupling represents the change in active power output from the inertia response of all new energy units and synchronous generators within the region. The specific expression is as follows: in, , These represent the combined inertia constant and equivalent damping coefficient under the synchronous machine and new energy virtual inertia control methods, respectively, within this region; This represents the sum of the mechanical power of all generators within the region; The coefficient representing the change in electromagnetic power of each equivalent generator is a constant in the inertial response time scale; This represents the frequency change of the equivalent synchronous machine in that region. Based on the principle of the frequency response model, it can be expressed as... express ; For the first The power angle of the generator; For the first The power angle of the generator; The frequency response model of the multi-source inertia coupling is constructed based on the frequency response characteristics of a single generator. The specific equation expression of the frequency response characteristics of a single generator is as follows: in, For the first The power angle of the generator; Indicates the first The derivative of the generator's power angle with respect to time; For the first The change in angular velocity corresponding to each generator; For the first The inertial time constant of the generator; For the first The change in the mechanical active power of the generator; For the first The change in electromagnetic active power of the generator; For the first Damping coefficient of the generator; Based on the frequency response characteristic expression of a single generator, a multi-source inertia coupled frequency response model is constructed: If all new energy generating units and generators in the region are equivalent to a single generator, the equivalent power angle can be expressed as: ; The sum of the active power changes output by the inertial response of all new energy generating units and generators in the region is Among them, the comprehensive inertia constant of the new energy power system Due to the generator's own inertia and new energy units The equivalent inertia is composed of the following expression: For the electromagnetic active power change portion of the inertial response output of all new energy units and generators within the region, under the time scale of the inertial response, it is generally assumed that the generator terminal voltage and inter-generator impedance are constants. The electromagnetic active power change of the generator has a sinusoidal relationship with the inter-generator power angle difference, specifically: in, The coefficient representing the change in electromagnetic power of each equivalent generator is a constant in the inertial response time scale; For the first The power angle of the generator; For the first The power angle of the generator; n represents the number of generators in the system; therefore, the change in electromagnetic active power output of all new energy units and generators in the region can be equivalent to a single power generation unit, i.e. ; The change in mechanical active power output of all new energy generating units and generators within the region is the sum of the changes in mechanical active power output of multiple new energy generating units and generators. The active power variation affected by damping of all new energy units and generators within the region, and the equivalent damping coefficient of the new energy power system. Damped by synchronous generator Equivalent damping provided by new energy units (through control) Composition, expressed as follows: 。 3. The method for real-time estimation of integrated inertia of a VSG containing a structured network according to claim 1, characterized in that, The state-space equations and observation equations in the improved unscented Kalman filter algorithm model are as follows: The state-space equations are expressed as follows: in, , This represents the system state variables to be estimated. and These represent the previous time and the current time, respectively. Indicates the time step; , and These represent Gaussian noise for the system state variables; , , express Estimates of the power angle, frequency deviation, and combined inertia constant at each moment; , They represent Estimates of the power angle, frequency deviation, and combined inertia constant at each moment; Indicates the first Mechanical power of the generator; Represents the inertial time constant; Indicates the first generator Active power at any given moment; The observation equation, specifically expressed as follows: in, , This represents the observation vector in the unscented Kalman filter algorithm; frequency change. electromagnetic power change As an observation; The aforementioned state-space equations and observation equations are improvements upon the state-space equations and observation equations in the unscented Kalman filter algorithm model. The general forms of the state-space equations and observation equations in the unscented Kalman filter algorithm model are specifically expressed as follows: in, and These are state variables, input variables, and observation variables; and These are the state-space equation and the observation equation, respectively. The state-space equation of the unscented Kalman filter is improved by modeling the system's integrated inertia as a state variable with time-varying characteristics. Based on the frequency response model, a state-space equation with the system's power angle deviation, frequency deviation, and integrated inertia as state variables is constructed and discretized to obtain the improved state-space equation of the unscented Kalman filter model. The specific expression is as follows: in, , This represents the system state variables to be estimated. and These represent the previous time and the current time, respectively. Indicates the time step; , and These represent Gaussian noise for the system state variables; express Estimates of the power angle, frequency deviation, and combined inertia constant at each moment; They represent Estimates of the power angle, frequency deviation, and combined inertia constant at each moment; Indicates the first Mechanical power of the generator; Represents the inertial time constant; Indicates the first generator Active power at any given moment; Based on the state-space equations of the power system, the frequency changes and electromagnetic power changes of each generator are selected as observations. Considering the influence of system observation noise, the observation equations of the system are constructed, and the specific expressions are as follows: in, , This represents the observation in the unscented Kalman filter algorithm; frequency change. electromagnetic power change As an observation; The method of recursively solving the equations in the unscented Kalman filter algorithm model using unscented transformation and weighted averaging yields one-step predictions and state updates of state variables, and generates real-time estimates of the overall inertia of the power system, including: Based on power system measurement data and past experience, initial values ​​are set for the state variables and the estimation error covariance matrix: Using the first State estimate at time 1 The state variables at time k are sampled using a proportionally symmetric sampling method to obtain the total... Sigma points are used as sampling points, denoted as . ,in , Representing the dimensions of system state variables; reuse Calculation of the state quantity sigma point set at time 1 The sigma point set for the predicted state variables at time t is calculated as follows: in, The state-space equations are as follows; The sigma point set is used as the initial value of the state variable, and is set to 0~ based on past experience. The constants within the range are used as the initial values ​​of the estimation error covariance matrix. Then, through unscented transformation and weighted averaging, the one-step predicted values ​​of the state variables are obtained. The specific expression is as follows: in, and They represent the first The one-step prediction of the state variables and the state covariance matrix at time 1; This represents the set of sigma sampling points for the predicted state value after unscented transformation; This represents the deviation between the predicted value at one step and the set of sampled points of the state variables; and These are the weights of the sampling points in the mean and covariance matrices, respectively. The covariance matrix represents the process noise; According to the The improved unscented Kalman filter gain is calculated from the actual observations at time 1, resulting in the inertia estimate after state update. The specific expression is as follows: in, Indicates the first The predicted observations of the sigma point set at time t; Indicates the first The predicted average of observations at any given time; This indicates the deviation between the predicted observation mean and the predicted observation value at the sigma point; This indicates the deviation between the actual observed value and the predicted observed mean; Indicates the first Actual observed value at any given time; At the same time, according to the first The improved unscented Kalman filter gain is calculated from the actual observations at time 1, yielding the estimated state covariance after state update. The specific expression is as follows: in, Indicates the first Kalman gain at time step; The covariance matrix representing the observation noise; This represents the updated system state estimate; This represents the updated state covariance matrix.

