A power grid energy storage configuration optimization method and system based on holomorphic embedding method

Abstract

The present application belongs to the field of power distribution energy storage optimization, and particularly relates to a power grid energy storage configuration optimization method and system based on full pure embedding method, aiming to solve the problem of how to improve the calculation efficiency and accuracy under a large number of scenarios in energy storage configuration planning. The method of the present application comprises: obtaining historical data and preprocessing; constructing an energy storage configuration double-layer planning model; generating a large number of scenarios by Monte Carlo simulation based on the preprocessed wind, light and load historical data according to the initial configuration scheme; clustering the large number of scenarios to determine typical scenarios; performing HEM power flow calculation on the typical scenarios; modifying the complex variables of each node based on the difference between the typical scenarios and their corresponding large number of similar scenarios, and then substituting into the power flow analytical expression to realize fast power flow calculation; performing multi-level optimization screening on each configuration scheme to determine the energy storage configuration scheme. The present application significantly improves the calculation efficiency and accuracy under a large number of scenarios, and ensures the reliability of the configuration scheme.

Description

Technical Field

[0001] This invention belongs to the field of power distribution energy storage optimization, and specifically relates to a power grid energy storage configuration optimization method, system and equipment based on the fully embedded method. Background Technology

[0002] Energy storage technology, as a key regulation tool in the power system, possesses unique advantages such as rapid response and bidirectional regulation, playing an irreplaceable role in addressing the power supply-demand imbalance caused by the integration of new energy sources into the grid. In particular, the large-scale integration of renewable energy sources into the grid has made the demand for flexible regulation capabilities in the power system increasingly urgent. However, the high investment cost of energy storage systems remains a major bottleneck restricting their large-scale application. Therefore, how to achieve economical allocation of energy storage capacity while ensuring system reliability has become a critical issue that urgently needs to be addressed.

[0003] The complexity of energy storage configuration planning is mainly reflected in three aspects: First, the randomness and volatility of renewable energy lead to diverse operating scenarios for the power system, making it difficult for traditional deterministic planning methods to effectively address these challenges. Second, energy storage systems exhibit significant temporal coupling characteristics across different time scales, which dramatically increases the scale of the optimization problem. Finally, the introduction of electricity market mechanisms further increases the complexity of the problem, requiring the simultaneous consideration of uncertainties in multiple dimensions, such as price fluctuations and market clearing. These characteristics make energy storage configuration planning a high-dimensional, strongly coupled, and complex optimization problem.

[0004] Currently, the mainstream solutions for energy storage configuration planning can be divided into two categories: exact algorithms based on mathematical programming and intelligent algorithms based on heuristics. While mathematical programming methods can guarantee optimal solutions, they often require significant simplification of the model, which can lead to substantial deviations between the planning results and actual operating conditions. For example, linear programming methods require linearizing the energy storage charging and discharging characteristics, while mixed-integer programming often ignores the dynamic characteristics of energy storage. On the other hand, heuristic methods such as genetic algorithms and particle swarm optimization can handle complex nonlinear and nonconvex problems well, but their solution efficiency drops significantly when dealing with large-scale scenario sets, making it difficult to meet the needs of practical engineering applications.

[0005] In particular, with the continuous expansion of the power system and the sustained increase in renewable energy penetration, the number of typical scenarios that need to be considered in energy storage configuration planning is growing exponentially. Traditional solution methods, when dealing with such a massive set of scenarios, either fail to converge due to excessive computational burden or suffer from inaccuracy due to oversimplification. Therefore, how to efficiently handle massive scenarios while ensuring the reliability of the planning scheme has become a key focus and challenge in current research. Summary of the Invention

[0006] To address the aforementioned problems in existing technologies, namely, how to improve the computational efficiency and accuracy in massive scenarios of energy storage configuration planning, the first aspect of this invention proposes a grid energy storage configuration optimization method based on a fully embedded method, used to solve for rapid calculations in massive scenarios of energy storage configuration planning. The method includes the following steps:

