Energy storage optimization configuration method, device, equipment and medium considering inertia constraints

Abstract

The present invention relates to the technical field of grid-side energy storage optimization configuration, and in particular to a method, device, equipment and medium for optimizing energy storage configuration considering inertia constraints, the method comprising: converting the frequency requirements for safe operation of the power system into system inertia constraints; generating typical wind and solar output scenarios based on SNGAN, and reducing the generated scenarios through AP‑DTW‑K‑medoids; setting the coordinated operation mode of pumped storage and energy storage batteries under a high-proportion new energy system; establishing the objective function and constraint conditions for optimizing the energy storage configuration of a high-proportion new energy system grid considering system inertia constraints with the lowest comprehensive operation cost of the power system, and constructing a mixed integer linear programming model; solving the mixed integer linear programming model to obtain an energy storage optimization configuration plan. The present invention can fully consider the uncertainty of wind and solar output, take into account the stability of the power grid frequency, and realize the complementary operation of energy storage and thermal power and new energy in the power grid.

Description

Technical Field

[0001] The present invention relates to the technical field of grid-side energy storage optimization configuration, and in particular to an energy storage optimization configuration method, device, equipment and medium considering inertia constraints. Background Art

[0002] At present, the installed capacity and power generation of new energy sources such as wind power and photovoltaic power are constantly increasing. Due to the inherent intermittent, volatile and uncertain nature of wind power and photovoltaic power, direct access poses challenges to the safe and stable operation of the power grid. On the one hand, with the continuous increase in the installed capacity of wind power and photovoltaic power generation, the large-scale grid connection of renewable energy has brought about power imbalance and the phenomenon of wind and solar power abandonment. On the other hand, in the context of a high proportion of new energy grid connection, the installed capacity share of thermal power has been continuously replaced by wind power and photovoltaic power generation, resulting in a continuous decline in the system inertia level and anti-interference ability, which will have a great impact on the frequency stability of the system.

[0003] Energy storage technology plays an important role in the power system. It can not only cooperate with pumped storage units to achieve peak load shaving and valley filling in the power system and improve the consumption rate of new energy, but also provide virtual inertia to stabilize the grid frequency. Therefore, energy storage technology is one of the effective measures to solve this problem.

[0004] In the prior art, research has been conducted on how to configure the energy storage capacity required in a power grid containing a high proportion of renewable energy. However, the multi-beam research has not considered the impact of inertia constraints on the capacity configuration results. When a large power disturbance occurs in the power grid, there is a possibility of power exceeding the limit.

[0005] The information disclosed in this background technology section is only intended to deepen the understanding of the overall background technology of the present invention, and should not be regarded as acknowledging or suggesting in any form that the information constitutes the prior art already known to those skilled in the art. Summary of the invention

[0006] The present invention provides a method, device, equipment and medium for optimizing energy storage configuration considering inertia constraints, thereby effectively solving the problems in the background technology.

[0007] In order to achieve the above object, the technical solution adopted by the present invention is: a method for optimizing energy storage configuration considering inertia constraints, comprising the following steps:

[0008] Convert the frequency requirements for safe operation of the power system into system inertia constraints;

[0009] Generate typical wind and solar power output scenarios based on the spectrum normalization generative adversarial network (SNGAN) method, and reduce the generated scenarios using the AP-DTW-K-medoids algorithm based on affinity propagation and dynamic time warping distance.

[0010] Set up the coordinated operation mode of pumped storage and energy storage batteries under a high proportion of new energy systems;

[0011] With the lowest comprehensive operation cost of the power system, the objective function and constraint conditions for the optimal configuration of power grid energy storage in a high-proportion new energy system considering the inertia constraint of the system are established, and a mixed integer linear programming model is constructed;

[0012] The mixed integer linear programming model is solved to obtain an energy storage optimization configuration scheme.

[0013] Furthermore, the step of converting the frequency requirement in the safe operation of the power system into the system inertia constraint comprises the following steps:

[0014] The system frequency response characteristics are expressed using the rotor motion equation:

[0015] ;

[0016] In the formula, is the damping coefficient during the system frequency response; is the system rated frequency; Real-time frequency change of the system; The frequency modulation power for all scheduling resources; The power shortage of the generator;

[0017] Calculate the maximum frequency change rate of the system:

[0018] ;

[0019] In the formula, Indicates the maximum frequency change rate of the system; Indicates the standard value of system frequency; is the real-time inertia of the system;

[0020] Calculate the minimum inertia based on the minimum frequency change rate constraint:

[0021] ;

[0022] Combined with the frequency response mode of each dispatching resource in the high-proportion renewable energy power system and the rotor motion equation, the maximum system frequency deviation level after the fault occurs is calculated, which needs to meet the system transient frequency deviation extreme value constraints, including:

[0023] ;

[0024] In the formula, is the frequency extreme value deviation; It is the maximum frequency variation allowed by the system.

[0025] Furthermore, after converting the frequency requirement in the safe operation of the power system into the system inertia constraint, the system frequency response process is simulated using the piecewise linearization technology, including the following steps:

[0026] The rotor motion equation from 0 to Integrate and solve to obtain:

[0027] ;

[0028] Where: and They are primary frequency modulation control dead zone and action time respectively;

[0029] The system primary frequency modulation rate is used to represent the time from the action of the stabilization control measures to the frequency reaching the transient extreme value. The system frequency reaching the extreme value time is expressed as:

[0030] ;

[0031] The rotor motion equation from 0 to Integrate and solve using the linearization method to obtain:

[0032] ;

[0033] Considering the above three formulas comprehensively, we can get:

[0034] ;

[0035] The minimum inertia requirement of the system is:

[0036] ;

[0037] The system real-time inertia includes:

[0038] ;

[0039] In the formula, , Respectively represent the total number of conventional thermal power units and the total number of pumped storage units; , They represent the rated power and inertia time constant of the synchronous unit respectively; , They represent the rated power and inertia time constant of the pumped storage unit respectively; , They represent the rated power and inertia time constant of the energy storage device respectively.

[0040] Furthermore, the spectrum normalization generative adversarial network (SNGAN) method is used to generate typical wind and solar power output scenarios, including:

[0041] Modify the traditional generative adversarial network GAN:

[0042] Assume the input is , where the expression of the discriminator is , then for a randomly given ,have:

[0043] ;

[0044] In the formula, For function The spectral norm of

[0045] Find the weights and biases of each layer in the discriminator:

[0046] ;

[0047] In the formula, is the weight of each layer in the discriminator, is the bias of each layer in the discriminator, For the discriminator layer;

[0048] According to the above formula, we can get:

[0049] ;

[0050] In the formula, is the discriminator output matrix, and , we can get:

[0051] ;

[0052] Normalizing the weights, we can get: :

[0053] ;

[0054] The global 1-Lipschitz constraint can be implemented by calculating the maximum singular value of the weight matrix of each layer.

[0055] Furthermore, the generated scene is reduced by using the center point segmentation clustering algorithm AP-DTW-K-medoids based on affinity propagation and dynamic time warping distance, comprising the following steps:

[0056] Enter the initial scene matrix , and given the number of scenes that need to be clustered , select the reference type of the affinity propagation AP algorithm and give the threshold ;

[0057] The similarity of the scenes is calculated by the given initial scene matrix, reference type and threshold parameters. The calculation formula is:

[0058] ;

[0059] In the formula, is the Euclidean distance, ;

[0060] Calculate the mutual attraction of each sample scene, including:

[0061] ;

[0062] Calculate the attribution degree of each sample scene, including:

[0063] ;

[0064] Update the belongingness and attractiveness of sample scenarios, including:

[0065] ;

[0066] In the formula, is the damping coefficient;

[0067] Iterate the loop, the cut-off criterion is that the cluster center no longer changes or exceeds the specified number of iterations, and then output the final cluster center , then the number of elements in the cluster center is , satisfy ;

[0068] Calculate the dynamic time warping (DTW) distances between the output cluster centers one by one to form a similarity matrix B;

[0069] Based on the obtained similarity matrix B, a minimum spanning tree is established, and the weight descending principle is used to remove The branches with larger weights are finally obtained. A data matrix is ​​used as the cluster center for the K-medoids algorithm to partition clusters by center points.