4. The method for real-time estimation of integrated inertia of a VSG containing a structured network according to claim 1, characterized in that, The modulus of the innovation sequence is: in, Indicates the modulus of the innovation sequence; Indicates in The actual observed value of the observation at a given time; express The expected value of the predicted value of the observation at time step; the information sequence modulus is defined by standard based on the modulus of the difference between the true value and the estimated value of the observation in the improved unscented Kalman filter model; After defining the information sequence modulus, set an appropriate threshold for the information sequence modulus. When the modulus of the new information sequence exceeds the set threshold, i.e. At that time, it is determined that a sudden change in inertia has occurred in the system, triggering the filter reset mechanism to correct the state covariance matrix for the next time step, specifically as follows: in, This represents the initial value of the state error covariance; After using the state variable mutation detection mechanism and the filter reset mechanism, the comprehensive inertia of the next moment is estimated, and finally the real-time estimation curve of the comprehensive inertia of the power system is obtained. Introducing the maximum inertia estimation error And mean inertia estimation error As an error analysis index for system inertia estimation, the calculation method is as follows: in, This represents the maximum value among all inertia estimates within the estimated time period; This represents the true value of inertia; This indicates the number of all inertia estimates within the estimated time period; Indicates the first The estimated value of inertia at time t; This represents the average inertia over the estimated time period.

5. A real-time estimation system for integrated inertia including a structured network VSG, characterized in that, include: The system model building module is configured to: build a multi-source inertial coupling frequency response model to characterize the dynamic relationship between system state variables and observations of the power system at the inertial response time scale; The inertia constant state transformation module is configured to: model the system's comprehensive inertia constant as a state variable with time-varying characteristics, and use the modeled state variable comprehensive inertia constant as the system state variable; construct the state space equation in the improved unscented Kalman filter algorithm model based on the frequency response model of the multi-source inertia coupling; then use the unscented transformation and weighted average method to obtain the one-step prediction value of the system state variable, and update it based on the real observations, thereby generating a real-time estimate of the comprehensive inertia of the power system; The mutation detection and correction module is configured to: after obtaining the real-time estimate of the power system's overall inertia, introduce a state variable mutation detection mechanism: determine whether a mutation in the system's overall inertia has occurred by defining the innovation sequence modulus and a set threshold; once a mutation is detected, immediately trigger the filter reset mechanism to correct the initial value of the filter at the next moment in order to obtain the real-time estimate curve of the power system's overall inertia.

