Method and device for identifying inertia and damping of power system

Abstract

The present invention belongs to the technical field of power systems, and provides a method for identifying the inertia and damping of a power system. By respectively inputting the real-time PMU sequence, environmental information, and system load of the target power system into a black-box model and a white-box model, the first inertia and the first damping output by the black-box model, and the second inertia and the second damping output by the white-box model are obtained; the inertia of the target power system is determined based on the first inertia and the second inertia, and the damping of the target power system is determined based on the first damping and the second damping. The present invention can obtain more accurate identification results of inertia and damping, improve the real-time online situation awareness and prediction ability of the power grid, effectively improve the frequency characteristic perception ability in areas with a high proportion of new energy power sources and the receiving end areas of DC interconnected power grids, further improve the active power control ability and frequency safety of the system, ensure the safe and stable operation of the power grid, and provide a basis for power system operation decision-making in an uncertain environment.

Description

Technical Field

[0001] The present invention relates to the technical field of power systems, and in particular, to a method and device for identifying the inertia and damping of a power system. Background Art

[0002] Inertia and damping are the inherent properties of a power system and are the basis for ensuring the safe and stable operation of the power system. If the system inertia is too low, it will cause the grid frequency to quickly drop to the threshold of under-frequency load shedding in the event of a fault, and a large power outage disaster will occur before the system frequency regulation service has time to act. If the system damping is too low, it will cause large oscillations in the grid frequency or power angle, resulting in the disconnection or damage of power plants.

[0003] Currently, a large number of distributed new energy sources are connected to the grid. These asynchronous power sources are decoupled from the grid frequency and cannot actively provide inertia support for the system under active power disturbances, which will lead to a reduction in the inertia and damping levels of the power system, deteriorate the system's anti-disturbance ability and stability characteristics.

[0004] To avoid the above problems, domestic and foreign researchers have proposed many methods, such as using system reserve capacity and battery energy storage to smooth the volatility of new energy power generation, or using virtual inertia control technology to improve the system's equivalent inertia. However, there is currently a lack of effective means to identify the inertia and damping of a power system, making it difficult to evaluate the effects of the above methods. Summary of the Invention

[0005] Embodiments of the present invention provide a method and device for identifying the inertia and damping of a power system to solve the problem of identifying the inertia and damping of a power system.

[0006] In a first aspect, embodiments of the present invention provide a method for identifying the inertia and damping of a power system, including:

[0007] Obtain a black-box model and a white-box model of the target power system; wherein, the black-box model is constructed based on an LSTM neural network, the white-box model is constructed based on the frequency dynamic process mechanism model of the target power system, and the inputs of the black-box model and the white-box model are both PMU sequences, environmental information, and system load, and the outputs are both inertia and damping;

[0008] Input the real-time PMU sequence, environmental information, and system load of the target power system into the black-box model and the white-box model respectively to obtain the first inertia and the first damping output by the black-box model, and the second inertia and the second damping output by the white-box model;

[0009] Determine the inertia of the target power system based on the first inertia and the second inertia, and determine the damping of the target power system based on the first damping and the second damping.

[0010] In a possible implementation, determining the inertia of the target power system based on the first inertia and the second inertia, and determining the damping of the target power system based on the first damping and the second damping includes:

[0011] If the difference between the first inertia and the second inertia is less than a preset inertia threshold, and the difference between the first damping and the second damping is less than a preset damping threshold, then use the first inertia as the inertia of the target power system and the first damping as the damping of the target power system.

[0012] In a possible implementation, determining the inertia of the target power system based on the first inertia and the second inertia, and determining the damping of the target power system based on the first damping and the second damping further includes:

[0013] If the difference between the first inertia and the second inertia is greater than a preset inertia threshold, or the difference between the first damping and the second damping is greater than a preset damping threshold, then obtain the gray-box model of the target power system; wherein, the gray-box model is constructed based on the LSTM neural network and the frequency dynamic process mechanism model of the target power system, the inputs of the gray-box model are the PMU sequence, environmental information, and system load, and the outputs are inertia and damping;

[0014] Input the real-time PMU sequence, environmental information, and system load of the target power system into the gray-box model, and use the inertia and damping output by the gray-box model as the inertia and damping of the target power system.

[0015] In a possible implementation, the gray-box model includes a first sub-model and a second sub-model. The first sub-model is constructed based on the frequency dynamic process mechanism model of the target power system. The input of the first sub-model is the PMU sequence, environmental information, and system load, and the output is the state matrix of the target power system. The second sub-model is constructed based on the LSTM neural network. The input of the second sub-model is the state matrix of the target power system, and the outputs are inertia and damping.

[0016] In a possible implementation, obtaining the black-box model of the target power system includes:

[0017] Obtain the full sample set; wherein, the full sample set includes multiple samples, each sample is the PMU sequence, environmental information, and system load of the target power system at different historical moments, and the label of each sample is the inertia and damping corresponding to the sample;

[0018] Select the samples corresponding to the typical time points in the full sample set to construct a typical sample set, and construct a normal sample set based on the remaining samples; wherein, the typical time points refer to the time points when active power disturbances occur in the target power system;

[0019] Randomly allocate the typical sample set and the normal sample set according to a preset ratio to obtain a training sample set and a validation sample set;

[0020] Train the LSTM neural network based on the training sample set, and verify the trained LSTM neural network based on the validation sample set to obtain the black-box model of the target power system.

