Electric energy storage power station virtual inertia optimization method considering inertia space-time distribution characteristics

Abstract

The invention provides an electric energy storage power station virtual inertia optimization method considering inertia space-time distribution characteristics, and the method comprises the steps: carrying out the virtual inertia optimization of an electric energy storage power station under the scene of considering the non-uniform inertia distribution of an electric power system under the high-proportion new energy access; constructing a source measurement frequency response characteristic model considering load power mapping and a load node frequency response characteristic model considering source measurement frequency mapping; the system small interference stability based on the characteristic root method is used as a frequency safety constraint condition; taking the minimum inertia and frequency modulation energy of the energy storage power station as a target, considering the physical constraint, the frequency security constraint and the system stability constraint of the energy storage power station, and proposing a network tracking / constructing energy storage power station inertia coordination optimization strategy considering the frequency dynamic space-time distribution; wherein the objective function is that the sum of the integral quantity of the load node frequency deviation and the inertia kinetic energy of the following / constructing network energy storage power station is minimum. According to the method, aiming at the characteristic of frequency space distribution differentiation in a novel power system, it is ensured that all nodes can meet double security constraints of frequency deviation and frequency change rate.

Description

Technical Field

[0001] This invention relates to the field of new energy utilization technology, and in particular to a virtual inertia optimization method for power storage power stations that considers the spatiotemporal distribution characteristics of inertia. Background Technology

[0002] Inertia is an inherent property of power systems, manifested as the system's resistance to energy fluctuations caused by external disturbances, and is a fundamental guarantee for the safe and stable operation of the system. The large-scale grid connection of new energy sources and the rapid decline in the proportion of synchronous generators have led to profound changes in the spatiotemporal distribution characteristics and support forms of power system inertia.

[0003] With the increasing penetration of various new energy sources and energy storage power generation units with different inertia characteristics into the power system, the new power frequency characteristic model of the power system based solely on the traditional single center of inertia (COI) can no longer accurately characterize the dynamic spatiotemporal distribution of frequency at different nodes, making it difficult to accurately assess the dynamic frequency response of each node. Therefore, the improvement effect of the multi-power station inertia coordination optimization strategy proposed based on this principle on the system's dynamic frequency response may not truly meet the frequency security requirements. Extensive research has been conducted on inertia optimization control of power systems, primarily focusing on two aspects. The first is inertia optimization strategies based on multi-center-of-inertia models. This simplifies the power system to a system frequency response (SFR) model, and the calculated equivalent inertia represents the system's central inertia, reflecting the overall system's frequency immunity. However, existing analytical methods for the frequency response of inertia center models have limitations; dimensionality reduction modeling errors caused by simplification of power source dynamics and network structures degrade the control accuracy of inertia optimization strategies. The second aspect is inertia optimization strategies considering the spatiotemporal characteristics of system frequencies. This involves considering the spatiotemporal distribution of frequencies and quantitatively analyzing the frequency characteristics of nodes. However, most current inertia optimization strategies focus on grid-connected power sources. The methods for realizing the inertia of power electronic power sources connected to the grid vary, and optimizing only the virtual inertia parameters for grid-connected control cannot meet the needs of new power systems. Summary of the Invention

[0004] To address the aforementioned issues, this invention aims to propose a virtual inertia optimization method for power storage stations that considers the spatiotemporal distribution characteristics of inertia. By optimizing the operating inertia of the power storage station, taking into account the frequency response characteristics of different nodes and the differences in inertia resources of grid-connected power storage stations, this method ensures the frequency security of the new power system under uneven inertia distribution, reduces the fluctuations in frequency and RoCoF of each node in the system after disturbances, and improves the small-disturbance stability of the new power system.

[0005] To achieve the above objectives, the technical solution of the present invention is implemented as follows: A method for optimizing the virtual inertia of a power storage substation considering the spatiotemporal distribution characteristics of inertia includes the following steps: S1. In the scenario of uneven power system inertia distribution under the consideration of high proportion of new energy access, construct source-measured frequency response characteristic model and load node frequency response characteristic model considering load power mapping. S2. The system small-disturbance stability based on the characteristic root method is used as a frequency security constraint. S3. Taking the minimum inertia and frequency regulation energy of the energy storage power station as the objective, and considering the physical constraints, frequency safety constraints and system stability constraints of the energy storage power station, a coordinated optimization strategy for the inertia of the grid-connected / networked energy storage power station is proposed, taking into account the dynamic spatiotemporal distribution of frequency. The objective function is to minimize the sum of the integral of the load node frequency deviation and the inertial kinetic energy of the grid-connected / networked energy storage power station.