[0070] The clustering evaluation index is involved in the obtained The data matrix of each scene is evaluated. If the preset indicators are met, the obtained scene data can be used as the final typical scenery scene.

[0071] Furthermore, the setting of the coordinated operation mode of pumped storage and energy storage batteries under a high proportion of new energy system includes:

[0072] When the gap between supply and demand of the power system is large, it is necessary to prioritize the safety of system operation. Under the premise of meeting system constraints, energy storage batteries will respond first, and pumped storage will take over the response. When the gap between supply and demand of the system is small, under the premise of meeting system constraints, energy storage batteries, pumped storage and thermal power units will respond to power shortages or surpluses in accordance with the principle of optimizing the economic benefits of the overall operation of the system.

[0073] Furthermore, the objective function includes:

[0074] ;

[0075] ;

[0076] ;

[0077] ;

[0078] ;

[0079] In the formula, The comprehensive operating cost of the power system; Indicates The fuel cost of a conventional thermal power unit on a typical day; Indicates The output power of a conventional thermal power unit during the period t at s on a typical day; Indicates The start-up and shutdown costs of a conventional thermal power unit on a typical day are: Indicates The cost of deep peak load regulation of conventional thermal power units in a typical day is: Indicates the unit operating status (1 means running, 0 means shutdown); It represents the unit cost of wind and solar power abandonment; , They represent the predicted output of wind power at time t on a typical day s; represents the capacity of the energy storage configuration, represents the maximum power of the energy storage configuration, Indicates the virtual inertia of the added energy storage; Represents the typical day probability.

[0080] Furthermore, the constraints include:

[0081] System minimum inertia constraint:

[0082] ;

[0083] In the formula, It represents the current inertia of the system during the period t under the typical day s;

[0084] Node power balance constraints:

[0085] ;

[0086] In the formula, The node for a typical day With Node The active power between is the node at time t on a typical day s The load size;

[0087] DC power flow constraints:

[0088] ;

[0089] In the formula, For Node and nodes The admittance between and They are the nodes in the period t under the typical day s and nodes The phase angle difference between and They are the nodes connected in the period t under the typical day s and The lower and upper limits of the active power of the line;

[0090] Phase Angle Constraint:

[0091] ;

[0092] In the formula, and are the lower and upper limits of the phase angle of node b respectively;

[0093] Output upper and lower limits of thermal power units:

[0094] ;

[0095] ;

[0096] ;

[0097] In the formula, , Represents conventional thermal power The upper and lower limits of output; , Respectively represent The normal peak-shaving and deep peak-shaving output of conventional thermal power units at time t on a typical day; and Respectively The normal peak load regulation and deep peak load regulation output status of a conventional thermal power unit at time t on a typical day. If the unit is in the startup state, it is 1, otherwise it is in the shutdown state, it is 0;

[0098] Other constraints:

[0099] Minimum start and stop time constraints for thermal power units, climbing constraints for thermal power units, wind and solar output constraints, pumped storage unit constraints and energy storage battery constraints.

[0100] Furthermore, the mixed integer linear programming model is solved to obtain an energy storage optimization configuration scheme, including:

[0101] The mixed integer linear programming model is solved by calling the Gurobi solver using Yalmip to obtain the maximum capacity, power and virtual inertia of each energy storage configuration in a high-proportion new energy grid.

[0102] The present invention also includes an energy storage optimization configuration device considering inertia constraints, using the above method, including:

[0103] Inertia constraint unit, used to convert the frequency requirements for safe operation of the power system into system inertia constraints;

[0104] The scenario generation unit is used to generate typical wind and solar power output scenarios based on the spectrum normalization generative adversarial network (SNGAN) method, and reduce the generated scenarios through the center point segmentation clustering algorithm AP-DTW-K-medoids based on affinity propagation and dynamic time warping distance;

[0105] A coordinated operation unit is used to set the coordinated operation mode of pumped storage and energy storage batteries under a high proportion of new energy systems;

[0106] A modeling unit, for establishing an objective function and constraint conditions for optimizing the configuration of power grid energy storage of a high-proportion new energy system taking into account the inertia constraint of the system with the lowest comprehensive operation cost of the power system, and constructing a mixed integer linear programming model;

[0107] A solving unit is used to solve the mixed integer linear programming model to obtain an energy storage optimization configuration plan.

[0108] The present invention also includes a computer device, including a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein when the processor executes the computer program, the method described above is implemented.

[0109] The present invention also includes a storage medium on which a computer program is stored. When the computer program is executed by a processor, the method described above is implemented.

[0110] The beneficial effects of the present invention are as follows: the system inertia is taken into consideration in the present invention, and the frequency requirements in the safe operation of the power system are converted into constraints on the system inertia, which can fully consider the uncertainty of wind and solar power output, take into account the stability of the power grid frequency, and realize the complementary operation of energy storage and thermal power and new energy in the power grid. The generated scenarios are reduced by the AP-DTW-K-medoids method, which is a center point segmentation clustering algorithm based on affinity propagation and dynamic time warping distance, thereby improving the accuracy of energy storage capacity and virtual inertia planning, and reducing the impact of repeated scenarios on the planning results. BRIEF DESCRIPTION OF THE DRAWINGS

[0111] In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the drawings required for use in the embodiments or the description of the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments recorded in the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying creative work.

[0112] Figure 1 is a flow chart of the method in Example 1;

[0113] Figure 2 It is a structural schematic diagram of the device in Example 1;

[0114] Figure 3 Schematic diagram of the process in Example 2;

[0115] Figure 4 Wind power output diagrams for five typical days obtained after scene generation and cluster reduction in Example 2;

[0116] Figure 5 The photovoltaic output diagrams of five typical days obtained after scene generation and cluster reduction in Example 2;

[0117] Figure 6 The load output diagrams of five typical days are obtained after the scene generation and cluster reduction in Example 2;

[0118] Figure 7 This is a graph showing the output results of 5 system units on a typical day in Example 2;

[0119] Figure 8 This is a graph showing the startup results of 5 thermal power units on a typical day in Example 2;

[0120] Fig. 9 This is a graph of the system inertia results for a typical day 5 in Example 2;

[0121] Fig.10 A schematic diagram of the structure of a computer device. DETAILED DESCRIPTION

[0122] The technical solutions in the embodiments of the present invention will be described clearly and completely below in conjunction with the drawings in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, rather than all the embodiments.

[0123] Embodiment 1:

[0124] like Figure 1 As shown: A method for optimizing energy storage configuration considering inertia constraints includes the following steps:

[0125] Convert the frequency requirements for safe operation of the power system into system inertia constraints;

[0126] Generate typical wind and solar power output scenarios based on the spectral normalization generative adversarial network (SNGAN) method, and reduce the generated scenarios through the center point segmentation clustering algorithm (K-medoids) based on affinity propagation (AP) and dynamic time warping distance (DTW), namely the AP-DTW-K-medoids method;

[0127] Set up the coordinated operation mode of pumped storage and energy storage batteries under a high proportion of new energy systems;

[0128] With the lowest comprehensive operating cost of the power system, the objective function and constraint conditions for the optimal configuration of power grid energy storage in a high-proportion new energy system considering system inertia constraints are established, and a mixed integer linear programming model is constructed;

[0129] The mixed integer linear programming model is solved to obtain the optimal energy storage configuration solution.

[0130] Taking system inertia into consideration, the frequency requirements for safe operation of the power system are converted into system inertia constraints, which can fully consider the uncertainty of wind and solar power output, take into account the stability of the power grid frequency, and realize the complementary operation of energy storage and thermal power and renewable energy in the power grid. The generated scenarios are reduced by the AP-DTW-K-medoids method, which is a center point segmentation clustering algorithm based on affinity propagation and dynamic time warping distance, to improve the accuracy of energy storage capacity and virtual inertia planning, while reducing the impact of repeated scenarios on planning results.