[0021] In a possible implementation, obtaining the white-box model of the target power system includes:

[0022] Obtain the full sample set; where the full sample set includes multiple samples, each sample is the PMU sequence, environmental information, and system load of the target power system at different historical moments, and the label of each sample is the inertia and damping corresponding to the sample;

[0023] Select the samples corresponding to the typical time points in the full sample set to construct a typical sample set, and construct a normal sample set based on the remaining samples; where the typical time points refer to the time points when active power disturbances occur in the target power system;

[0024] Randomly allocate the typical sample set and the normal sample set according to a preset ratio to obtain a training sample set and a validation sample set;

[0025] Train the frequency dynamic process mechanism model of the target power system based on the training sample set and the parameter estimation algorithm, and verify the trained frequency dynamic process mechanism model based on the validation sample set to obtain the white-box model of the target power system.

[0026] In a second aspect, an embodiment of the present invention provides a device for identifying the inertia and damping of a power system, including:

[0027] An acquisition module, configured to acquire the black-box model and the white-box model of the target power system; where the black-box model is constructed based on the LSTM neural network, the white-box model is constructed based on the frequency dynamic process mechanism model of the target power system, the inputs of the black-box model and the white-box model are both the PMU sequence, environmental information, and system load, and the outputs are both inertia and damping;

[0028] A prediction module, configured to respectively input the real-time PMU sequence, environmental information, and system load of the target power system into the black-box model and the white-box model to obtain the first inertia and the first damping output by the black-box model, and the second inertia and the second damping output by the white-box model;

[0029] A determination module, configured to determine the inertia of the target power system based on the first inertia and the second inertia, and determine the damping of the target power system based on the first damping and the second damping.

[0030] In a third aspect, an embodiment of the present invention provides a terminal, including a memory, a processor, and a computer program stored in the memory and executable on the processor. When the processor executes the computer program, the steps of the method described in the first aspect above or any possible implementation manner of the first aspect are implemented.

[0031] In a fourth aspect, an embodiment of the present invention provides a computer-readable storage medium storing a computer program, and when the computer program is executed by a processor, the steps of the method described in the first aspect above or any possible implementation manner of the first aspect are implemented.

[0032] The beneficial effects of the method and device for identifying the inertia and damping of the power system provided by the embodiments of the present invention are as follows:

[0033] The real-time PMU sequence can reflect the power-frequency noise signals of each node in the target power system. The environmental information and system load will affect the operating state of the target power system. The present invention expands the data basis for calculating inertia and damping for a power system with new energy power sources, and differentiates the training samples according to whether there is active power disturbance before model training, enhancing the generalization ability of the model. At the same time, by combining the black-box model and the white-box model, the inertia and damping of the target power system are determined, obtaining a more accurate identification result, improving the real-time online situation awareness and prediction ability of the power grid, effectively improving the frequency characteristic perception ability in areas with a high proportion of new energy power sources and the receiving end areas of DC interconnected power grids, further improving the active power control ability and frequency safety of the system, ensuring the safe and stable operation of the power grid, providing a basis for power system operation decision-making in an uncertain environment, and making up for the shortcoming of insufficient perception of time-varying frequency modulation characteristics in current power grid regulation, which is crucial for accurately grasping the power and frequency regulation characteristics of each link of the system. BRIEF DESCRIPTION OF THE DRAWINGS

[0034] To more clearly illustrate the technical solutions in the embodiments of the present invention, the following will briefly introduce the drawings required for use in the embodiments or the description of the prior art. Obviously, the drawings in the following description are only some embodiments of the present invention. For those of ordinary skill in the art, without creative efforts, other drawings can be obtained based on these drawings.

[0035] Figure 1 is a flowchart of the implementation of the method for identifying the inertia and damping of the power system provided by an embodiment of the present invention;

[0036] Figure 2 is a schematic structural diagram of the device for identifying the inertia and damping of the power system provided by an embodiment of the present invention;

[0037] Figure 3is a schematic diagram of a terminal provided by an embodiment of the present invention. DETAILED DESCRIPTION

[0038] In the following description, specific details such as specific system structures, technologies, etc. are provided for the purpose of illustration rather than limitation, so as to provide a thorough understanding of the embodiments of the present invention. However, it should be clear to those skilled in the art that the present invention may be implemented in other embodiments without these specific details. In other cases, detailed descriptions of well-known systems, devices, circuits, and methods are omitted to prevent unnecessary details from obstructing the description of the present invention.

[0039] In order to make the purpose, technical solutions and advantages of the present invention more clear, specific embodiments will be described below in conjunction with the accompanying drawings.

[0040] Vigorously developing renewable green energy and enhancing wide-area power transmission capacity will be the main trend of future power system development. The characteristics of "high proportion of renewable energy and high proportion of power electronic interface" will become more and more obvious in the future power grid. Due to the increase in the proportion of new energy such as wind power and photovoltaic power and the development of DC ultra-high voltage technology, the proportion of conventional synchronous generators has gradually decreased, the system inertia has decreased, the damping characteristics have changed, and the system's active balance ability has been weakened.