[0006] Furthermore, the construction of the source-measurement frequency response characteristic model considering load power mapping is as follows: For a power network with m generating nodes and n load nodes, neglecting voltage amplitude fluctuations at each node and line resistance, the active power in the network is described using a DC model, and the power flow equations for the active power injected into each node are as follows: (1) In the formula, ΔPe=[ΔPe.1,ΔPe.2,…,ΔPe.m]T, Δδ=[Δδ1,Δδ2,…,Δδm]T are the column vectors of the power injection change and the phase angle change of each power generation node, respectively; ΔPL=[ΔPL1,ΔPL2,…,ΔPLn]T, Δθ=[Δθ1,Δθ2,…,Δθn]T are the column vectors of the power injection change and the phase angle change of each load node, respectively; BGG is the diagonal matrix composed of the reactance of the power generation nodes, BLL is the correlation matrix between the power change and the angle change of the load nodes, and BLG and BGL are the correlation matrices between the power generation nodes and the load nodes; The power frequency characteristic models for synchronous machines and grid-connected energy storage generation units are as follows: (2) (3) Where Hm, Dm, HGFM, and DGFM represent the diagonal matrices composed of the inertial time constant and damping coefficient of each synchronous generator and grid-type energy storage power station, respectively, i.e., Hm=diag(Hm1,Hm2,…), Dm=diag(Dm1,Dm2,…), HGFM=diag(HGFM.1, HGFM.2,…), DGFM=diag(DGFM.1, DGFM.2,…); Source-measured frequency response characteristic model considering load power mapping: (4) HGFL is a diagonal matrix composed of the virtual inertia time constants of each energy storage power station using grid-based virtual inertia control, i.e., HGFL=diag(HGFL.1,HGFL.2,…), and DGFL is a diagonal matrix composed of the damping coefficients of each energy storage power station using grid-based virtual inertia control, i.e., DGFL=diag(DGFL1,DGFL2,…). The construction of the load node frequency response characteristic model considering source-measuring frequency mapping: Under the assumptions of unit voltage amplitude and pure reactance transmission lines, and without considering the dynamic differences within the energy storage power station, the energy storage power station adopts the capacity equalization aggregation method. The network nodes are then renumbered in the order of synchronous generators, grid-connected energy storage power stations, and grid-connected energy storage power stations. Assuming that the reference power ΔPset.GFM of the grid-connected energy storage power station remains constant during the frequency response, the new power system frequency dynamic characteristic equation considering the spatiotemporal distribution characteristics of frequency can be expressed as: (5) Following the Kron simplification approach, only the dynamics at the power generation nodes are retained, while load node variables are eliminated. Power disturbances occurring at load nodes are equivalently distributed to the power generation nodes. Except for the generation nodes, the injected current at all nodes is zero. The network equation can be expressed as: (6) in, ∈Rm×1 is a column vector consisting of the current injected into the system by all power generation nodes in the network; ∈Rm×1 and Let Rn×1 be a column vector composed of the potential of the generating nodes and the voltage of the load nodes in the entire network, where the elements are respectively... =Emejδ, =Unejδ; The load node voltage can be expressed as the electromotive force within the entire grid's generating units: (7) In the formula, K represents the complex matrix relating the admittances of generating nodes and load nodes. The load node k and all generating nodes have the following relationship: (8) (9) In the formula, Kk,i is the element in the k-th row and i-th column of K; Simplifying, we obtain the frequency expression for the load nodes. It can be seen that the load node frequency is a weighted superposition of the generator node frequencies with respect to the network topology, i.e.: (10) Equations (5) and (10) together form a new frequency-space-time dynamic response model for power systems.

[0007] Furthermore, the system stability constraint refers to: determining whether the system satisfies the small disturbance stability criterion by solving the eigenvalues ​​of the system state matrix, i.e., the damping ratio of all eigenvalues ​​is greater than 3% to 5%; the system stability constraint is achieved in the following way: under the assumption of unit voltage amplitude and pure reactive transmission line, the system state matrix is ​​constructed; according to Lyapunov's first method, the eigenvalues ​​of the system state matrix are solved to determine whether the system satisfies the small disturbance stability criterion.

[0008] Furthermore, the frequency safety constraint refers to: considering the frequency of all nodes in the system and the frequency change rate RoCoF constraint, ensuring that the frequency and RoCoF of each node are within a safe range after the disturbance.

[0009] Furthermore, the frequency security constraint is implemented in the following way: Set upper and lower limits for the frequency to ensure that the frequency of each node does not exceed the limits after the disturbance; Set the maximum value of RoCoF to ensure that the rate of frequency change of each node is within a safe range after the disturbance.

[0010] Furthermore, the physical constraints of the energy storage power station refer to the physical performance constraints of the energy storage power station, including the range of the equivalent backup inertia and the maximum power output capacity.

[0011] Furthermore, the physical constraints of the energy storage power station are achieved in the following ways: Consider the range of equivalent backup inertia of energy storage power stations to avoid the inability of energy storage power stations to respond to optimized control commands; Consider the maximum power output capacity of the energy storage power station to ensure that the energy storage power station does not exceed its physical limitations during the optimization process.

[0012] Furthermore, the construction of the source-measured frequency response characteristic model and the load node frequency response characteristic model considering load power mapping specifically includes: The network nodes are renumbered in the order of synchronous generators, grid-connected energy storage power stations, and grid-connected energy storage power stations; Assuming that the reference power of the grid-connected energy storage power station remains constant during the frequency response, a new power system frequency dynamic characteristic equation considering the spatiotemporal distribution characteristics of frequency is constructed.

[0013] Furthermore, the novel power system frequency dynamic characteristic equation can be expressed as: (11) Where x is the system's state variable, u is the system's disturbance variable, and K L K R A R This corresponds to the system matrix; Linearizing equation (11) at the equilibrium point, the state matrix of the system is constructed as follows: (12) In equation (12), As is the system state matrix; According to Lyapunov's first method, if all eigenvalues ​​have negative real parts, then the system is stable; by solving for the eigenvalues ​​of the system's state matrix, the eigenvalues ​​of the system can be obtained. Let the i-th eigenvalue be... Then the damping ratio expression for the i-th mode is: (13) To ensure system stability under small disturbances, each eigenvalue should have positive damping. When the minimum damping ratio is greater than 3% to 5%, the system is considered to satisfy the small disturbance stability criterion. ζ represents the set of damping ratios of all eigenvalues, set as follows: (14) The specific constraints on node frequency are as follows: (15) In the formula, ω max and ω min These are the upper and lower bounds of the frequency limit, respectively; dω max / dt i The maximum value of RoCoF at node i, RoCoF max It is the limit value for the rate of change of frequency; The physical constraint expressions for energy storage power stations are as follows: (16) Among them, P GFM.min and P GFM.max These represent the upper and lower bounds of the active power output of a grid-connected energy storage power station, P. GFM.i The active power output of the i-th grid-connected energy storage power station; P GFL.min and P GFL.maxThese represent the upper and lower bounds of the active power output of a grid-connected energy storage power station, P. GFL.i The active power output is for the i-th grid-connected energy storage power station.