[0131] In this embodiment, the frequency requirement for safe operation of the power system is converted into a system inertia constraint, including the following steps:

[0132] The system frequency response characteristics are expressed using the rotor motion equation:

[0133] ;

[0134] In the formula, is the damping coefficient during the system frequency response; is the system rated frequency; Real-time frequency change of the system; The frequency modulation power for all scheduling resources; The power shortage of the generator;

[0135] Calculate the maximum frequency change rate of the system:

[0136] ;

[0137] In the formula, Indicates the maximum frequency change rate of the system; Indicates the standard value of system frequency; is the real-time inertia of the system;

[0138] Calculate the minimum inertia based on the minimum frequency change rate constraint:

[0139] ;

[0140] Combined with the frequency response mode of each dispatching resource in the high-proportion renewable energy power system and the rotor motion equation, the maximum system frequency deviation level after the fault occurs is calculated, which needs to meet the system transient frequency deviation extreme value constraints, including:

[0141] ;

[0142] In the formula, is the frequency extreme value deviation; It is the maximum frequency variation allowed by the system.

[0143] After the frequency requirements for safe operation of the power system are converted into system inertia constraints, the piecewise linearization technology is used to simulate the system frequency response process, including the following steps:

[0144] The rotor motion equation from 0 to Integrate and solve to obtain:

[0145] ;

[0146] Where: and They are primary frequency modulation control dead zone and action time respectively;

[0147] The system primary frequency modulation rate is used to represent the time from the action of the stabilization control measures to the frequency reaching the transient extreme value. The system frequency reaching the extreme value time is expressed as:

[0148] ;

[0149] The rotor motion equation from 0 to Integrate and solve using the linearization method to obtain:

[0150] ;

[0151] Considering the above three formulas comprehensively, we can get:

[0152] ;

[0153] The minimum inertia requirement of the system is:

[0154] ;

[0155] The system real-time inertia includes:

[0156] ;

[0157] In the formula, , Respectively represent the total number of conventional thermal power units and the total number of pumped storage units; , They represent the rated power and inertia time constant of the synchronous unit respectively; , They represent the rated power and inertia time constant of the pumped storage unit respectively; , They represent the rated power and inertia time constant of the energy storage device respectively.

[0158] Among them, the typical wind and solar power output scenarios are generated based on the spectrum normalization generative adversarial network (SNGAN) method, including:

[0159] Modify the traditional generative adversarial network GAN:

[0160] Assume the input is , where the expression of the discriminator is , then for a randomly given ,have:

[0161] ;

[0162] In the formula, For function The spectral norm of

[0163] Find the weights and biases of each layer in the discriminator:

[0164] ;

[0165] In the formula, is the weight of each layer in the discriminator, is the bias of each layer in the discriminator, For the discriminator layer;

[0166] According to the above formula, we can get:

[0167] ;

[0168] In the formula, is the discriminator output matrix, and , we can get:

[0169] ;

[0170] Normalizing the weights, we can get: :

[0171] ;

[0172] The global 1-Lipschitz constraint can be implemented by calculating the maximum singular value of the weight matrix of each layer.

[0173] As a preferred embodiment of the above, the generated scene is reduced by a center point segmentation clustering algorithm based on affinity propagation and dynamic time warping distance, namely, AP-DTW-K-medoids method, including the following steps:

[0174] Enter the initial scene matrix , and given the number of scenes that need to be clustered , select the reference type of the affinity propagation AP algorithm and give the threshold ;

[0175] The similarity of the scenes is calculated by the given initial scene matrix, reference type and threshold parameters. The calculation formula is:

[0176] ;

[0177] In the formula, is the Euclidean distance, ;

[0178] Calculate the mutual attraction of each sample scene, including:

[0179] ;

[0180] Calculate the attribution degree of each sample scene, including:

[0181] ;

[0182] Update the belongingness and attractiveness of sample scenarios, including:

[0183] ;

[0184] In the formula, is the damping coefficient;

[0185] Iterate the loop, the cut-off criterion is that the cluster center no longer changes or exceeds the specified number of iterations, and then output the final cluster center , then the number of elements in the cluster center is , satisfy ;

[0186] Calculate the dynamic time warping (DTW) distances between the output cluster centers one by one to form a similarity matrix B;

[0187] Based on the obtained similarity matrix B, a minimum spanning tree is established, and the weight descending principle is used to remove The branches with larger weights are finally obtained. A data matrix is ​​used as the cluster center for the K-medoids algorithm to partition clusters by center points.

[0188] The clustering evaluation index is involved in the obtained The data matrix of each scene is evaluated. If the preset indicators are met, the obtained scene data can be used as the final typical scenery scene.

[0189] In this embodiment, the coordinated operation mode of pumped storage and energy storage batteries under a high proportion of new energy system is set, including:

[0190] When the gap between supply and demand of the power system is large, it is necessary to prioritize the safety of system operation. Under the premise of meeting system constraints, energy storage batteries will respond first, and pumped storage will take over the response. When the gap between supply and demand of the system is small, under the premise of meeting system constraints, energy storage batteries, pumped storage and thermal power units will respond to power shortages or surpluses in accordance with the principle of optimizing the economic benefits of the overall operation of the system.

[0191] The objective function includes:

[0192] ;

[0193] ;

[0194] ;

[0195] ;

[0196] ;

[0197] In the formula, The comprehensive operating cost of the power system; Indicates The fuel cost of a conventional thermal power unit on a typical day; Indicates The output power of a conventional thermal power unit during the period t at s on a typical day; Indicates The start-up and shutdown costs of a conventional thermal power unit on a typical day are: Indicates The cost of deep peak load regulation of conventional thermal power units in a typical day is: Indicates the unit operating status (1 means running, 0 means shutdown); It represents the unit cost of wind and solar power abandonment; , They represent the predicted output of wind power at time t on a typical day s; represents the capacity of the energy storage configuration, represents the maximum power of the energy storage configuration, Indicates the virtual inertia of the added energy storage; Represents the typical day probability.

[0198] The constraints include:

[0199] System minimum inertia constraint:

[0200] ;

[0201] In the formula, It represents the current inertia of the system during the period t under the typical day s;

[0202] Node power balance constraints:

[0203] ;

[0204] In the formula, The node for a typical day With Node The active power between is the node at time t on a typical day s The load size;

[0205] DC power flow constraints:

[0206] ;

[0207] In the formula, For Node and nodes The admittance between and They are the nodes in the period t under the typical day s and nodes The phase angle difference between and They are the nodes connected in the period t under the typical day s and The lower and upper limits of the active power of the line;

[0208] Phase Angle Constraint:

[0209] ;

[0210] In the formula, and are the lower and upper limits of the phase angle of node b respectively;

[0211] Output upper and lower limits of thermal power units:

[0212] ;

[0213] ;

[0214] ;

[0215] In the formula, , Represents conventional thermal power The upper and lower limits of output; , Respectively represent The normal peak-shaving and deep peak-shaving output of conventional thermal power units at time t on a typical day; and Respectively The normal peak load regulation and deep peak load regulation output status of a conventional thermal power unit at time t on a typical day. If the unit is in the startup state, it is 1, otherwise it is in the shutdown state, it is 0;

[0216] Other constraints:

[0217] Minimum start and stop time constraints for thermal power units, climbing constraints for thermal power units, wind and solar output constraints, pumped storage unit constraints and energy storage battery constraints.

[0218] As a preferred embodiment of the above, solving the mixed integer linear programming model to obtain an energy storage optimization configuration scheme includes:

[0219] Yalmip is used to call the Gurobi solver to solve the mixed integer linear programming model, and the maximum capacity, power and virtual inertia of each energy storage configuration in a high-proportion new energy grid are obtained.