6. The integrated inertia real-time estimation system for a structured network VSG according to claim 5, characterized in that, The multi-source inertial coupling frequency response model represents the change in active power output from the inertial response of all new energy units and synchronous generators within the region. The specific expression is as follows: in, , These represent the combined inertia constant and equivalent damping coefficient under the synchronous machine and new energy virtual inertia control methods, respectively, within this region; This represents the sum of the mechanical power of all generators within the region; The coefficient representing the change in electromagnetic power of each equivalent generator is a constant in the inertial response time scale; This represents the frequency change of the equivalent synchronous machine in that region. Based on the principle of the frequency response model, it can be expressed as... express ; For the first The power angle of the generator; For the first The power angle of the generator; The frequency response model of the multi-source inertia coupling is constructed based on the frequency response characteristics of a single generator. The specific equation expression of the frequency response characteristics of a single generator is as follows: in, For the first The power angle of the generator; Indicates the first The derivative of the generator's power angle with respect to time; This represents the change in angular velocity corresponding to the generator node; For the first The change in the mechanical active power of the generator; For the first The change in electromagnetic active power of the generator; For the first Damping coefficient of the generator; Based on the frequency response characteristic expression of a single generator, a multi-source inertia coupled frequency response model is constructed: If all new energy generating units and generators in the region are equivalent to a single generator, the equivalent power angle can be expressed as: ; The sum of the active power changes output by the inertial response of all new energy generating units and generators in the region is Among them, the comprehensive inertia constant of the new energy power system Due to the generator's own inertia and new energy units The equivalent inertia is composed of the following expression: For the electromagnetic active power change portion of the inertial response output of all new energy units and generators within the region, under the time scale of the inertial response, it is generally assumed that the generator terminal voltage and inter-generator impedance are constants. The electromagnetic active power change of the generator has a sinusoidal relationship with the inter-generator power angle difference, specifically: in, The coefficient representing the change in electromagnetic power of each equivalent generator is a constant in the inertial response time scale; For the first The power angle of the generator; For the first The power angle of the generator; n represents the number of generators in the system; therefore, the change in electromagnetic active power output of all new energy units and generators in the region can be equivalent to a single power generation unit, i.e. ; The change in mechanical active power output of all new energy generating units and generators within the region is the sum of the changes in mechanical active power output of multiple new energy generating units and generators. The active power variation affected by damping of all new energy units and generators within the region, and the equivalent damping coefficient of the new energy power system. Damped by synchronous generator Equivalent damping provided by new energy units (through control) Composition, expressed as follows: 。 7. The integrated inertia real-time estimation system for a structured network VSG according to claim 5, characterized in that, The state-space equations and observation equations in the improved unscented Kalman filter algorithm model are as follows: The state-space equations are expressed as follows: in, , This represents the system state variables to be estimated. and These represent the previous time and the current time, respectively. Indicates the time step; , and These represent Gaussian noise for the system state variables; express Estimates of the power angle, frequency deviation, and combined inertia constant at each moment; They represent Estimates of the power angle, frequency deviation, and combined inertia constant at each moment; Indicates the first Mechanical power of the generator; Represents the inertial time constant; Indicates the first generator Active power at any given moment; The observation equation, specifically expressed as follows: in, , This represents the observation vector in the unscented Kalman filter algorithm; frequency change. electromagnetic power change As an observation; The aforementioned state-space equations and observation equations are improvements upon the state-space equations and observation equations in the unscented Kalman filter algorithm model. The general forms of the state-space equations and observation equations in the unscented Kalman filter algorithm model are specifically expressed as follows: in, and These are state variables, input variables, and observation variables; and These are the state-space equation and the observation equation, respectively. The state-space equation of the unscented Kalman filter is improved by modeling the system's integrated inertia as a state variable with time-varying characteristics. Based on the frequency response model, a state-space equation with the system's power angle deviation, frequency deviation, and integrated inertia as state variables is constructed and discretized to obtain the state-space equation in the improved unscented Kalman filter algorithm model. The specific expression is as follows: in, , This represents the system state vector to be estimated; and These represent the previous time and the current time, respectively. Indicates the time step; , and These represent Gaussian noise for the system state variables; They represent Estimates of the power angle, frequency deviation, and combined inertia constant at each moment; They represent Estimates of the power angle, frequency deviation, and combined inertia constant at each moment; Indicates the first Mechanical power of the generator; Represents the inertial time constant; Indicates the first generator Active power at any given moment; Based on the state-space equations of the power system, the frequency changes and electromagnetic power changes of each generator are selected as observations. Considering the influence of system observation noise, the observation equations of the system are constructed, and the specific expressions are as follows: in, , This represents the observation vector in the unscented Kalman filter algorithm; frequency change. electromagnetic power change As an observation; The method of recursively solving the equations in the unscented Kalman filter algorithm model using unscented transformation and weighted averaging yields one-step predictions and state updates of state variables, and generates real-time estimates of the overall inertia of the power system, including: Initial values ​​for state variables and the estimated error covariance matrix are set based on power system measurement data and past experience. Using the first State estimate at time 1 The state variables at time k are sampled using a proportionally symmetric sampling method to obtain the total... Sigma points are used as sampling points, denoted as . ,in , Representing the dimensions of system state variables; reuse Calculation of the state quantity sigma point set at time 1 The sigma point set for the predicted state variables at time t is calculated as follows: in, The state-space equations are as follows; The sigma point set is used as the initial value of the state variable, and is set to 0~ based on past experience. A constant within a certain range is set as the initial value of the estimation error covariance matrix. Then, through unscented transformation and weighted averaging, the one-step predicted value of the state variable is obtained. The specific expression is as follows: in, and They represent the first The one-step prediction of the state variables and the state covariance matrix at time 1; This represents the set of sigma sampling points for the predicted state value after unscented transformation; This represents the deviation between the predicted value at one step and the set of sampled points of the state variables; and These are the weights of the sampling points in the mean and covariance matrices, respectively. The covariance matrix represents the process noise; According to the The improved unscented Kalman filter gain is calculated from the actual observations at time 1, resulting in the inertia estimate after state update. The specific expression is as follows: in, Indicates the first The predicted observations of the sigma point set at time t; Indicates the first The predicted average of observations at any given time; This indicates the deviation between the predicted observation mean and the predicted observation value at the sigma point; This indicates the deviation between the actual observed value and the predicted observed mean; Indicates the first Actual observed value at any given time; At the same time, according to the first The improved unscented Kalman filter gain is calculated from the actual observations at time 1, yielding the estimated state covariance after state update. The specific expression is as follows: in, Indicates the first Kalman gain at time step; The covariance matrix representing the observation noise; This represents the updated system state estimate; This represents the updated state covariance matrix.