[0041] At present, there are more than 3,000 Phasor Measurement Units (PMUs) in the domestic high-voltage power grid, covering all 500kV nodes, some 220kV nodes and important power plants. At the same time, when the power system is operating normally, due to load changes and switch switching, electrical quantities such as line power and bus frequency show small fluctuations similar to noise. This type of noise-like signal contains rich dynamic characteristics of the system and exists at all times during the normal operation of the system. Therefore, based on the power frequency noise signals of each node of the power system collected in real time by PMU, the modeling and identification of the power system can be carried out, so as to realize the real-time online identification of the inertia and damping constant and the spatiotemporal characteristics of the power system.

[0042] In the related work on inertia and damping identification, it can be summarized into the following two research directions: 1) The methods for estimating the inertia of the power system, including estimating the equivalent inertia constant of the system or the equivalent inertia constant of the demand side, mostly use load shedding or setting short-circuit faults to stimulate the transient characteristics of the system, and then complete the estimation with the help of the generator dynamic equation. 2) The impact of the reduction of the system's rotational inertia on the system's frequency stability after the increase in the new energy penetration rate. By analyzing the relationship between two indicators, namely the system frequency change rate and the lowest frequency point, and the change of the system inertia, the impact of the reduction of the system inertia on the frequency stability is revealed. The first research direction belongs to the white-box model of model plus data, and the second direction belongs to a sensitivity analysis method, which is a mapping relationship between single factors. It lacks comprehensive consideration of the complex relationships at the system level. On the other hand, the analysis of the overall frequency response characteristics of complex large power grid systems is mainly divided into four categories: numerical simulation method, equivalent model method, linearization analysis method, and artificial intelligence. Among them, the numerical simulation method is based on the single-machine models of power sources and loads, and combines numerical integration methods to solve the accurate time-domain solution of complex models. However, in this method, the frequency response parameters are often given in advance as boundary conditions, making it difficult to directly identify. The equivalent model method establishes a low-order frequency change process description equation for the whole system to calculate the system frequency change, but ignores the problem of identifying the frequency response parameters of each part within the system. The current system inertia identification basically continues the thinking paradigm of the "white-box" model, calculating the unified inertia constant of the system according to strict theoretical formulas, or using sensitivity analysis methods to analyze the impact of a certain factor on the local operating point of inertia, lacking the unified characteristic analysis of various types of variables in the system.

[0043] See Figure 1 , which shows the implementation flowchart of the method for identifying the inertia and damping of the power system provided by the embodiment of the present invention, and is described in detail as follows:

[0044] Step 101, obtain the black-box model and white-box model of the target power system; wherein, the black-box model is constructed based on the LSTM neural network, and the white-box model is constructed based on the frequency dynamic process mechanism model of the target power system. The inputs of the black-box model and the white-box model are both PMU sequences, environmental information, and system loads, and the outputs are both inertia and damping.

[0045] In this embodiment, the parameters of each part in the power system have temporal continuity, that is, the current inertia and damping of the power system are affected not only by the current parameters but also by the parameters in the past period of time. Based on this feature, the black-box model in this embodiment is constructed based on the Long Short Term Memory (LSTM) neural network to utilize the parameters of the power system within a certain time period. The white-box model is constructed based on the frequency dynamic process mechanism model of the target power system. The frequency dynamic process mechanism model can represent the key characteristics of the frequency dynamic process of the power system, such as the lowest frequency point, the frequency change rate, the recovery time, and the frequency steady-state value, providing a tool for the analysis of the frequency dynamic process of the power system, thus realizing the "bottom-up" research paradigm from equipment to system.

[0046] In addition, the inertia and damping of the power system are jointly affected by controllable power sources, new energy sources, and loads. The new energy sources and loads are greatly affected by external factors such as meteorological conditions and energy consumption patterns. To improve the accuracy of system inertia and damping identification, this embodiment adds environmental information and system load to the PMU sequence as input parameters of the identification model.

[0047] Step 102: Input the real-time PMU sequence, environmental information, and system load of the target power system into the black-box model and the white-box model respectively to obtain the first inertia and the first damping output by the black-box model, and the second inertia and the second damping output by the white-box model.

[0048] In this embodiment, the white-box model is a model with a completely clear internal calculation process, also known as a mechanism model, while the black-box model is a model with an unclear internal calculation process, only having a mapping of input and output parameters, also known as an empirical model. Currently, the electromechanical transient process model of synchronous units is relatively mature. However, due to the diverse control strategies of new energy, it is difficult to form a mechanism model. The load model has long relied on methods with weak representation capabilities such as regression for construction. If only the white-box model is used to identify the inertia and damping of the target power system, the part related to new energy is difficult to be accurately calculated by the white-box system. If only the black-box model is used for identification, it is overly dependent on empirical data. When the target power system is too complex, it is difficult to ensure that the empirical data covers all possible situations. This embodiment aims to integrate the advantages of the two types of models.

[0049] Step 103: Determine the inertia of the target power system based on the first inertia and the second inertia, and determine the damping of the target power system based on the first damping and the second damping.