[0220] like Figure 2As shown, this embodiment also includes an energy storage optimization configuration device considering inertia constraints, using the above method, including:

[0221] Inertia constraint unit, used to convert the frequency requirements for safe operation of the power system into system inertia constraints;

[0222] The scenario generation unit is used to generate typical wind and solar power output scenarios based on the spectrum normalization generative adversarial network (SNGAN) method, and reduce the generated scenarios through the center point segmentation clustering algorithm AP-DTW-K-medoids based on affinity propagation and dynamic time warping distance;

[0223] A coordinated operation unit is used to set the coordinated operation mode of pumped storage and energy storage batteries under a high proportion of new energy systems;

[0224] A modeling unit is used to establish the objective function and constraint conditions for optimizing the configuration of power grid energy storage of a high-proportion new energy system considering the system inertia constraint with the lowest comprehensive operation cost of the power system, and to construct a mixed integer linear programming model;

[0225] The solving unit is used to solve the mixed integer linear programming model to obtain the energy storage optimization configuration plan.

[0226] Embodiment 2:

[0227] like Figure 3 As shown, this embodiment includes a method for optimizing energy storage configuration considering inertia constraints, including the following steps:

[0228] S110: Convert the frequency requirements for safe operation of the power system into system inertia constraints;

[0229] S120: Generate typical wind and solar power output scenarios based on the SNGAN method, and reduce the generated scenarios using the AP-DTW-K-medoids method;

[0230] S130: Explore the coordinated operation mode of pumped storage and energy storage batteries under a high proportion of new energy systems;

[0231] S140: To minimize the comprehensive operation cost of the power system, establish the objective function and constraint conditions for the optimal configuration of power grid energy storage in a high-proportion new energy system considering inertia constraints;

[0232] S150: Constructing a mixed integer linear programming model for optimizing energy storage configuration of a high-proportion new energy power grid considering inertia constraints based on the objective function and constraint conditions; solving the mixed integer linear programming model to obtain a method for optimizing energy storage configuration of a high-proportion new energy power grid considering inertia constraints.

[0233] Specifically, the frequency requirements for safe operation of the power system are converted into system inertia constraints, including:

[0234] The system frequency response characteristics can be expressed by the rotor motion equation:

[0235] ;

[0236] in, is the damping coefficient during the system frequency response; is the system rated frequency; Real-time frequency change of the system; The frequency modulation power for all scheduling resources; The power shortage of the generator;

[0237] The extreme value of the system frequency change rate usually occurs at the moment when the system fails. According to the rotor motion equation, the calculation expression of the maximum frequency change rate of the system can be calculated:

[0238] ;

[0239] in, Indicates the maximum frequency change rate of the system; Indicates the standard value of system frequency; is the real-time inertia of the system; further, the minimum inertia based on the minimum frequency change rate constraint can be obtained as:

[0240] ;

[0241] Combined with the frequency response mode of each dispatching resource in the high-proportion renewable energy power system and the rotor motion equation, the maximum system frequency deviation level after the fault occurs is calculated, which needs to meet the system transient frequency deviation extreme value constraints, including:

[0242] ;

[0243] in, is the frequency extreme value deviation; is the maximum frequency variation deviation allowed by the system; the above formula is a nonlinear constraint. In previous studies, the SFR model is often established for time domain simulation or the SFR model is simplified by order reduction to solve it. However, as the number of units in the system increases, the control strategies of new energy units are also different, which leads to an increase in the order of the SFR model. The model establishment is very complicated, which will also restrict the efficiency of model solution. This paper uses piecewise linearization technology to simulate the system frequency response process, including:

[0244] The rotor motion equation from 0 to Integrate and solve to obtain:

[0245] ;

[0246] in: and They are primary frequency modulation control dead zone and action time respectively;

[0247] In addition, the system primary frequency modulation rate is used to represent the time from the action of the stabilization control measures to the frequency reaching the transient extreme value. The system frequency reaching the extreme value time can be expressed as:

[0248] ;

[0249] The rotor motion equation from 0 to Integrate and solve using linearization method, we have:

[0250] ;

[0251] Considering the above three formulas comprehensively, we can get:

[0252] ;

[0253] The system inertia must be higher than both the minimum inertia based on the minimum frequency change rate constraint and the minimum inertia based on the system transient frequency deviation extreme value constraint to ensure the frequency stability during the operation of the power system. Therefore, the system minimum inertia requirement is:

[0254] ;

[0255] The system real-time inertia includes:

[0256] ;

[0257] in, , Respectively represent the total number of conventional thermal power units and the total number of pumped storage units; , They represent the rated power and inertia time constant of the synchronous unit respectively; , They represent the rated power and inertia time constant of the pumped storage unit respectively; , They represent the rated power and inertia time constant of the energy storage device respectively.

[0258] Specifically, the generation of the typical time sequence scenarios of new energy mainly includes the generation and reduction of typical scenarios of wind and solar power output, including:

[0259] The generation of typical wind and solar power scenarios is achieved based on the spectral normalization generative adversarial network (SNGAN) method. The modifications to the traditional GAN ​​include:

[0260] Assume the input is , where the expression of the discriminator is , then for a randomly given ,have:

[0261] ;

[0262] in, For function Then, find the weights and biases of each layer in the discriminator:

[0263] ;

[0264] in, is the weight of each layer in the discriminator, is the bias of each layer in the discriminator, For the discriminator Layer. According to the above formula, we can get:

[0265] ;

[0266] in, is the discriminator output matrix, and , we can get:

[0267] ;

[0268] Normalizing the weights, we can get: :

[0269] ;

[0270] Using the above inequality, we only need to calculate the maximum singular value of the weight matrix of each layer to achieve the global 1-Lipschitz constraint.

[0271] The wind-solar output generated by SNGAN contains many output scenarios. In order to improve the accuracy of energy storage capacity and virtual inertia planning, and reduce the impact of repeated scenarios on planning results, it is necessary to remove the output scenarios with high similarity. Commonly used scenario reduction methods include affinity propagation, k-means, and simultaneous backward reduction algorithms. The scenario reduction method used in this paper is the AP-DTW-K-medoids scenario reduction method that combines the advantages of the three. The specific steps include:

[0272] 1) Input the initial scene matrix , and given the number of scenes that need to be clustered , select the reference type of AP algorithm and give the threshold ;

[0273] 2) Calculate the similarity of the scene using the given initial scene matrix, reference type and threshold parameters. The calculation formula is:

[0274] ;

[0275] in, is the Euclidean distance, .

[0276] 3) Calculate the mutual attraction of each sample scene, including:

[0277] ;

[0278] 4) Calculate the degree of belonging of each sample scene, including:

[0279] ;

[0280] 5) Update the belongingness and attractiveness of sample scenarios, including:

[0281] ;

[0282] in, is the damping coefficient.

[0283] 6) Iteration cycle, the cut-off criterion is that the cluster center no longer changes or exceeds the specified number of iterations, and then the final cluster center is output , then the number of elements in the cluster center is , satisfy .

[0284] 7) Calculate the DTW distance between the output cluster centers one by one to form a similarity matrix B.

[0285] 8) Establish a minimum spanning tree based on the obtained similarity matrix B, and remove The branches with larger weights are finally obtained. A data matrix is ​​used as the cluster center obtained by the K-medoids algorithm.

[0286] 9) Participate in clustering evaluation indexes to obtain The data matrix of each scene is evaluated. If the preset indicators are met, the obtained scene data can be used as the final typical scenery scene.

[0287] Specifically, the coordinated operation mode of pumped storage and new energy storage under a high proportion of new energy systems is explored, including:

[0288] Pumped storage is characterized by large capacity and strong regulation capability, but is subject to greater environmental restrictions; energy storage batteries are characterized by fast response and flexible regulation. By coordinating pumped storage and new energy storage, we can provide comprehensive and multi-dimensional support for high-proportion new energy systems, including power supply, emergency standby, and improving system regulation capabilities, thereby ensuring the stable operation of the power system.