[0014] Furthermore, the objective function is as follows: (17) In equation (17), T is the inertial response time, a L a G These are the weighting coefficients corresponding to the load node frequency deviation and the generator node kinetic energy deviation, respectively, with the aim of ensuring that the two are of similar order of magnitude.

[0015] Beneficial effects: This invention optimizes the operating inertia of energy storage power stations, considers the frequency response characteristics of different nodes and the differences in inertia resources of grid-connected energy storage power stations, thereby ensuring the frequency security of the new power system under uneven inertia distribution, reducing the fluctuations of frequency and RoCoF of each node in the system after disturbance, and improving the small-disturbance stability of the new power system. Attached Figure Description

[0016] The accompanying drawings, which form part of this invention, are used to provide a further understanding of the invention. The illustrative embodiments of the invention and their descriptions are used to explain the invention and do not constitute an undue limitation of the invention. In the drawings: Figure 1 This is a flowchart of the virtual inertia optimization method for power storage power stations that considers the spatiotemporal distribution characteristics of inertia, as described in an embodiment of the present invention. Figure 2 This is the IEEE-68 node power system network topology diagram in this embodiment; Figure 3 This is a comparison chart of the virtual inertia of the energy storage power station under the three schemes in this embodiment; Figure 4 This is a comparison diagram of the distribution of system matrix eigenvalues ​​under the three schemes in this embodiment; Figure 5a This is the frequency response waveform of the 68 nodes in the following embodiment; Figure 5b This is the frequency response waveform of 68 nodes under Scheme 2 in this embodiment; Figure 5c This is the frequency response waveform of 68 nodes under Scheme 3 in this embodiment; Figure 6a This is the spatial distribution of the frequency change rate of the 68 nodes in the scheme below in this embodiment; Figure 6b This is the spatial distribution of the frequency change rate of the 68 nodes under Scheme 2 in this embodiment; Figure 6c This is the spatial distribution of the frequency change rate of 68 nodes under Scheme 3 in this embodiment; Figure 7 These are the virtual inertia optimization results of energy storage power stations in scenarios 1 and 2 of this embodiment; Figure 8 This embodiment compares the distribution of system matrix eigenvalues ​​for scenarios 1 and 2. Figure 9 This is the 68-node frequency response waveform of scenario 2 in this embodiment; Figure 10 These are comparison waveforms of the frequency response of node 28 under two scenarios in this embodiment; Figure 11a These are the active power output curves of the energy storage power stations at nodes 7, 8, and 9 in scenario 1 of this embodiment; Figure 11b These are the active power output curves of the energy storage power stations at nodes 7, 8, and 9 in scenario 2 of this embodiment. Detailed Implementation

[0017] It should be noted that, unless otherwise specified, the embodiments and features described in the present invention can be combined with each other.

[0018] The present invention will now be described in detail with reference to the accompanying drawings and embodiments.

[0019] Design Concept: Inertia is an inherent property of power systems, manifesting as impedance to energy fluctuations caused by external disturbances, and is a fundamental guarantee for the safe and stable operation of the system. The large-scale grid connection of new energy sources and the rapid decline in the proportion of synchronous generators have profoundly changed the spatiotemporal distribution characteristics and support forms of power system inertia. Energy storage power stations based on power electronic interfaces differ from traditional power sources in their power supply characteristics, providing flexible and controllable virtual inertia. However, this also makes the frequency stability characteristics of each node more complex and variable, exacerbating the frequency fluctuations at each node caused by the uneven horizontal spatial distribution of power system inertia, and significantly impacting system frequency analysis, control, and protection.

[0020] Example 1 Based on the above design concept, see Figure 1 This embodiment of the method for optimizing the virtual inertia of a power storage power station considering the spatiotemporal distribution characteristics of inertia includes the following steps: S1. In the scenario of uneven power system inertia distribution under the consideration of high proportion of new energy access, construct source-measured frequency response characteristic model and load node frequency response characteristic model considering load power mapping. S2. The system small-disturbance stability based on the characteristic root method is used as a frequency security constraint. S3. Taking the minimum inertia and frequency regulation energy of the energy storage power station as the objective, and considering the physical constraints, frequency safety constraints and system stability constraints of the energy storage power station, a coordinated optimization strategy for the inertia of the grid-connected / networked energy storage power station is proposed, taking into account the dynamic spatiotemporal distribution of frequency. The objective function is to minimize the sum of the integral of the load node frequency deviation and the inertial kinetic energy of the grid-connected / networked energy storage power station.

[0021] The inertia coordination optimization strategy for energy storage power stations proposed in this embodiment, which considers the dynamic spatiotemporal distribution characteristics of frequency, can ensure that all nodes in the system can meet the dual safety constraints of system frequency deviation and frequency change rate, taking into account the characteristics of frequency spatial distribution differences in new power systems.

[0022] In a specific example, the construction of the source-measured frequency response characteristic model taking into account load power mapping is as follows: For a power network containing m generating nodes (including synchronous generators and grid-connected energy storage power stations participating in inertial response) and n load nodes (including electricity loads and new energy power stations without inertial response), neglecting voltage amplitude fluctuations at each node and line resistance, the active power in the network is described using a DC model, and the power flow equations for the active power injected into each node are as follows: (1) In the formula, ΔPe=[ΔPe.1,ΔPe.2,…,ΔPe.m]T, Δδ=[Δδ1,Δδ2,…,Δδm]T are the column vectors of the power injection change and the phase angle change of each power generation node, respectively; ΔPL=[ΔPL1,ΔPL2,…,ΔPLn]T, Δθ=[Δθ1,Δθ2,…,Δθn]T are the column vectors of the power injection change and the phase angle change of each load node, respectively; BGG is the diagonal matrix composed of the reactance of the power generation nodes, BLL is the correlation matrix between the power change and the angle change of the load nodes, and BLG and BGL are the correlation matrices between the power generation nodes and the load nodes.