8. The integrated inertia real-time estimation system for a structured network VSG according to claim 5, characterized in that, The novel sequence modulus in the mutation detection and correction module is: in, Indicates the modulus of the innovation sequence; Indicates in The actual observed value of the observation at a given time; express The expected value of the predicted value of the observation at time step; the information sequence modulus is defined by standard based on the modulus of the difference between the true value and the estimated value of the observation in the improved unscented Kalman filter model; After defining the information sequence modulus, set an appropriate threshold for the information sequence modulus. When the modulus of the new information sequence exceeds the set threshold, i.e. At that time, it is determined that a sudden change in inertia has occurred in the system, triggering the filter reset mechanism to correct the state covariance matrix for the next time step, specifically as follows: in, This represents the initial value of the state error covariance; After using the state variable mutation detection mechanism and the filter reset mechanism, the comprehensive inertia of the next moment is estimated, and finally the real-time estimation curve of the comprehensive inertia of the power system is obtained. Introducing the maximum inertia estimation error And mean inertia estimation error As an error analysis index for system inertia estimation, the calculation method is as follows: in, This represents the maximum value among all inertia estimates within the estimated time period; This represents the true value of inertia; This indicates the number of all inertia estimates within the estimated time period; Indicates the first The estimated value of inertia at time t; This represents the average inertia over the estimated time period.

9. A computer-readable storage medium having a program stored thereon, characterized in that, When executed by the processor, the program implements the steps in the real-time estimation method for the integrated inertia of a structured network VSG as described in any one of claims 1-4.

10. An electronic device comprising a memory, a processor, and a program stored in the memory and executable on the processor, characterized in that, When the processor executes the program, it implements the steps in the integrated inertia real-time estimation method of a structured network VSG as described in any one of claims 1-4.

Images