[0289] Specifically, it is necessary to adjust the coordinated operation mode of pumped storage and energy storage batteries based on the basic principles of system frequency safety and economy in combination with the demand for frequency regulation of high-proportion new energy systems. When the difference between power system supply and demand is large, it is necessary to prioritize the safety of system operation. Under the premise of meeting system constraints, energy storage batteries will respond first, and pumped storage will take over the response; when the difference between system supply and demand is small, under the premise of meeting system constraints, energy storage batteries, pumped storage and thermal power units will respond to power shortages or surpluses in accordance with the principle of optimal economic benefits of overall system operation, and pumped storage will usually respond first.

[0290] Specifically, the objective function of optimizing energy storage configuration in a power grid with a high proportion of new energy considering inertia constraints is established, including:

[0291] ;

[0292] ;

[0293] ;

[0294] ;

[0295] ;

[0296] in, The comprehensive operating cost of the power system; Indicates The fuel cost of a conventional thermal power unit on a typical day; Indicates The output power of a conventional thermal power unit during the period t at s on a typical day; Indicates The start-up and shutdown costs of a conventional thermal power unit on a typical day are: Indicates The cost of deep peak load regulation of conventional thermal power units in a typical day is: Indicates the unit operating status (1 means running, 0 means shutdown); It represents the unit cost of wind and solar power abandonment; , They represent the predicted output of wind power at time t on a typical day s; represents the capacity of the energy storage configuration, represents the maximum power of the energy storage configuration, Indicates the virtual inertia of the added energy storage; Represents the typical day probability.

[0297] Specifically, the capacity optimization configuration constraint conditions considering inertia constraint as the main consideration are established, including:

[0298] System minimum inertia constraint:

[0299] ;

[0300] in, It represents the current inertia of the system during the period t under the typical day s;

[0301] Node power balance constraints:

[0302] ;

[0303] in, The node for a typical day With Node The active power between is the node at time t on a typical day s The load size;

[0304] DC power flow constraints:

[0305] ;

[0306] in, For Node and nodes The admittance between and They are the nodes in the period t under the typical day s and nodes The phase angle difference between and They are the nodes connected in the period t under the typical day s and The lower and upper limits of the active power of the line;

[0307] Phase Angle Constraint:

[0308] ;

[0309] in, and are the lower and upper limits of the phase angle of node b respectively;

[0310] Output upper and lower limits of thermal power units:

[0311] ;

[0312] ;

[0313] ;

[0314] in, , Represents conventional thermal power The upper and lower limits of output; , Respectively represent The normal peak-shaving and deep peak-shaving output of conventional thermal power units at time t on a typical day; and Respectively The normal peak-shaving and deep peak-shaving output status of a conventional thermal power unit at time t on a typical day. If the unit is in the startup state, it is 1; otherwise, if it is in the shutdown state, it is 0.

[0315] Other constraints include minimum start and stop time constraints for thermal power units, thermal power unit climbing constraints, wind and solar output constraints, pumped storage unit constraints and energy storage battery constraints, which will not be repeated here.

[0316] Specifically, a mixed integer linear programming model for optimizing energy storage configuration of a high-proportion renewable energy grid considering inertia constraints is solved to obtain an optimizing energy storage configuration method for a high-proportion renewable energy grid considering inertia constraints, including:

[0317] Yalmip is used to call the Gurobi solver to solve the mixed integer linear programming model for the optimal configuration of energy storage in a high-proportion new energy grid considering inertia constraints, and the maximum capacity, power and virtual inertia of each energy storage configuration in the high-proportion new energy grid are obtained.

[0318] In order to demonstrate the effectiveness of the high-proportion renewable energy grid energy storage optimization configuration method considering inertia constraints provided by the present invention, it is described below in conjunction with specific application scenarios.

[0319] The embodiment of the present invention adopts a modified IEEE-39 node system to verify the accuracy of the model. The maximum frequency deviation of the system is set to 1 Hz, that is, the lower limit of the frequency in the system dynamic process is 49 Hz, the maximum frequency change rate of the system is 1 Hz / s, and the system power shortage is assumed to be 10% of the real-time load level.

[0320] Convert the frequency constraint to inertia constraint as follows:

[0321] The system frequency response characteristics can be expressed by the rotor motion equation as follows:

[0322] ;

[0323] in, is the damping coefficient during the system frequency response; is the system rated frequency; Real-time frequency change of the system; The frequency modulation power for all scheduling resources; The power shortage of the generator;

[0324] The extreme value of the system frequency change rate usually occurs at the moment when the system fails. According to the rotor motion equation, the calculation expression of the maximum frequency change rate of the system can be calculated:

[0325] ;

[0326] in, Indicates the maximum frequency change rate of the system; Indicates the standard value of system frequency; is the real-time inertia of the system; further, the minimum inertia based on the minimum frequency change rate constraint can be obtained as:

[0327] ;

[0328] Combined with the frequency response mode of each dispatching resource in the high-proportion renewable energy power system and the rotor motion equation, the maximum system frequency deviation level after the fault occurs is calculated, which needs to meet the system transient frequency deviation extreme value constraints, including:

[0329] ;

[0330] in, is the frequency extreme value deviation; is the maximum frequency variation deviation allowed by the system; the above formula is a nonlinear constraint. In previous studies, the SFR model is often established for time domain simulation or the SFR model is simplified by order reduction to solve it. However, as the number of units in the system increases, the control strategies of new energy units are also different, which leads to an increase in the order of the SFR model. The model establishment is very complicated, which will also restrict the efficiency of model solution. This paper uses piecewise linearization technology to simulate the system frequency response process, including:

[0331] The rotor motion equation from 0 to Integrate and solve to obtain:

[0332] ;

[0333] in: and They are primary frequency modulation control dead zone and action time respectively;

[0334] In addition, the system primary frequency modulation rate is used to represent the time from the action of the stabilization control measures to the frequency reaching the transient extreme value. The system frequency reaching the extreme value time can be expressed as:

[0335] ;

[0336] The rotor motion equation from 0 to Integrate and solve using linearization method, we have:

[0337] ;

[0338] Considering the above three formulas comprehensively, we can get:

[0339] ;

[0340] The system inertia must be higher than both the minimum inertia based on the minimum frequency change rate constraint and the minimum inertia based on the system transient frequency deviation extreme value constraint to ensure the frequency stability during the operation of the power system. Therefore, the system minimum inertia requirement is:

[0341] ;

[0342] The system real-time inertia includes:

[0343] ;

[0344] in, , Respectively represent the total number of conventional thermal power units and the total number of pumped storage units; , They represent the rated power and inertia time constant of the synchronous unit respectively; , They represent the rated power and inertia time constant of the pumped storage unit respectively; , They represent the rated power and inertia time constant of the energy storage device respectively.

[0345] Next is the generation of typical time series scenarios for new energy, which mainly includes the generation and reduction of typical scenarios for wind and solar power output, as follows:

[0346] The generation of typical wind and solar power scenarios is achieved based on the spectral normalization generative adversarial network (SNGAN) method. The modifications to the traditional GAN ​​include:

[0347] Assume the input is , where the expression of the discriminator is , then for a randomly given ,have:

[0348] ;

[0349] in, For function Then, find the weights and biases of each layer in the discriminator:

[0350] ;

[0351] in, is the weight of each layer in the discriminator, is the bias of each layer in the discriminator, For the discriminator Layer. According to the above formula, we can get:

[0352] ;

[0353] in, is the discriminator output matrix, and , we can get:

[0354] ;

[0355] Normalizing the weights, we can get: :

[0356] ;

[0357] Using the above inequality, we only need to calculate the maximum singular value of the weight matrix of each layer to achieve the global 1-Lipschitz constraint.