[0023] The power frequency characteristic models for synchronous machines and grid-connected energy storage generation units are as follows: (2) (3) Wherein, Hm, Dm, HGFM, and DGFM represent the diagonal matrix composed of the inertial time constant and damping coefficient of each synchronous generator and grid-type energy storage power station, respectively, i.e., Hm=diag(Hm1,Hm2,…), Dm=diag(Dm1,Dm2,…), HGFM=diag(HGFM.1, HGFM.2,…), and DGFM=diag(DGFM.1, DGFM.2,…).

[0024] Because the power frequency characteristic model of a grid-connected energy storage generator can directly reflect the power angle characteristics and spatial network mapping relationship (mapped in power angle and electromagnetic power) similar to those of a synchronous generator, while the power frequency characteristic model of a grid-connected energy storage generator is only affected by the power angle of other nodes in the power network, it cannot construct a power frequency model that reflects the spatiotemporal characteristics of frequency. In order to obtain a unified power frequency characteristic model that takes into account the load mapping relationship, this invention proposes a method to replace the power angle change of the grid-connected energy storage generator with the power angle relationship of other non-grid-connected generator nodes, forming a power frequency dynamic characteristic model similar to that of synchronous generators and grid-connected energy storage generators, thereby ensuring that the power frequency characteristic model of the grid-connected energy storage generator can reflect the spatial frequency dynamic characteristics.

[0025] (4) HGFL is a diagonal matrix composed of the virtual inertia time constants of each energy storage power station using grid-following virtual inertia control, i.e., HGFL=diag(HGFL.1,HGFL.2,…), and DGFL is a diagonal matrix composed of the damping coefficients of each energy storage power station using grid-following virtual inertia control, i.e., DGFL=diag(DGFL1,DGFL2,…).

[0026] The construction of the load node frequency response characteristic model considering source-measuring frequency mapping: Under the assumptions of unit voltage amplitude and pure reactance transmission lines, and without considering the dynamic differences within the energy storage power station, the energy storage power station adopts the capacity equalization aggregation method. The network nodes are then renumbered in the order of synchronous generators, grid-connected energy storage power stations, and grid-connected energy storage power stations. Assuming that the reference power ΔPset.GFM of the grid-connected energy storage power station remains constant during the frequency response, the new power system frequency dynamic characteristic equation considering the spatiotemporal distribution characteristics of frequency can be expressed as: (5) Following the Kron simplification approach, only the dynamics at the power generation nodes are retained, while load node variables are eliminated. Power disturbances occurring at load nodes are equivalently distributed to the power generation nodes. Except for the generation nodes, the injected current at all nodes is zero. The network equation can be expressed as: (6) in, ∈Rm×1 is a column vector consisting of the current injected into the system by all power generation nodes in the network; ∈Rm×1 and Let Rn×1 be a column vector composed of the potential of the generating nodes and the voltage of the load nodes in the entire network, where the elements are respectively... =Emejδ, =Unejδ The load node voltage can be expressed as the electromotive force within the entire grid's generating units: (7) In the formula, K represents the complex matrix relating the admittances of generating nodes and load nodes. The load node k and all generating nodes have the following relationship: (8) (9) In the formula, Kk,i is the element in the k-th row and i-th column of K.

[0027] Simplifying, we obtain the frequency expression for the load nodes. It can be seen that the load node frequency is a weighted superposition of the generator node frequencies with respect to the network topology, i.e.: (10) Equations (5) and (10) together form a new frequency-space-time dynamic response model for power systems.

[0028] In a specific example, the system stability constraint refers to determining whether the system satisfies the small disturbance stability criterion by solving the eigenvalues ​​of the system state matrix, i.e., the damping ratio of all eigenvalues ​​is greater than 3% to 5%. The system stability constraint is implemented in the following way: Under the assumptions of unit voltage amplitude and pure reactive transmission line, construct the state matrix of the system; According to Lyapunov's first method, the eigenvalues ​​of the system state matrix are solved to determine whether the system satisfies the small disturbance stability criterion.

[0029] This embodiment is based on the stability analysis of the eigenvalue method: by constructing the system's state matrix and using the eigenvalue method to analyze the system's small-disturbance stability, it ensures that the system operates under the condition that all eigenvalues ​​have negative real parts (i.e., positive damping), thereby enhancing the system's stability. When the minimum damping ratio is greater than 3% to 5%, the system is considered to meet the small-disturbance stability criterion.

[0030] In a specific example, the frequency safety constraint refers to: considering the frequencies of all nodes in the system and the RoCoF constraint, ensuring that the frequencies and RoCoF of each node remain within a safe range after the disturbance. This frequency safety constraint is achieved through the following methods: Set upper and lower limits for the frequency to ensure that the frequency of each node does not exceed the limits after the disturbance; Set the maximum value of RoCoF to ensure that the rate of frequency change of each node is within a safe range after the disturbance.

[0031] The physical constraints of the energy storage power station refer to the physical performance constraints of the energy storage power station, including the range of the equivalent reserve inertia and the maximum power output capability. These physical constraints are achieved through the following methods: Consider the range of equivalent backup inertia of energy storage power stations to avoid the inability of energy storage power stations to respond to optimized control commands; Consider the maximum power output capacity of the energy storage power station to ensure that the energy storage power station does not exceed its physical limitations during the optimization process.