[0050] In this embodiment, both the white-box model and the black-box model are relatively accurate models. The first inertia and the first damping output by the black-box model, as well as the second inertia and the second damping output by the white-box model, are also relatively close to the actual inertia and damping of the target power system. In view of the characteristics of the white-box model and the black-box model, the present invention synthesizes the identification results obtained by the two types of models to determine the inertia and damping of the target power system, which can be closer to the actual inertia and damping of the target power system. Specifically, the first inertia and the second inertia can be used as the value range of the inertia of the target power system, and the first damping and the second damping can be used as the value range of the damping of the target power system. According to the actual situation, the endpoint value, the midpoint value or the value at a specified position can be selected as the inertia and damping of the target power system.

[0051] In a possible implementation manner, determining the inertia of the target power system based on the first inertia and the second inertia, and determining the damping of the target power system based on the first damping and the second damping includes:

[0052] If the difference between the first inertia and the second inertia is less than a preset inertia threshold, and the difference between the first damping and the second damping is less than a preset damping threshold, then the first inertia is used as the inertia of the target power system, and the first damping is used as the damping of the target power system.

[0053] In this embodiment, if the difference between the first inertia and the second inertia is less than a preset inertia threshold, and the difference between the first damping and the second damping is less than a preset damping threshold, it means that the identification result of the white-box model is very close to the identification result of the black-box model, indicating that both sets of identification results are relatively accurate. At this time, the empirical data used by the black-box model can achieve a better identification effect. Therefore, the identification result of the black-box model can be selected from the two sets of identification results as the inertia and damping of the target power system.

[0054] In a possible implementation manner, determining the inertia of the target power system based on the first inertia and the second inertia, and determining the damping of the target power system based on the first damping and the second damping further includes:

[0055] If the difference between the first inertia and the second inertia is greater than a preset inertia threshold, or the difference between the first damping and the second damping is greater than a preset damping threshold, then a grey-box model of the target power system is obtained; wherein, the grey-box model is constructed based on the LSTM neural network and the frequency dynamic process mechanism model of the target power system, the input of the grey-box model is the PMU sequence, the environmental information and the system load, and the output is the inertia and the damping;

[0056] The real-time PMU sequence, environmental information and system load of the target power system are input into the grey-box model, and the inertia and damping output by the grey-box model are used as the inertia and damping of the target power system.

[0057] In this embodiment, if the difference between the first inertia and the second inertia is greater than a preset inertia threshold, or the difference between the first damping and the second damping is greater than a preset damping threshold, it indicates that the identification results of the white-box model and the black-box model differ significantly, and it is difficult to determine which of the two sets of identification results is more accurate. At this time, a gray-box model can be introduced, adding a new technology idea driven by data on the basis of mechanism model analysis to achieve a hybrid analysis of the "white-box" and "black-box" models for inertia and damping identification.

[0058] In a possible implementation, the gray-box model includes a first sub-model and a second sub-model. The first sub-model is constructed based on the frequency dynamic process mechanism model of the target power system. The input of the first sub-model is the PMU sequence, environmental information, and system load, and the output is the state matrix of the target power system. The second sub-model is constructed based on the LSTM neural network. The input of the second sub-model is the state matrix of the target power system, and the output is inertia and damping.

[0059] In this embodiment, the gray-box model is a model that combines a mechanism model and an empirical model. The state matrix is used to represent the state information of each node of the target power system at a certain moment, or to represent the state information of a certain node of the target power system over a period of time. The state information can be electrical quantity characteristics of the power system such as voltage, current, and switch status. The detailed operating state of the target power system can be determined through the state matrix. The operating state of the power system can be solved by the mechanism model, but the mapping relationship between the operating state and the system inertia and damping is more complex and needs to be solved by the empirical model for the target power system. Therefore, in this embodiment, the state matrix of the target power system is accurately calculated by using parameter identification based on the "white-box" model, and the mapping relationship between the electrical quantity characteristics of the system and the system inertia and damping is learned by using the LSTM neural network in the characteristic analysis part. Compared with a single mechanism model or empirical model, better inertia and damping identification effects can be obtained.

[0060] In a possible implementation, obtaining the black-box model of the target power system includes:

[0061] Obtaining a full sample set; where the full sample set includes multiple samples, each sample being the PMU sequence, environmental information, and system load of the target power system at different historical moments, and the label of each sample being the inertia and damping corresponding to the sample;

[0062] Selecting the samples corresponding to the typical time points in the full sample set to construct a typical sample set, and constructing a normal sample set based on the remaining samples; where the typical time point refers to the time point when the target power system has an active power disturbance;

[0063] Randomly allocating the typical sample set and the normal sample set according to a preset ratio to obtain a training sample set and a validation sample set;

[0064] Train the LSTM neural network based on the training sample set, and verify the trained LSTM neural network based on the validation sample set to obtain the black-box model of the target power system.