[0358] The wind-solar output generated by SNGAN contains many output scenarios. In order to improve the accuracy of energy storage capacity and virtual inertia planning, and reduce the impact of repeated scenarios on planning results, it is necessary to remove the output scenarios with high similarity. Commonly used scenario reduction methods include affinity propagation, k-means, and simultaneous backward reduction algorithms. The scenario reduction method used in this paper is the AP-DTW-K-medoids scenario reduction method that combines the advantages of the three. The specific steps include:

[0359] 1) Input the initial scene matrix , and given the number of scenes that need to be clustered , select the reference type of AP algorithm and give the threshold ;

[0360] 2) Calculate the similarity of the scene using the given initial scene matrix, reference type and threshold parameters. The calculation formula is:

[0361] ;

[0362] in, is the Euclidean distance, .

[0363] 3) Calculate the mutual attraction of each sample scene, including:

[0364] ;

[0365] 4) Calculate the degree of belonging of each sample scene, including:

[0366] ;

[0367] 5) Update the belongingness and attractiveness of sample scenarios, including:

[0368] ;

[0369] in, is the damping coefficient.

[0370] 6) Iteration cycle, the cut-off criterion is that the cluster center no longer changes or exceeds the specified number of iterations, and then the final cluster center is output , then the number of elements in the cluster center is , satisfy .

[0371] 7) Calculate the DTW distance between the output cluster centers one by one to form a similarity matrix B.

[0372] 8) Establish a minimum spanning tree based on the obtained similarity matrix B, and remove The branches with larger weights are finally obtained. A data matrix is ​​used as the cluster center obtained by the K-medoids algorithm.

[0373] 9) Participate in clustering evaluation indexes to obtain The data matrix of each scene is evaluated. If the preset indicators are met, the obtained scene data can be used as the final typical scenery scene.

[0374] The probability of working on different typical days is shown in Table 1.

[0375] Table 1 Typical daily output probability

[0376] typical day 1 2 3 4 5 Probability 0.1726 0.2822 0.2959 0.1096 0.1397

[0377] Next, we explore the coordinated operation of pumped storage and new energy storage in a high-proportion new energy system, as follows:

[0378] Pumped storage is characterized by large capacity and strong regulation capability, but is subject to greater environmental restrictions; energy storage batteries are characterized by fast response and flexible regulation. By coordinating pumped storage and new energy storage, we can provide comprehensive and multi-dimensional support for high-proportion new energy systems, including power supply, emergency standby, and improving system regulation capabilities, thereby ensuring the stable operation of the power system.

[0379] Specifically, combined with the needs of high-proportion new energy systems for frequency regulation, the coordinated operation mode of pumped storage and energy storage batteries is adjusted based on the basic principles of system frequency safety and economy. When the difference between power supply and demand is large, it is necessary to prioritize the safety of system operation. Under the premise of meeting system constraints, energy storage batteries will respond first, and pumped storage will take over the response; when the difference between system supply and demand is small, under the premise of meeting system constraints, energy storage batteries, pumped storage and thermal power units will respond to power shortages or surpluses in accordance with the principle of optimal economic benefits of overall system operation, and pumped storage will usually respond first.

[0380] Then, the objective function of optimizing energy storage configuration in a power grid with a high proportion of new energy considering inertia constraints is established as follows:

[0381] ;

[0382] ;

[0383] ;

[0384] ;

[0385] ;

[0386] in, The comprehensive operating cost of the power system; Indicates The fuel cost of a conventional thermal power unit on a typical day; Indicates The output power of a conventional thermal power unit during the period t at s on a typical day; Indicates The start-up and shutdown costs of a conventional thermal power unit on a typical day are: Indicates The cost of deep peak load regulation of conventional thermal power units in a typical day is: Indicates the unit operating status (1 means running, 0 means shutdown); It represents the unit cost of wind and solar power abandonment; , They represent the predicted output of wind power at time t on a typical day s; represents the capacity of the energy storage configuration, represents the maximum power of the energy storage configuration, Indicates the virtual inertia of the added energy storage; Represents the typical day probability.

[0387] Then, the capacity optimization configuration constraint conditions considering inertia constraint are established as follows:

[0388] System minimum inertia constraint:

[0389] ;

[0390] in, It represents the current inertia of the system during the period t under the typical day s;

[0391] Node power balance constraints:

[0392] ;

[0393] in, The node for a typical day With Node The active power between is the node at time t on a typical day s The load size;

[0394] DC power flow constraints:

[0395] ;

[0396] in, For Node and nodes The admittance between and They are the nodes in the period t under the typical day s and nodes The phase angle difference between and They are the nodes connected in the period t under the typical day s and The lower and upper limits of the active power of the line;

[0397] Phase Angle Constraint:

[0398] ;

[0399] in, and are the lower and upper limits of the phase angle of node b respectively;

[0400] Output upper and lower limits of thermal power units:

[0401] ;

[0402] ;

[0403] ;

[0404] in, , Represents conventional thermal power The upper and lower limits of output; , Respectively represent The normal peak-shaving and deep peak-shaving output of conventional thermal power units at time t on a typical day; and Respectively The normal peak-shaving and deep peak-shaving output status of a conventional thermal power unit at time t on a typical day. If the unit is in the startup state, it is 1; otherwise, if it is in the shutdown state, it is 0.

[0405] The parameters of thermal power units are shown in Table 2:

[0406] Table 2 Thermal power unit parameters

[0407] Thermal power station number Installed Node Maximum output of the unit (MW) Minimum output of normal peak load regulation (MW) Minimum output of deep peak load regulation (MW) Unit inertia time constant (S) 1 31 677 474.5 271.1 5.8 2 32 650 455 260 5.0 3 33 632 442.4 252.8 5.4 4 34 508 355.6 203.2 5.0 5 35 650 455 260 5.6 6 36 560 392 224 5.7 7 37 540 378 216 4.7 8 38 830 581 332 3.8

[0408] Other constraints include minimum start and stop time constraints for thermal power units, thermal power unit climbing constraints, wind and solar output constraints, pumped storage unit constraints and energy storage battery constraints, which will not be repeated here.

[0409] The parameters of the pumped storage unit are shown in Table 3:

[0410] Table 3 Pumped storage unit parameters

[0411] Unit No. Maximum power generation output (MW) Minimum power generation output (MW) Maximum pumping output (MW) Minimum pumping output (MW) 1 160 80 160 160 2 160 80 160 160 3 160 80 160 160 4 160 80 160 160

[0412] Energy storage power station parameters are shown in Table 4:

[0413] Table 4 Energy storage power station configuration parameters

[0414] parameter Energy storage power station Unit power cost / (yuan / KW) 350 Unit capacity cost / (yuan / KWh) 1800 Unit virtual inertia cost / (yuan / S) 1500 discount rate / % 5 Charging efficiency 0.9 Discharge efficiency 0.9 Maximum state of charge 0.9 Minimum state of charge 0.1 Design service life 10

[0415] Figure 4 , Figure 5 and Figure 6 The wind power output diagram, photovoltaic output diagram, and load output diagram of five typical days obtained after scenario generation and cluster reduction are shown respectively;

[0416] Figure 7 , Figure 8 , Fig. 9 The output results of each unit in the power grid, the startup results of thermal power units and the results of system inertia are displayed respectively.