[0032] This embodiment optimizes the inertia configuration of energy storage power stations. Virtual inertia is flexible and controllable: It takes into account the physical constraints of the energy storage power station (such as the range of equivalent backup inertia and maximum power output capability), as well as frequency security constraints and system stability constraints. By optimizing the operating inertia of the energy storage power station, it reduces the fluctuations of frequency and frequency change rate of each node in the system after disturbance.

[0033] Coordination and optimization of grid-connected / network-connected energy storage power stations: An inertia coordination and optimization strategy applicable to the hybrid access of grid-connected / network-connected energy storage power stations is proposed, which can more effectively improve the stability of the system; the access optimization of network-connected energy storage power stations shows better results in improving system stability.

[0034] In a specific example, the construction of the source-measured frequency response characteristic model and the load node frequency response characteristic model considering load power mapping specifically includes: The network nodes are renumbered in the order of synchronous generators, grid-connected energy storage power stations, and grid-connected energy storage power stations; Assuming that the reference power of the grid-connected energy storage power station remains constant during the frequency response, a new power system frequency dynamic characteristic equation considering the spatiotemporal distribution characteristics of frequency is constructed.

[0035] In a specific example, the novel power system frequency dynamic characteristic equation can be expressed as: (11) Where x is the system's state variable, u is the system's disturbance variable, and K L K R A R This corresponds to the system matrix; Linearizing equation (11) at the equilibrium point, the state matrix of the system is constructed as follows: (12) In equation (12), As is the system state matrix; According to Lyapunov's first method, if all eigenvalues ​​have negative real parts, then the system is stable; by solving for the eigenvalues ​​of the system's state matrix, the eigenvalues ​​of the system can be obtained. Let the i-th eigenvalue be... Then the damping ratio expression for the i-th mode is: (13) To ensure system stability under small disturbances, each eigenvalue should have positive damping. When the minimum damping ratio is greater than 3% to 5%, the system is considered to satisfy the small disturbance stability criterion. ζ represents the set of damping ratios of all eigenvalues, set as follows: (14) Protection devices triggered by frequency stability indicators (such as the lowest frequency point or RoCoF) are installed on specific busbars. Considering the significant differences in frequency spatial distribution in new power systems, this invention considers the frequencies of all nodes in the system and RoCoF constraints to ensure the safe and stable operation of the system. According to the basic principles of power system dynamic analysis, the maximum value of the dynamic frequency after a disturbance occurs at the instant the disturbance occurs, and is determined by system inertia and power flow distribution. The specific constraints on node frequencies are as follows: (15) In the formula, ω max and ω min These are the upper and lower bounds of the frequency limit, respectively; dω max / dt i The maximum value of RoCoF at node i, RoCoF max It is the limit value for the rate of change of frequency; The inertia support capability of energy storage power stations for the power system is limited by the physical performance of the energy storage power stations themselves. The physical constraints of energy storage mainly consider the range of the equivalent reserve inertia of the energy storage power station and its maximum power output capability, in order to avoid the energy storage power station being unable to respond to the inertia commands issued by the optimization control, which would affect the overall frequency characteristics of the system. The physical constraint expression of the energy storage power station is as follows: (16) Among them, P GFM.min and P GFM.max These represent the upper and lower bounds of the active power output of a grid-connected energy storage power station, P. GFM.i The active power output of the i-th grid-connected energy storage power station; P GFL.min and P GFL.max These represent the upper and lower bounds of the active power output of a grid-connected energy storage power station, P. GFL.i The active power output is for the i-th grid-connected energy storage power station.

[0036] In a specific example, when an energy storage power station participates in grid inertia regulation, it is desirable for the frequencies of each load node to be closer to the base value, and for the required kinetic energy of inertia to be minimized. Therefore, the objective function is to minimize the sum of the integral of the load node frequency deviation and the inertial kinetic energy of the grid-connected energy storage power station. The specific objective function is as follows: (17) In equation (1), T is the inertial response time, a L a GThese are the weighting coefficients corresponding to the load node frequency deviation and the generator node kinetic energy deviation, respectively, with the aim of ensuring that the two are of similar order of magnitude.

[0037] In practical implementation, to verify the effectiveness of the virtual inertia optimization method for power storage power stations considering the spatiotemporal distribution characteristics of inertia proposed in this embodiment, a novel 68-node power system simulation system was built in MATLAB, such as... Figure 2 As shown, the following modifications are made to the IEEE-68 node system: nodes 11-13 remain SG nodes with a corresponding inertia time constant HSG = 40s; the remaining 13 generating nodes are all grid-type energy storage power station generating nodes. The variable to be optimized in the algorithm is the virtual inertia parameter of each energy storage power station. It is assumed that a 2p.u. load surge occurs at node 28 in the system at 1.2s.

[0038] Simulation 1: Comparative Analysis of the Effects of Virtual Inertia Coordination Optimization in Energy Storage Power Stations Using ±0.5Hz and ±2Hz / s as constraints for frequency deviation and frequency change rate, three different virtual inertia optimization comparison schemes are presented: (1) Option 1: Low inertia setting of energy storage power station without optimization strategy; (2) Scheme 2: Energy storage inertia optimization based on COI model; (3) Option 3: The energy storage power station inertia optimization strategy proposed in this paper.

[0039] The optimization results of the virtual inertia of the energy storage power station corresponding to the three schemes are as follows: Figure 3 As shown.

[0040] Figure 4 The system matrix As is distributed under three strategies. In Scheme 1, the dominant eigenvalues ​​are generally closer to the imaginary axis, and there are two sets of modal eigenvalues ​​less than 5%. In this case, the system is prone to oscillation after being disturbed. Compared with Scheme 2, the dominant eigenvalues ​​of the system matrix in Scheme 3 are more biased to the left of the imaginary axis. The angle of the eigenvalue distribution in Scheme 3 is smaller than that in Scheme 2, and the damping ratio of the corresponding oscillation mode is also larger.