[0065] In this embodiment, the inertia and damping characteristics of the system are mainly reflected in the process of large-range dynamic changes in frequency. In the classical power system stability theory, the transition time between different steady-state frequencies, the lowest frequency point, etc. depend on the system inertia, damping, and the frequency modulation characteristics of the power source and load. When the power grid is operating in a normal steady state, the active power is in a small disturbance or even no disturbance state, making it difficult to identify the system inertia. Existing system inertia calculation methods must rely on specially designed large active power disturbance experiments. As the system uncertainty continues to increase and the diversity of operating modes increases sharply, the change of system inertia is difficult to predict. Under real conditions, it is impossible to conduct disturbance tests under various working conditions to analyze the inertia constant. How to propose an inertia and damping identification method with strong generalization ability under small sample conditions in a power system with variable operating modes and increasing uncertain factors is the current difficulty.

[0066] In response to this, this embodiment differentiates the samples in the full sample set, selects the samples corresponding to the typical active power frequency event points as the typical sample set, and uses the remaining samples as the ordinary sample set. Training and validating the black-box model using the two sample sets simultaneously can improve the generalization ability of the model.

[0067] In one possible implementation, obtaining the white-box model of the target power system includes:

[0068] Obtain the full sample set; where the full sample set includes multiple samples, each sample is the PMU sequence, environmental information, and system load of the target power system at different historical moments, and the label of each sample is the inertia and damping corresponding to the sample;

[0069] Select the samples corresponding to the typical time points in the full sample set to construct the typical sample set, and construct the ordinary sample set based on the remaining samples; where the typical time point refers to the time point when the target power system experiences active power disturbance;

[0070] Randomly allocate the typical sample set and the ordinary sample set according to a preset ratio to obtain the training sample set and the validation sample set;

[0071] Train the frequency dynamic process mechanism model of the target power system based on the training sample set and the parameter estimation algorithm, and verify the trained frequency dynamic process mechanism model based on the validation sample set to obtain the white-box model of the target power system.

[0072] In this embodiment, a parameter identification method can also be used to solve the system inertia and damping in typical scenarios, and then the sample enhancement technology is used to expand the sample size to achieve the identification of inertia and damping in a small-sample and complex unknown system.

[0073] In the embodiment of the present invention, the real-time PMU sequence can reflect the power-frequency type noise signals of each node in the target power system. The environmental information and system load will affect the operating state of the target power system. The present invention expands the data basis for calculating inertia and damping for a power system with new energy power sources, and differentiates the training samples according to whether there is active disturbance before model training, enhancing the generalization ability of the model. At the same time, combining the black-box model and the white-box model to determine the inertia and damping of the target power system, obtaining a more accurate identification result, improving the real-time online situation awareness and prediction ability of the power grid, effectively improving the frequency characteristic perception ability in areas with a high proportion of new energy power sources and the receiving end areas of DC interconnected power grids, further improving the active power control ability and frequency safety of the system, ensuring the safe and stable operation of the power grid, providing a basis for power system operation decision-making in an uncertain environment, making up for the shortcoming of insufficient perception of time-varying frequency modulation characteristics in current power grid regulation, and being crucial for accurately grasping the power and frequency regulation characteristics of each link of the system.

[0074] It should be understood that the magnitudes of the sequence numbers of the steps in the above embodiments do not mean the order of execution. The execution order of each process should be determined according to its function and internal logic, and should not constitute any limitation to the implementation process of the embodiments of the present invention.

[0075] The following is the device embodiment of the present invention. For the details not described in detail therein, reference can be made to the corresponding method embodiments above.

[0076] Figure 2 The structural schematic diagram of the device for identifying the inertia and damping of a power system provided by the embodiment of the present invention is shown. For the sake of convenience of description, only the parts related to the embodiment of the present invention are shown and are described in detail as follows:

[0077] As Figure 2 shown, the device 2 for identifying the inertia and damping of a power system includes:

[0078] An acquisition module 21, configured to acquire the black-box model and the white-box model of the target power system; wherein, the black-box model is constructed based on the LSTM neural network, the white-box model is constructed based on the frequency dynamic process mechanism model of the target power system, the inputs of the black-box model and the white-box model are both the PMU sequence, environmental information and system load, and the outputs are both inertia and damping;

[0079] A prediction module 22, configured to input the real-time PMU sequence, environmental information, and system load of the target power system into a black-box model and a white-box model respectively, to obtain a first inertia and a first damping output by the black-box model, and a second inertia and a second damping output by the white-box model;

[0080] A determination module 23, configured to determine the inertia of the target power system based on the first inertia and the second inertia, and determine the damping of the target power system based on the first damping and the second damping.

[0081] In a possible implementation manner, the determination module 23 is specifically configured to:

[0082] When the difference between the first inertia and the second inertia is less than a preset inertia threshold, and the difference between the first damping and the second damping is less than a preset damping threshold, use the first inertia as the inertia of the target power system and the first damping as the damping of the target power system.

[0083] In a possible implementation manner, the determination module 23 is further configured to:

[0084] When the difference between the first inertia and the second inertia is greater than the preset inertia threshold, or the difference between the first damping and the second damping is greater than the preset damping threshold, obtain a grey-box model of the target power system; wherein, the grey-box model is constructed based on an LSTM neural network and a frequency dynamic process mechanism model of the target power system, the input of the grey-box model is the PMU sequence, environmental information, and system load, and the output is inertia and damping;

[0085] Input the real-time PMU sequence, environmental information, and system load of the target power system into the grey-box model, and use the inertia and damping output by the grey-box model as the inertia and damping of the target power system.