[0007] Acquire historical data on wind, solar and lotus flowers, and perform data preprocessing on the historical data on wind, solar and lotus flowers;

[0008] Construct a two-layer planning model for energy storage configuration, based on The algorithm solves the model:

[0009] Step S100: Initialize the population and randomly generate M configuration schemes, where M is the population size for the genetic algorithm;

[0010] Step S200: Based on the capacity ratio of each node in each configuration scheme, and combined with the preprocessed historical wind-solar-load data, generate massive scenarios corresponding to each configuration scheme and perform parallel power flow calculations.

[0011] The power flow calculation method for massive scenarios under each configuration scheme is as follows:

[0012] Step S210: Perform cluster analysis on the massive number of scenes to obtain the optimal number of clusters, and obtain the clustered typical scenes based on the optimal number of clusters;

[0013] Step S220: Perform HEM power flow calculation on the typical scenarios to obtain the power flow analytical expression corresponding to each typical scenario. The power flow analytical expression includes the power series coefficient matrix of state variables and the complex variables of each node.

[0014] Step S230: Based on the scenario difference analysis, the complex variable correction mechanism corrects the complex variables of each node in the massive number of similar scenarios corresponding to the typical scenario. The corrected complex variables are substituted into the power flow analysis expression to obtain the power flow calculation result.

[0015] Step S300: Based on the power flow analytical expression and power flow calculation results, perform multi-level optimization and screening of each configuration scheme, and output the configuration scheme that meets the screening conditions as a reasonable energy storage configuration scheme.

[0016] In some preferred embodiments, the energy storage configuration two-layer planning model includes an upper-layer planning model and a lower-layer operation model. The upper-layer planning model takes minimizing the total system cost as the objective function, while the lower-layer operation model takes minimizing the system operating cost as the objective function.

[0017] In some preferred embodiments, the typical scenarios after clustering are obtained by the following method:

[0018] Clustering algorithms are used to cluster massive amounts of data, and the cluster centers that minimize the objective function are output. , k To ensure both the adaptability of HEM and the overall computational efficiency of the number of clusters, the scenario corresponding to each cluster center is taken as a typical scenario; the clustering includes the K-means clustering algorithm.

[0019] In some preferred embodiments, the HEM power flow calculation is performed as follows:

[0020] Step S221: Construct the holomorphic embedded form of the power balance equation and recursively solve for the power series coefficients of the state variables in each typical scenario;

[0021] Step S222: Calculate PA based on the power series coefficients of the state variables, set the complex variable s of the node to 1, and obtain an approximate solution to the power flow equation;

[0022] Step S223: Determine whether the approximate solution satisfies the termination condition;

[0023] If the condition is met, the approximate solution is the power flow steady-state solution; if not, the number of power series terms and the PA order are increased, and the process returns to step S222 to perform the next order of recursion until the termination condition is met.

[0024] In some preferred embodiments, the termination condition is the convergence accuracy:

[0025] When the power flow calculation result reaches the aforementioned convergence accuracy, a power series coefficient model is constructed based on the state variables. N×n Power series coefficient matrix of state variables ;in N The number of system nodes. n The number of iterations required for HEM to converge;

[0026] Otherwise, if the number of recursions n Once the maximum number of attempts is reached, a filtering signal will be sent to filter out the typical scenario and its corresponding similar scenarios.