[0041] Figures 5(a), (b), and (c) compare the frequency response waveforms of 68 nodes under three schemes. In Scheme 1, the frequency fluctuation amplitude of each node is the largest, with the frequency drop at node 28, where the disturbance occurs, reaching 59.46Hz. The frequency fluctuation amplitude of the other nodes is also relatively large. In Scheme 2, the optimization strategy adopts the COI model, which can also ensure that the frequency of each node in the system does not exceed the limit, and the frequency fluctuation amplitude of each node is smaller compared with Scheme 1. However, the frequency of each node still fluctuates significantly after the disturbance occurs. In Scheme 3, based on the strategy proposed in this chapter, the maximum frequency deviation of the nodes in the system is the smallest among the three schemes, ensuring the frequency safety of the system. The frequency fluctuation of each node is not significantly different, and the uneven distribution of inertia and the frequency dynamic characteristics of the system are improved. The steady-state frequency of each node under all three schemes is 59.62Hz, because the steady-state frequency deviation of the system is mainly related to the disturbance power, primary frequency modulation, and secondary frequency modulation, and is not affected by the inertia distribution.

[0042] To more intuitively illustrate the optimization effect of this embodiment on a novel power system with significant differences in frequency spatial distribution, Figures 6(a), (b), and (c) show the distribution of the maximum frequency change rate within the system in the form of temperature difference diagrams. The inertia of nodes 11-12 is mainly provided by synchronous generators, and the inertia level in adjacent areas is relatively abundant, resulting in a smaller frequency change rate in these areas. Generating node 9 is electrically far from the other inertia sources, therefore the frequency change rate in the area near node 9 is larger than in other areas. In Scheme 1, the system has weak resistance to frequency changes. After an active power disturbance occurs at node 28, the frequency change rates of nodes 4, 5, 6, 9, 20, and 25-29 significantly exceed the 2Hz / s limit, and the frequency change rates of most nodes are close to the dangerous range of 2Hz / s, failing to meet frequency safety requirements. Scheme 2 can ensure that the frequency change rate in most areas meets the conditions, but it cannot guarantee that the local frequency response index meets the requirements; the frequency change rate indexes of nodes 9, 27, 28, and 29 still exceed the limit. Scheme 3 considers the characteristics of frequency spatial distribution, ensuring that the frequency change rate indexes of all nodes meet the requirements.

[0043] Simulation 2: Comparative Analysis of Virtual Inertia Optimization for Grid-Based Energy Storage Power Stations To compare and analyze the characteristics and differences of the virtual inertia provided by grid-based and grid-connected energy storage power stations, the optimization scenario corresponding to Scheme 3 in Simulation 1 is defined as Scenario 1; the generation nodes 7, 8, and 9 near the disturbance occurrence node 28 in Scheme 3 are changed from grid-based energy storage power stations to grid-connected energy storage power stations, and this scenario is defined as Scenario 2. The network configuration, power disturbance magnitude, and occurrence time remain consistent with Simulation 1. The optimal inertia optimization results for the corresponding energy storage power station are determined according to the optimization process, such as... Figure 7 As shown.

[0044] Figure 8The distribution of eigenvalues ​​of the system matrix As in scenarios 1 and 2 is shown. The spectral distribution of the eigenvalues ​​of the state matrix in scenario 1 exhibits a narrower fan-shaped region compared to scenario 2, indicating that the system has better damping characteristics under the inertial support of the grid-type energy storage power station.

[0045] Figure 9 The proposed strategy can also improve the dynamic frequency performance of system nodes under grid-connected energy storage power station access, ensuring that each node meets the frequency security requirements.

[0046] Figure 10 A comparison of the frequency response waveforms of node 28 in scenarios 1 and 2 is presented. The lowest frequency drop point in scenario 2 is 59.6Hz. Furthermore, throughout the entire frequency dynamic process at node 28, due to the delayed response of the grid-connected system, the frequency drop rate in scenario 2 is faster than in scenario 1 after the disturbance occurs. The maximum frequency difference between node 28 in scenarios 1 and 2 is 0.1Hz at 3.2s. It is clear that grid-connected energy storage power stations have a greater advantage in supporting the inertia of the power system.

[0047] Figure 11 compares the active power output curves of grid-based and grid-connected energy storage stations. In both scenarios, the energy storage station with the highest active power output is located at node 9. In scenario 1, the maximum active power change of the grid-based energy storage station is 0.089 pu, approximately 1.5 times that of the grid-connected energy storage station in scenario 2 (0.058 pu). Furthermore, the maximum output power of the grid-connected energy storage station is not at the instant of the disturbance. Under the same inertia optimization strategy, the grid-based energy storage station can respond to power changes more promptly, resulting in a larger instantaneous output power. Integrating the active power output curves from the instant of the disturbance to 4 seconds, the energy provided by the energy storage station in scenario 2 during the disturbance process is 0.87 kWh more than that in scenario 1, indicating that the grid-connected energy storage station outputs more energy during the inertia response process. The maximum response power of the grid-connected energy storage station is relatively smaller than that of the grid-based energy storage station, resulting in a slower response speed and, correspondingly, a longer power oscillation time. Therefore, the inertia support effect of grid-type energy storage power stations is more conducive to mitigating frequency oscillations after disturbances to new power systems.

[0048] Simulation verification of effectiveness Multi-scenario simulation analysis: By building a novel 68-node power system simulation system, different virtual inertia optimization schemes were compared and analyzed, verifying the effectiveness of the proposed optimization method. Simulation results show that the proposed optimization method can significantly improve the frequency dynamic performance of the system and ensure frequency security.