[0086] In a possible implementation manner, the grey-box model includes a first sub-model and a second sub-model. The first sub-model is constructed based on the frequency dynamic process mechanism model of the target power system. The input of the first sub-model is the PMU sequence, environmental information, and system load, and the output is the state matrix of the target power system. The second sub-

[0087] model is constructed based on an LSTM neural network. The input of the second sub-model is the state matrix of the target power system, and the output is inertia and damping.

[0088] In a possible implementation manner, the obtaining module 21 is specifically configured to:

[0089] Obtain a full sample set; wherein, the full sample set includes multiple samples, each sample is the PMU sequence, environmental information, and system load of the target power system at different historical moments, and the label of each sample is the inertia and damping corresponding to the sample;

[0090] 5 Select samples corresponding to typical time points from the full sample set, construct a typical sample set, and construct a normal sample set based on the remaining samples; wherein, the typical time point refers to the time point when active power disturbances occur in the target power system;

[0091] Randomly allocate the typical sample set and the normal sample set according to a preset ratio to obtain a training sample set and a validation sample set;

[0092] 0 Train the LSTM neural network based on the training sample set, and verify the trained LSTM neural network based on the validation sample set to obtain a black-box model of the target power system.

[0093] In a possible implementation, the obtaining module 21 is specifically configured to:

[0094] Obtain the full sample set; wherein, the full sample set includes multiple samples, each sample being the PMU sequence, environmental information, and system load of the target power system at different historical moments, and the label of each sample being the inertia and damping corresponding to the sample 5;

[0095] Select samples corresponding to typical time points from the full sample set, construct a typical sample set, and construct a normal sample set based on the remaining samples; wherein, the typical time point refers to the time point when active power disturbances occur in the target power system;

[0096] Randomly allocate the typical sample set and the normal sample set according to a preset ratio to obtain a training sample set and a validation sample set;

[0097] Train the frequency dynamic process mechanism model of the target power system based on the training sample set and a parameter estimation algorithm, and verify the trained frequency dynamic process mechanism model based on the validation sample set to obtain a white-box model of the target power system.

[0098] In the embodiments of the present invention, the real-time PMU sequence can reflect the power frequency noise signals of each node in the target power system. The environmental information and system load will affect the operating state of the target power system. The present invention expands the data basis for calculating inertia and damping for a power system with new energy power sources, and differentiates training samples according to the presence or absence of active power disturbances before model training, enhancing the generalization ability of the model. At the same time, by combining the black-box model and the white-box model, the inertia and damping of the target power system are determined to obtain a more accurate identification result, improving the real-time online situation awareness and prediction ability of the power grid, effectively improving the frequency characteristic perception ability in areas with a high proportion of new energy power sources and the receiving end areas of DC interconnected power grids, further improving the active power control ability and frequency safety of the system, ensuring the safe and stable operation of the power grid, providing a basis for power system operation decision-making in an uncertain environment, and making up for the shortcoming of insufficient perception of time-varying frequency modulation characteristics in current power grid regulation, which is crucial for accurately grasping the power and frequency regulation characteristics of each link in the system.

[0099] Figure 3 It is a schematic diagram of the terminal provided by the embodiments of the present invention. As Figure 3 shown, the terminal 3 of this embodiment includes: a processor 30, a memory 31, and a computer program 32 stored in the memory 31 and executable on the processor 30. When the processor 30 executes the computer program 32, it implements the steps in the embodiments of the above-mentioned various methods for identifying the inertia and damping of the power system, such as Figure 1 the steps 101 to 103 shown. Alternatively, when the processor 30 executes the computer program 32, it implements the functions of each module / unit in the above-mentioned device embodiments, such as Figure 2 the functions of the modules / units 21 to 23 shown.

[0100] Exemplarily, the computer program 32 can be divided into one or more modules / units. The one or more modules / units are stored in the memory 31 and executed by the processor 30 to complete the present invention. The one or more modules / units can be a series of computer program instruction segments capable of performing specific functions, and these instruction segments are used to describe the execution process of the computer program 32 in the terminal 3. For example, the computer program 32 can be divided into Figure 2 the modules / units 21 to 23 shown.

[0101] The terminal 3 can be a computing device such as a desktop computer, a notebook, a palm computer, and a cloud server. The terminal 3 may include, but is not limited to, a processor 30 and a memory 31. Those skilled in the art can understand, Figure 3This is only an example of the terminal 3, which does not constitute a limitation on the terminal 3. It may include more or fewer components than shown in the figure, or combine certain components, or different components. For example, the terminal may also include input / output devices, network access devices, buses, etc.

[0102] The so-called processor 30 may be a central processing unit (CPU), or may also be other general-purpose processors, digital signal processors (DSPs), application specific integrated circuits (ASICs), field-programmable gate arrays (FPGAs), or other programmable logic devices, discrete gate or transistor logic devices, discrete hardware components, etc. The general-purpose processor may be a microprocessor, or the processor may also be any conventional processor, etc.