[0417] Finally, the mixed integer linear programming model of the optimal configuration of energy storage for a high-proportion new energy grid considering inertia constraints is solved to obtain the optimal configuration method of energy storage for a high-proportion new energy grid considering inertia constraints and the maximum capacity, power and virtual inertia of each energy storage configuration in the high-proportion new energy grid. The results are as follows:

[0418] The comprehensive operating costs of five typical scenarios with or without energy storage are shown in Table 5:

[0419] Table 5 Comprehensive operating costs for five typical days with and without energy storage configuration

[0420] Comprehensive operating cost (energy storage configuration) Comprehensive operating cost (without considering energy storage configuration) Typical day 1 20338864 21696406 Typical day 2 18817218 20736296 Typical day 3 28330524 28820237 Typical day 4 18899683 21245718 Typical day 5 20408024 20827422

[0421] The comprehensive operating costs of five typical scenarios with and without inertia constraints are shown in Table 6:

[0422] Table 6 Comprehensive operating costs for five typical days with and without inertia constraints

[0423] Comprehensive operating cost (inertia constraint) Comprehensive operating cost (without considering inertia constraints) Typical day 1 20338864 20303532 Typical day 2 18817218 18821074 Typical day 3 28330524 28178937 Typical day 4 18899683 18988899 Typical day 5 20408024 20399496

[0424] The energy storage configuration results with and without inertia constraints are shown in Table 7:

[0425] Table 7 Energy storage configuration results with and without inertia constraints

[0426] Consider inertia constraints Ignore inertia constraints Configure maximum capacity / MWh 759.8 759.8 Configured maximum power / MW 759.8 363.8

[0427] In summary, the solution in this embodiment can fully consider the uncertainty of wind and solar power output, take into account the stability of grid frequency, and realize the complementary operation of energy storage and thermal power and new energy in the grid.

[0428] See also Fig.10 A computer device 400 provided in an embodiment of the present application includes: a processor 410 and a memory 420, wherein the memory 420 stores a computer program executable by the processor 410, and when the computer program is executed by the processor 410, the above method is executed.

[0429] The embodiment of the present application further provides a storage medium 430 on which a computer program is stored. When the computer program is run by the processor 410, the above method is executed.

[0430] Among them, the storage medium 430 can be implemented by any type of volatile or non-volatile storage device or a combination thereof, such as static random access memory (SRAM), electrically erasable programmable read-only memory (EEPROM), erasable programmable read-only memory (EPROM), programmable red-only memory (PROM), read-only memory (ROM), magnetic storage, flash memory, disk or optical disk.

[0431] In the description of the present invention, the terms "first" and "second" are used for descriptive purposes only and should not be understood as indicating or implying relative importance or implicitly indicating the number of the indicated technical features. Therefore, the features defined as "first" and "second" may explicitly or implicitly include one or more of the features. "Multiple" means two or more, unless otherwise clearly and specifically defined.

[0432] In the present invention, unless otherwise clearly specified and limited, the terms "installed", "connected", "connected", "fixed" and the like should be understood in a broad sense, for example, it can be a fixed connection, a detachable connection, or an integral connection; it can be a mechanical connection or an electrical connection; it can be a direct connection or an indirect connection through an intermediate medium, it can be the internal connection of two elements or the interaction relationship between two elements. For ordinary technicians in this field, the specific meanings of the above terms in the present invention can be understood according to specific circumstances.

[0433] In the description of this specification, the description with reference to the terms "one embodiment", "some embodiments", "example", "specific example", or "some examples" etc. means that the specific features, structures, materials or characteristics described in conjunction with the embodiment or example are included in at least one embodiment or example of the present invention. In this specification, the schematic representations of the above terms do not necessarily refer to the same embodiment or example. Moreover, the specific features, structures, materials or characteristics described may be combined in any one or more embodiments or examples in a suitable manner. In addition, those skilled in the art may combine and combine the different embodiments or examples described in this specification and the features of the different embodiments or examples, unless they are contradictory.

[0434] Any process or method description in a flowchart or otherwise described herein may be understood to represent a module, segment or portion of code that includes one or more executable instructions for implementing the steps of a specific logical function or process, and the scope of the preferred embodiments of the present invention includes alternative implementations in which functions may not be performed in the order shown or discussed, including performing functions in a substantially simultaneous manner or in the reverse order depending on the functions involved, which should be understood by those skilled in the art to which the embodiments of the present invention belong.

[0435] The logic and / or steps represented in the flowchart or otherwise described herein, for example, can be considered as an ordered list of executable instructions for implementing logical functions, and can be embodied in any computer-readable medium for use by an instruction execution system, device or apparatus (such as a computer-based system, a system including a processor, or other system that can fetch instructions from an instruction execution system, device or apparatus and execute instructions), or in combination with these instruction execution systems, devices or apparatuses. For the purposes of this specification, "computer-readable medium" can be any device that can contain, store, communicate, propagate or transmit a program for use by an instruction execution system, device or apparatus, or in combination with these instruction execution systems, devices or apparatuses. More specific examples (non-exhaustive list) of computer-readable media include the following: an electrical connection with one or more wires (electronic device), a portable computer disk box (magnetic device), a random access memory (RAM), a read-only memory (ROM), an erasable and programmable read-only memory (EPROM or flash memory), an optical fiber device, and a portable compact disk read-only memory (CDROM). In addition, the computer-readable medium may even be paper or other suitable medium on which the program is printed, since the program may be obtained electronically, for example, by optically scanning the paper or other medium and then editing, interpreting or processing in other suitable ways if necessary, and then stored in a computer memory.

[0436] It should be understood that the various parts of the present invention can be implemented by hardware, software, firmware or a combination thereof. In the above-mentioned embodiments, multiple steps or methods can be implemented by software or firmware stored in a memory and executed by a suitable instruction execution system. For example, if implemented by hardware, as in another embodiment, it can be implemented by any one of the following technologies known in the art or their combination: a discrete logic circuit having a logic gate circuit for implementing a logic function for a data signal, a dedicated integrated circuit having a suitable combination of logic gate circuits, a programmable gate array (PGA), a field programmable gate array (FPGA), etc.

[0437] A person skilled in the art may understand that all or part of the steps in the method for implementing the above-mentioned embodiment may be completed by instructing related hardware through a program, and the program may be stored in a computer-readable storage medium, which, when executed, includes one or a combination of the steps of the method embodiment.

[0438] The storage medium mentioned above may be a read-only memory, a magnetic disk or an optical disk, etc. Although the embodiments of the present invention have been shown and described above, it can be understood that the above embodiments are exemplary and cannot be understood as limiting the present invention. A person of ordinary skill in the art may change, modify, replace and modify the above embodiments within the scope of the present invention.

Claims

1. A method for optimizing energy storage configuration considering inertia constraints, characterized in that: The steps include: Convert the frequency requirements for safe operation of the power system into system inertia constraints; Generate typical wind and solar power output scenarios based on spectrum normalization generative adversarial network (SNGAN), and reduce the generated scenarios through AP-DTW-K-medoids, a center point segmentation clustering algorithm based on affinity propagation and dynamic time warping distance. Set up the coordinated operation mode of pumped storage and energy storage batteries under a high proportion of new energy systems; With the lowest comprehensive operation cost of the power system, the objective function and constraint conditions for the optimal configuration of power grid energy storage in a high-proportion new energy system considering the inertia constraint of the system are established, and a mixed integer linear programming model is constructed; Solving the mixed integer linear programming model to obtain an energy storage optimization configuration plan; The step of converting the frequency requirement for safe operation of the power system into the system inertia constraint comprises the following steps: The system frequency response characteristics are expressed using the rotor motion equation: Where D sys is the damping coefficient during the system frequency response process; f N is the rated frequency of the system; Δf t is the real-time frequency change of the system; ΔP Σ is the frequency modulation power of all scheduling resources; ΔP loss The power shortage of the generator; Calculate the maximum frequency change rate of the system: In the formula, Rocof M Indicates the maximum frequency change rate of the system; f N Indicates the standard value of system frequency; H sys is the real-time inertia of the system; Calculate the minimum inertia based on the minimum frequency change rate constraint: Combined with the frequency response mode of each dispatching resource in the high-proportion renewable energy power system and the rotor motion equation, the maximum system frequency deviation level after the fault occurs is calculated, which needs to meet the system transient frequency deviation extreme value constraints, including: Δf NF ≤Δf max ; Where Δf NF is the frequency extreme value deviation; Δf max It is the maximum frequency variation allowed by the system.