[0049] The patent visually demonstrates the optimization effect through charts, comparing the system frequency, frequency change rate, and active power output of the energy storage power station before and after optimization, further proving the advantages of the proposed optimization method.

[0050] In summary, this embodiment, based on the low inertia characteristics of novel power systems leading to differentiated frequency spatial distributions, and addressing the frequency security requirements of the system, proposes a hybrid system state-space model and a node frequency spatiotemporal dynamic response model, including synchronous generators and virtual inertia-controlled energy storage power stations connected to / integrated with the grid. Based on these models, an inertia coordination optimization method for energy storage power stations, considering node frequency security and small-disturbance stability, is proposed, applicable to scenarios with uneven inertia in complex power networks. The proposed virtual inertia optimization method for energy storage power stations can effectively improve the dynamic frequency performance of system nodes with significant differences in frequency spatiotemporal distribution, ensuring that all nodes in the system can meet the dual security constraints of system frequency deviation and frequency change rate, thereby improving the overall stability of the system.

[0051] The optimization method proposed in this embodiment can meet the system frequency safety indicators under hybrid access of grid-connected and grid-connected energy storage stations, and improve the system frequency dynamic characteristics. Meanwhile, based on the simulation results comparing grid-connected and grid-connected access, although both optimized grid-connected and grid-connected energy storage stations can provide sufficient inertia support for the system, the optimization of grid-connected energy storage station access can more effectively improve system stability, and grid-connected energy storage stations have better frequency support for new power systems. However, the instantaneous output power of grid-connected energy storage stations during the inertia response process is higher, and this factor needs to be considered additionally when configuring the converter capacity of energy storage stations.

[0052] This embodiment proposes an effective method for optimizing the virtual inertia of power system energy storage stations by comprehensively considering multiple aspects such as dynamic spatiotemporal distribution of frequency, system stability under small disturbances, inertia configuration of energy storage power stations, and simulation verification. This provides strong support for improving the frequency security and stability of new power systems.

[0053] The above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of the present invention should be included within the protection scope of the present invention.

Claims

1. A method for optimizing the virtual inertia of a power storage station considering the spatiotemporal distribution characteristics of inertia, characterized in that, Includes the following steps: S1. In the scenario of uneven power system inertia distribution under the consideration of high proportion of new energy access, construct source-measured frequency response characteristic model and load node frequency response characteristic model considering load power mapping. S2. The system small-disturbance stability based on the characteristic root method is used as a frequency security constraint. S3. Taking the minimum inertia and frequency regulation energy of the energy storage power station as the objective, and considering the physical constraints, frequency safety constraints and system stability constraints of the energy storage power station, a coordinated optimization strategy for the inertia of the grid-connected / networked energy storage power station is proposed, taking into account the dynamic spatiotemporal distribution of frequency. The objective function is to minimize the sum of the integral of the load node frequency deviation and the inertial kinetic energy of the grid-connected / networked energy storage power station.

2. The virtual inertia optimization method for power storage power stations considering the spatiotemporal distribution characteristics of inertia according to claim 1, characterized in that, The construction of the source-measured frequency response characteristic model considering load power mapping: For a power network with m generating nodes and n load nodes, neglecting voltage amplitude fluctuations at each node and line resistance, the active power in the network is described using a DC model, and the power flow equations for the active power injected into each node are as follows: （1） In the formula, ΔPe=[ΔPe.1,ΔPe.2,…,ΔPe.m]T, Δδ=[Δδ1,Δδ2,…,Δδm]T are the column vectors of the power injection change and the phase angle change of each power generation node, respectively; ΔPL=[ΔPL1,ΔPL2,…,ΔPLn]T, Δθ=[Δθ1,Δθ2,…,Δθn]T are the column vectors of the power injection change and the phase angle change of each load node, respectively; BGG is the diagonal matrix composed of the reactance of the power generation nodes, BLL is the correlation matrix between the power change and the angle change of the load nodes, and BLG and BGL are the correlation matrices between the power generation nodes and the load nodes; The power frequency characteristic models for synchronous machines and grid-connected energy storage generation units are as follows: （2） （3） Where Hm, Dm, HGFM, and DGFM represent the diagonal matrices composed of the inertial time constant and damping coefficient of each synchronous generator and grid-type energy storage power station, respectively, i.e., Hm=diag(Hm1,Hm2,…), Dm=diag(Dm1,Dm2,…), HGFM=diag(HGFM.1, HGFM.2,…), DGFM=diag(DGFM.1, DGFM.2,…); Source-measured frequency response characteristic model considering load power mapping: （4） HGFL is a diagonal matrix composed of the virtual inertia time constants of each energy storage power station using grid-based virtual inertia control, i.e., HGFL=diag(HGFL.1,HGFL.2,…), and DGFL is a diagonal matrix composed of the damping coefficients of each energy storage power station using grid-based virtual inertia control, i.e., DGFL=diag(DGFL1,DGFL2,…). The construction of the load node frequency response characteristic model considering source-measuring frequency mapping: Under the assumptions of unit voltage amplitude and pure reactance transmission lines, and without considering the dynamic differences within the energy storage power station, the energy storage power station adopts the capacity equalization aggregation method. The network nodes are then renumbered in the order of synchronous generators, grid-connected energy storage power stations, and grid-connected energy storage power stations. Assuming that the reference power ΔPset.GFM of the grid-connected energy storage power station remains constant during the frequency response, the new power system frequency dynamic characteristic equation considering the spatiotemporal distribution characteristics of frequency can be expressed as: （5） Following the Kron simplification approach, only the dynamics at the power generation nodes are retained, while load node variables are eliminated. Power disturbances occurring at load nodes are equivalently distributed to the power generation nodes. Except for the generation nodes, the injected current at all nodes is zero. The network equation can be expressed as: （6） in, ∈Rm×1 is a column vector consisting of the current injected into the system by all power generation nodes in the network; ∈Rm×1 and Let Rn×1 be a column vector composed of the potential of the generating nodes and the voltage of the load nodes in the entire network, where the elements are respectively... =Emejδ, =Unejδ; The load node voltage can be expressed as the electromotive force within the entire grid's generating units: （7） In the formula, K represents the complex matrix relating the admittances of generating nodes and load nodes. The load node k and all generating nodes have the following relationship: （8） （9） In the formula, Kk,i is the element in the k-th row and i-th column of K; Simplifying, we obtain the frequency expression for the load nodes. It can be seen that the load node frequency is a weighted superposition of the generator node frequencies with respect to the network topology, i.e.: （10） Equations (5) and (10) together form a new frequency-space-time dynamic response model for power systems.