[0103] The memory 31 may be an internal storage unit of the terminal 3, such as the hard disk or memory of the terminal 3. The memory 31 may also be an external storage device of the terminal 3, such as a plug-in hard disk, a smart media card (SMC), a secure digital (SD) card, a flash card, etc. equipped on the terminal 3. Further, the memory 31 may also include both the internal storage unit and the external storage device of the terminal 3. The memory 31 is used to store the computer program and other programs and data required by the terminal. The memory 31 may also be used to temporarily store the data that has been output or will be output.

[0104] Those skilled in the art can clearly understand that, for the convenience and brevity of description, only the above division of each functional unit and module is used for illustration. In actual applications, the above functions can be allocated to different functional units and modules according to needs, that is, the internal structure of the device can be divided into different functional units or modules to complete all or part of the functions described above. Each functional unit and module in the embodiment may be integrated into a processing unit, or each unit may exist physically alone, or two or more units may be integrated into one unit. The above integrated unit may be implemented in the form of hardware or in the form of a software functional unit. In addition, the specific names of each functional unit and module are only for the convenience of mutual distinction and do not limit the protection scope of this application. The specific working process of the units and modules in the above system can refer to the corresponding process in the foregoing method embodiment and will not be elaborated here.

[0105] In the above embodiments, the descriptions of the various embodiments have their own emphases. For parts not described in detail or recorded in a certain embodiment, reference may be made to the relevant descriptions of other embodiments.

[0106] Those of ordinary skill in the art can realize that the units and algorithm steps of each example described in combination with the embodiments disclosed herein can be implemented by electronic hardware, or a combination of computer software and electronic hardware. Whether these functions are executed in a hardware or software manner depends on the specific application and design constraints of the technical solution. Professional technicians can use different methods to implement the described functions for each specific application, but such implementation should not be considered to exceed the scope of the present invention.

[0107] In the embodiments provided by the present invention, it should be understood that the disclosed device / terminal and method can be implemented in other ways. For example, the device / terminal embodiments described above are merely illustrative. For example, the division of the modules or units is only a logical function division, and there may be other division methods in actual implementation. For example, multiple units or components can be combined or integrated into another system, or some features can be ignored or not executed. Another point is that the displayed or discussed couplings or direct couplings or communication connections to each other can be through some interfaces, and the indirect couplings or communication connections of the devices or units can be in electrical, mechanical or other forms.

[0108] The units described as separate components may or may not be physically separated, and the components displayed as units may or may not be physical units, that is, they may be located in one place, or may be distributed to multiple network units. Some or all of the units can be selected according to actual needs to achieve the purpose of the solution of this embodiment.

[0109] In addition, the functional units in the various embodiments of the present invention can be integrated in a processing unit, or each unit can exist physically alone, or two or more units can be integrated in one unit. The above integrated units can be implemented in the form of hardware or in the form of software functional units.

[0110] When the integrated module / unit is implemented in the form of a software functional unit and sold or used as an independent product, it can be stored in a computer-readable storage medium. Based on this understanding, to implement all or part of the processes in the above-described embodiment methods of the present invention, it can also be completed by a computer program instructing relevant hardware. The computer program can be stored in a computer-readable storage medium. When the computer program is executed by a processor, the steps of the above-described embodiments of various methods for identifying the inertia and damping of a power system can be implemented. Among them, the computer program includes computer program code, and the computer program code can be in the form of source code, object code, executable file, or some intermediate form, etc. The computer-readable medium can include: any entity or device capable of carrying the computer program code, recording medium, USB flash drive, mobile hard disk, magnetic disk, optical disc, computer memory, read-only memory (ROM), random access memory (RAM), electrical carrier signal, telecommunication signal, and software distribution medium, etc. It should be noted that the content included in the computer-readable medium can be appropriately increased or decreased according to the requirements of legislation and patent practice in the jurisdiction. For example, in some jurisdictions, according to legislation and patent practice, the computer-readable medium does not include electrical carrier signals and telecommunication signals.

[0111] The above-described embodiments are only used to illustrate the technical solutions of the present invention and are not intended to limit them. Although the present invention has been described in detail with reference to the foregoing embodiments, those of ordinary skill in the art should understand that they can still modify the technical solutions described in the foregoing embodiments, or perform equivalent replacements for some of the technical features. These modifications or replacements do not cause the essence of the corresponding technical solutions to deviate from the spirit and scope of the technical solutions of the various embodiments of the present invention and should all be included in the protection scope of the present invention.