[0027] In some preferred embodiments, the fast power flow calculation results are obtained by:

[0028] Step S231, for the first i A typical scenario For the corresponding massive number of similar scenarios, change the value of the complex variable s of the corresponding node, and perform power flow calculation on the massive number of similar scenarios:

[0029] ;

[0030] ;

[0031] In the formula, Indicates the first i Nodes in similar scenarios j The voltage holomorphic function; Represents the nodes in the i-th scenario j A reactive holomorphic function; For nodes j Embedded complex variables;

[0032] Step S232: Correct the complex variables of the massive number of similar scenes, using typical scenes. The complex variable takes the value 1 as the baseline. It varies with the generator output and load of each node; the specific value of the single-input complex variable of each node corresponds proportionally to the net load of each node in the current scenario and the net load of each node in a typical scenario:

[0033] ;

[0034] In the formula, For nodes in similar scenarios j The complex variable to be determined; Nodes in typical scenarios j A known complex variable with a value of 1; , Nodes in similar scenarios j The load and generator power; , Nodes in typical scenarios j The load and generator power;

[0035] Step S233: Substitute the corrected complex variables of the massive number of similar scenes into the corresponding typical clustering scenarios. The power flow calculation results are obtained from the power series coefficient matrix of the state variables.

[0036] In some preferred embodiments, the multi-level optimization screening includes scenario optimization screening and solution optimization screening:

[0037] The method for optimizing and filtering the scenarios is as follows:

[0038] For each scenario under each configuration scheme, a penalty term is obtained by constraining the node voltage, and this term is added to the target value of the upper-level planning model.

[0039] The optimization and screening method for the proposed scheme is as follows:

[0040] For each configuration scheme after scenario optimization and screening, if the penalty term corresponding to each configuration scheme is less than the penalty term threshold and the number of remaining scenarios is greater than the set scenario number threshold, then the configuration scheme is output as a reasonable energy storage configuration scheme; otherwise, based on sensitivity analysis, key nodes are determined, and the energy storage capacity configured at the key nodes is adjusted to optimize the configuration schemes that do not meet the requirements. After optimization, steps S200-S300 are executed again to delete unreasonable energy storage configuration schemes; the penalty term of the configuration scheme is the sum of the penalty terms corresponding to the remaining scenarios under the configuration scheme.

[0041] In some preferred embodiments, the node voltage is constrained by means of:

[0042] Step S301: Based on the power flow calculation results, determine the voltage limit of each node, identify and record the voltage limit-over-limit nodes, and add penalty items to the scenarios corresponding to the voltage limit-over-limit nodes;

[0043] Step S302: Adjust the complex variable value of the voltage over-limit node until the target voltage returns to the allowable range, thereby correcting the voltage over-limit node;

[0044] Step S303: After all voltage over-limit nodes have been corrected, return to step S230 based on the adjusted complex variable values ​​of each node, recalculate and update the power flow calculation results;

[0045] Step S304: Perform another voltage limit judgment on the updated power flow calculation results. If there are still voltage limit exceeding nodes in the updated scenario, add the set large value penalty item to the scenario or filter them out directly.

[0046] In some preferred embodiments, the method for determining key nodes based on sensitivity analysis is as follows:

[0047] Differentiating the power flow analytical expression determines the relationship between nodes. p The new energy storage node with the greatest voltage impact q As a key node.

[0048] A second aspect of the present invention proposes a grid energy storage configuration optimization system based on a fully embedded method, the system comprising:

[0049] The data processing module is configured to acquire historical wind, solar and load data and perform data preprocessing on the historical wind, solar and load data.

[0050] The model building module is configured to build a two-layer planning model for energy storage configuration.

[0051] Energy storage configuration module, used to randomly generate configuration schemes;

[0052] The scene generation module is configured to generate massive scenes from preprocessed historical wind, solar and load data through Monte Carlo simulation and perform power flow calculations, cluster the generated massive scenes, and generate typical scenes after clustering.

[0053] The power flow calculation module is configured to perform power flow calculations on various typical scenarios to obtain the power series coefficient matrix of state variables and the power flow analytical expression. Based on the complex variable correction mechanism of scenario difference analysis, the complex variables of each node in the massive number of similar scenarios corresponding to the typical scenario are corrected. The corrected complex variables are substituted into the power flow analytical expression to obtain the power flow calculation results.