3. The virtual inertia optimization method for power storage power stations considering the spatiotemporal distribution characteristics of inertia according to claim 1, characterized in that, The system stability constraint refers to determining whether the system satisfies the small disturbance stability criterion by solving the eigenvalues ​​of the system state matrix, i.e., the damping ratio of all eigenvalues ​​is greater than 3% to 5%. The system stability constraint is achieved in the following way: under the assumption of unit voltage amplitude and pure reactive transmission line, the system state matrix is ​​constructed; according to Lyapunov's first method, the eigenvalues ​​of the system state matrix are solved to determine whether the system satisfies the small disturbance stability criterion.

4. The virtual inertia optimization method for power storage power stations considering the spatiotemporal distribution characteristics of inertia according to claim 1, characterized in that, The frequency safety constraint refers to: considering the frequency of all nodes in the system and the frequency change rate RoCoF constraint, ensuring that the frequency and RoCoF of each node are within a safe range after the disturbance.

5. The virtual inertia optimization method for power storage power stations considering the spatiotemporal distribution characteristics of inertia according to claim 4, characterized in that, The frequency security constraints are implemented in the following ways: Set upper and lower limits for the frequency to ensure that the frequency of each node does not exceed the limits after the disturbance; Set the maximum value of RoCoF to ensure that the rate of frequency change of each node is within a safe range after the disturbance.

6. The virtual inertia optimization method for power storage power stations considering the spatiotemporal distribution characteristics of inertia according to claim 1, characterized in that, The physical constraints of the energy storage power station refer to the physical performance constraints of the energy storage power station, including the range of the equivalent backup inertia and the maximum power output capacity.

7. The virtual inertia optimization method for power storage power stations considering the spatiotemporal distribution characteristics of inertia according to claim 6, characterized in that, The physical constraints of the energy storage power station are achieved in the following ways: Consider the range of equivalent backup inertia of energy storage power stations to avoid the inability of energy storage power stations to respond to optimized control commands; Consider the maximum power output capacity of the energy storage power station to ensure that the energy storage power station does not exceed its physical limitations during the optimization process.

8. The virtual inertia optimization method for power storage power stations considering the spatiotemporal distribution characteristics of inertia according to claim 1, characterized in that, The construction of the source-measured frequency response characteristic model and the load node frequency response characteristic model considering load power mapping specifically includes: The network nodes are renumbered in the order of synchronous generators, grid-connected energy storage power stations, and grid-connected energy storage power stations; Assuming that the reference power of the grid-connected energy storage power station remains constant during the frequency response, a new power system frequency dynamic characteristic equation considering the spatiotemporal distribution characteristics of frequency is constructed.

9. The virtual inertia optimization method for power storage power stations considering the spatiotemporal distribution characteristics of inertia according to claim 8, characterized in that, The novel power system frequency dynamic characteristic equation can be expressed as: （11） Where x is the system's state variable, u is the system's disturbance variable, and K L K R A R This corresponds to the system matrix; Linearizing equation (11) at the equilibrium point, the state matrix of the system is constructed as follows: （12） In equation (12), As is the system state matrix; According to Lyapunov's first method, if all eigenvalues ​​have negative real parts, then the system is stable; by solving for the eigenvalues ​​of the system's state matrix, the eigenvalues ​​of the system can be obtained. Let the i-th eigenvalue be... Then the damping ratio expression for the i-th mode is: （13） To ensure system stability under small disturbances, each eigenvalue should have positive damping. When the minimum damping ratio is greater than 3% to 5%, the system is considered to satisfy the small disturbance stability criterion. ζ represents the set of damping ratios of all eigenvalues, set as follows: （14） The specific constraints on node frequency are as follows: （15） In the formula, ω max and ω min These are the upper and lower bounds of the frequency limit, respectively; dω max / dt i The maximum value of RoCoF at node i, RoCoF max It is the limit value for the rate of change of frequency; The physical constraint expressions for energy storage power stations are as follows: （16） Among them, P GFM.min and P GFM.max These represent the upper and lower bounds of the active power output of a grid-connected energy storage power station, P. GFM.i The active power output of the i-th grid-connected energy storage power station; P GFL.min and P GFL.max These represent the upper and lower bounds of the active power output of a grid-connected energy storage power station, P. GFL.i The active power output is for the i-th grid-connected energy storage power station.

10. The virtual inertia optimization method for power storage power stations considering the spatiotemporal distribution characteristics of inertia according to claim 1, characterized in that, The objective function is as follows: （17） In equation (17), T is the inertial response time, a L a G These are the weighting coefficients corresponding to the load node frequency deviation and the generator node kinetic energy deviation, respectively, with the aim of ensuring that the two are of similar order of magnitude.

Images