2. The energy storage optimization configuration method considering inertia constraints according to claim 1 is characterized in that: After the frequency requirements for safe operation of the power system are converted into system inertia constraints, the system frequency response process is simulated using piecewise linearization technology, including the following steps: The rotor motion equation from 0 to t db Integrate and solve to obtain: Where: f db and t db They are primary frequency modulation control dead zone and action time respectively; The system primary frequency modulation rate is used to represent the time from the action of the stabilization control measures to the frequency reaching the transient extreme value. The system frequency reaching the extreme value time is expressed as: The rotor motion equation from 0 to t NF Integrate and solve using the linearization method to obtain: Considering the above three formulas comprehensively, we can get: The minimum inertia requirement of the system is: The system real-time inertia includes: In the formula, Gen_n and ps_n represent the total number of conventional thermal power units and the total number of pumped storage units respectively; P Gen，i , H Gen，i Respectively represent the rated power and inertia time constant of the synchronous unit; P S，i , H PS，i Respectively represent the rated power and inertia time constant of the pumped storage unit; P ES , H ES They represent the rated power and inertia time constant of the energy storage device respectively.

3. The energy storage optimization configuration method considering inertia constraints according to claim 1 is characterized in that: The spectrum normalization generative adversarial network (SNGAN) method generates typical wind and solar power output scenarios, including: Modify the traditional generative adversarial network GAN: Let the input be x, where the discriminator expression is f(x) = W x , then for a randomly given x, we have: In the formula, δ(W x ) is the function W x The spectral norm of Find the weights and biases of each layer in the discriminator: Where W l is the weight of each layer in the discriminator, a l is the bias of each layer in the discriminator, l is the lth layer of the discriminator; According to the above formula, we can get: Where D is the discriminator output matrix, and δ(A)δ(B)≥δ(AB), we can get: Normalizing the weights, we can get: For any X: The global 1-Lipschitz constraint can be implemented by calculating the maximum singular value of the weight matrix of each layer.

4. The energy storage optimization configuration method considering inertia constraints according to claim 3 is characterized in that: The generated scene is reduced by using the center point segmentation clustering algorithm AP-DTW-K-medoids based on affinity propagation and dynamic time warping distance, including the following steps: Input initial scene matrix N = {n1; n2; n3…n d }, given the number of scenes m that need to be clustered, select the reference type of the affinity propagation AP algorithm, and give the threshold η∈[0,1]; The similarity of the scenes is calculated by the given initial scene matrix, reference type and threshold parameters. The calculation formula is: x(i,c)=-d 2 (n i ,n c )=-||n i -n c || 2 Where d is the Euclidean distance, i, c∈d; Calculate the mutual attraction of each sample scene, including: y(i,c)=x(i,c)-max{a(i,c′)+x(i,c′)}, c≠c′; Calculate the attribution degree of each sample scene, including: Update the belongingness and attractiveness of sample scenarios, including: Where λ is the damping coefficient; Iterate the loop, the cut-off criterion is that the cluster center no longer changes or exceeds the specified number of iterations, and then output the final cluster center V. At this time, the number of elements in the cluster center is k, and k satisfies k>m; Calculate the dynamic time warping (DTW) distances between the output cluster centers one by one to form a similarity matrix B; The minimum spanning tree is established based on the obtained similarity matrix B. By using the weight descending principle, m-1 branches with larger weights are removed, and finally m data matrices are obtained as the cluster centers required by the center point segmentation clustering K-medoids algorithm. The clustering evaluation index is used to evaluate the data matrix of the m scenes obtained. If the preset index is met, the obtained scene data is used as the final typical scenery scene.

5. The energy storage optimization configuration method considering inertia constraints according to claim 1 is characterized in that: The coordinated operation mode of pumped storage and energy storage batteries under a high proportion of new energy systems includes: When the gap between supply and demand of the power system is large, it is necessary to prioritize the safety of system operation. Under the premise of meeting system constraints, energy storage batteries will respond first, and pumped storage will take over the response. When the gap between supply and demand of the system is small, under the premise of meeting system constraints, energy storage batteries, pumped storage and thermal power units will respond to power shortages or surpluses in accordance with the principle of optimizing the economic benefits of the overall operation of the system.

6. The energy storage optimization configuration method considering inertia constraints according to claim 1 is characterized in that: The objective function includes: min F=C1+C2+C3+C4; Where, F is the comprehensive operating cost of the power system; f i，s represents the fuel cost of the i-th conventional thermal power unit on a typical day s; P Gi，s (t) represents the output power of the i-th conventional thermal power unit in the period t on a typical day s; S i，t，s represents the start-up and shutdown cost of the i-th conventional thermal power unit on a typical day s, A i，t，s represents the deep peak load regulation cost of the i-th conventional thermal power unit on a typical day s, u i，t，s Indicates the running status of the unit, 1 indicates running, 0 indicates shutdown; C cur It represents the unit cost of wind and solar power abandonment; They represent the predicted output of wind power at time t on a typical day s; E bess Represents the capacity of energy storage configuration, P bess Indicates the maximum power of the energy storage configuration, ΔH bess Represents the virtual inertia of the additional energy storage; α s represents the typical day probability.

7. The energy storage optimization configuration method considering inertia constraints according to claim 6 is characterized in that: The constraints include: System minimum inertia constraint: In the formula, It represents the current inertia of the system during the period t under the typical day s; Node power balance constraints: In the formula, is the active power between node b and node b′ on a typical day s; is the load size of node b at time t on a typical day s; DC power flow constraints: In the formula, B bb′ is the admittance between node b and node b′; θ b，t，s and θ b′，t，s are the phase angle differences between node b and node b′ during period t on a typical day s; and are the lower and upper limits of the active power of the line connecting nodes b and b′ during period t on a typical day s, respectively; Phase Angle Constraint: In the formula, and are the lower and upper limits of the phase angle of node b respectively; Output upper and lower limits of thermal power units: Where P Gi，max , P Gi，min They represent the upper and lower limits of the output of conventional thermal power i respectively; They represent the normal peak-shaving and deep peak-shaving outputs of the i-th conventional thermal power unit at time t on a typical day s; and They are respectively the normal peak-shaving and deep peak-shaving output states of the i-th conventional thermal power unit at time t on a typical day s. If the unit is in the startup state, it is 1, otherwise it is in the shutdown state, it is 0; Other constraints: Minimum start and stop time constraints for thermal power units, climbing constraints for thermal power units, wind and solar output constraints, pumped storage unit constraints and energy storage battery constraints.

8. The energy storage optimization configuration method considering inertia constraints according to claim 1 is characterized in that: The mixed integer linear programming model is solved to obtain an energy storage optimization configuration scheme, including: The mixed integer linear programming model is solved by calling the Gurobi solver using Yalmip to obtain the maximum capacity, power and virtual inertia of each energy storage configuration in a high-proportion new energy grid.

9. An energy storage optimization configuration device considering inertia constraints, characterized in that: Use of the method according to any one of claims 1 to 8, comprising: Inertia constraint unit, used to convert the frequency requirements for safe operation of the power system into system inertia constraints; The scenario generation unit is used to generate typical wind and solar power output scenarios based on the spectrum normalization generative adversarial network (SNGAN) method, and reduce the generated scenarios through the center point segmentation clustering algorithm AP-DTW-K-medoids based on affinity propagation and dynamic time warping distance; A coordinated operation unit is used to set the coordinated operation mode of pumped storage and energy storage batteries under a high proportion of new energy systems; A modeling unit, for establishing an objective function and constraint conditions for optimizing the configuration of power grid energy storage of a high-proportion new energy system taking into account the inertia constraint of the system with the lowest comprehensive operation cost of the power system, and constructing a mixed integer linear programming model; A solving unit is used to solve the mixed integer linear programming model to obtain an energy storage optimization configuration plan.

10. A computer device 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 method according to any one of claims 1 to 8 is implemented.

11. A storage medium having a computer program stored thereon, characterized in that: When the computer program is executed by a processor, the method according to any one of claims 1 to 8 is implemented.

Images