Claims

1. A method for identifying the inertia and damping of a power system, characterized in that, it includes: Obtain the black-box model and white-box model of the target power system; wherein, the black-box model is constructed based on the LSTM neural network, the white-box model is constructed based on the frequency dynamic process mechanism model of the target power system, and the inputs of the black-box model and the white-box model are both PMU sequences, environmental information and system load, and the outputs are both inertia and damping; Input the real-time PMU sequence, environmental information and system load of the target power system into the black-box model and the white-box model respectively to obtain the first inertia and the first damping output by the black-box model, and the second inertia and the second damping output by the white-box model; Determine the inertia of the target power system based on the first inertia and the second inertia, and determine the damping of the target power system based on the first damping and the second damping; The determining the inertia of the target power system based on the first inertia and the second inertia, and determining the damping of the target power system based on the first damping and the second damping includes: If the difference between the first inertia and the second inertia is less than the preset inertia threshold, and the difference between the first damping and the second damping is less than the preset damping threshold, then use the first inertia as the inertia of the target power system and the first damping as the damping of the target power system; The determining the inertia of the target power system based on the first inertia and the second inertia, and determining the damping of the target power system based on the first damping and the second damping further includes: If the difference between the first inertia and the second inertia is greater than the preset inertia threshold, or the difference between the first damping and the second damping is greater than the preset damping threshold, then obtain the gray-box model of the target power system; wherein, the gray-box model is constructed based on the LSTM neural network and the frequency dynamic process mechanism model of the target power system, and the input of the gray-box model is PMU sequence, environmental information and system load, and the output is inertia and damping; Input the real-time PMU sequence, environmental information and system load of the target power system into the gray-box model, and use the inertia and damping output by the gray-box model as the inertia and damping of the target power system.

2. The method for identifying the inertia and damping of a power system according to claim 1, characterized in that, The gray-box model includes a first sub-model and a second sub-model. The first sub-model is constructed based on the frequency dynamic process mechanism model of the target power system. The input of the first sub-model is PMU sequence, environmental information and system load, and the output is the state matrix of the target power system. The second sub-model is constructed based on the LSTM neural network. The input of the second sub-model is the state matrix of the target power system, and the output is inertia and damping.

3. The method for identifying the inertia and damping of a power system according to claim 1, characterized in that, Obtaining the black-box model of the target power system includes: Obtain the full sample set; wherein, the full sample set includes multiple samples, each sample being the PMU sequence, environmental information, and system load of the target power system at different historical moments, and the label of each sample being the inertia and damping corresponding to the sample; Select the samples corresponding to the typical time points in the full sample set, construct a typical sample set, and construct a normal sample set based on the remaining samples; wherein, the typical time point refers to the time point when the active power disturbance occurs in the target power system; Randomly allocate the typical sample set and the normal sample set according to a preset ratio to obtain a training sample set and a validation sample set; Train the LSTM neural network based on the training sample set, and verify the trained LSTM neural network based on the validation sample set to obtain the black-box model of the target power system.

4. The method for identifying the inertia and damping of a power system according to claim 1, characterized in that, Obtaining the white-box model of the target power system includes: Obtain the full sample set; wherein, the full sample set includes multiple samples, each sample being the PMU sequence, environmental information, and system load of the target power system at different historical moments, and the label of each sample being the inertia and damping corresponding to the sample; Select the samples corresponding to the typical time points in the full sample set, construct a typical sample set, and construct a normal sample set based on the remaining samples; wherein, the typical time point refers to the time point when the active power disturbance occurs in the target power system; Randomly allocate the typical sample set and the normal sample set according to a preset ratio to obtain a training sample set and a validation sample set; Train the frequency dynamic process mechanism model of the target power system based on the training sample set and the parameter estimation algorithm, and verify the trained frequency dynamic process mechanism model based on the validation sample set to obtain the white-box model of the target power system.

5. An apparatus for identifying the inertia and damping of a power system, characterized in that, comprising: An acquisition module, configured to acquire the black-box model and the white-box model of the target power system; wherein, the black-box model is constructed based on an LSTM neural network, the white-box model is constructed based on the frequency dynamic process mechanism model of the target power system, and the inputs of the black-box model and the white-box model are both the PMU sequence, environmental information, and system load, and the outputs are both inertia and damping; A prediction module, configured to input the real-time PMU sequence, environmental information, and system load of the target power system into the black-box model and the white-box model respectively to obtain the first inertia and the first damping output by the black-box model, and the second inertia and the second damping output by the white-box model; A determination module, configured to determine the inertia of the target power system based on the first inertia and the second inertia, and determine the damping of the target power system based on the first damping and the second damping; The determination module is specifically configured to: When the difference between the first inertia and the second inertia is less than a preset inertia threshold, and the difference between the first damping and the second damping is less than a preset damping threshold, the first inertia is taken as the inertia of the target power system, and the first damping is taken as the damping of the target power system; The determining module is further configured to: If the difference between the first inertia and the second inertia is greater than a preset inertia threshold, or the difference between the first damping and the second damping is greater than a preset damping threshold, obtain a gray-box model of the target power system; wherein, the gray-box model is constructed based on an LSTM neural network and a frequency dynamic process mechanism model of the target power system, the input of the gray-box model is a PMU sequence, environmental information and system load, and the output is inertia and damping; Input the real-time PMU sequence, environmental information and system load of the target power system into the gray-box model, and take the inertia and damping output by the gray-box model as the inertia and damping of the target power system.

6. A terminal, comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, Characterized in that, When the processor executes the computer program, the steps of the method according to any one of claims 1 to 4 above are implemented.

7. A computer-readable storage medium, the computer-readable storage medium stores a computer program, Characterized in that, When the computer program is executed by a processor, the steps of the method according to any one of claims 1 to 4 above are implemented.

Images