[0054] The optimization and filtering module is configured to perform multi-level optimization and filtering of each configuration scheme based on the power flow analysis expression and power flow calculation results, and output a reasonable energy storage configuration scheme.

[0055] The beneficial effects of this invention are:

[0056] 1) This invention constructs a two-layer planning model for energy storage configuration, uses clustering to extract typical scenarios from massive scenarios, solves typical scenarios based on a fully embedded power flow method, and derives power flow analytical expressions covering multiple scenarios to achieve rapid power flow calculation for massive scenarios; it significantly improves the computational efficiency and accuracy under massive scenarios, and further ensures the reliability and economy of the configuration scheme, providing a new technical path for energy storage planning under large-scale renewable energy grid connection conditions;

[0057] 2) The K-means clustering algorithm was used to achieve scientific dimensionality reduction of massive running scenarios and extract typical scenarios after clustering. This not only preserved the typical features of the scenarios, but also significantly reduced the scale of subsequent optimization calculations and improved planning efficiency. At the same time, a validity verification mechanism for typical scenarios was established to ensure the reliability of the clustering results.

[0058] 3) The HEM method is introduced into the power flow calculation stage to perform non-iterative fast power flow calculation; by establishing the analytical relationship between state variables and complex variables, fast power flow calculation is performed on massive scenarios, realizing rapid evaluation of system state, avoiding the computational burden of repeated power flow calculation in traditional methods, and significantly improving computational efficiency, especially when dealing with large-scale scenarios.

[0059] 4) The genetic algorithm is innovatively combined with the HEM method to establish a complete voltage amplitude constraint and correction system, ensuring the safety and reliability of the final output scheme in actual operation; the voltage amplitude is constrained and corrected by adjusting complex variables; and the power flow analytical expression is analyzed by methods such as differentiation to accurately identify the key nodes in the system with the most significant correction effect on power flow over-limit, providing a theoretical basis for the optimization and adjustment of energy storage configuration schemes. Attached Figure Description

[0060] Other features, objects, and advantages of this application will become more apparent from the following detailed description of non-limiting embodiments with reference to the accompanying drawings:

[0061] Figure 1 This is a flowchart of the grid energy storage configuration optimization method based on the fully embedded method of the present invention;

[0062] Figure 2 This is a framework diagram of the two-layer planning model for energy storage configuration in an embodiment of the present invention;

[0063] Figure 3 This is a flowchart for solving the planning model based on NSGA-II;

[0064] Figure 4 This is a flowchart of the solution method after incorporating the HEM optimization method in this embodiment of the invention. Detailed Implementation

[0065] The present application will now be described in further detail with reference to the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are for illustrative purposes only and are not intended to limit the invention. Furthermore, it should be noted that, for ease of description, only the parts relevant to the invention are shown in the accompanying drawings.

[0066] It should be noted that, unless otherwise specified, the embodiments and features described in this application can be combined with each other. This application will now be described in detail with reference to the accompanying drawings and embodiments.

[0067] This invention constructs a two-layer planning model for energy storage configuration, combines historical wind and solar load data with scheme configuration results, and uses the Monte Carlo method to generate massive scenarios. These massive scenarios are then clustered to extract typical scenarios. The Heteromorphic Embedding Method (HEM) is introduced into the planning model for non-iterative fast power flow calculation. By analyzing the differences between typical scenarios and their corresponding massive number of similar scenarios, the complex variables of each node in the system are precisely corrected, achieving fast power flow calculation for massive scenarios. This allows for efficient power flow solutions when dealing with massive scenarios, improving the efficiency of power system planning and ensuring the safety and reliability of the final scheme.

[0068] To more clearly explain the grid energy storage configuration optimization method based on the fully embedded method of this invention, the following will be combined with... Figure 1 The steps in the embodiments of the present invention will be described in detail below.

[0069] A grid energy storage configuration optimization method based on the fully embedded method according to the first embodiment of the present invention includes the following steps:

[0070] Step 1: Obtain historical data on wind, solar and lotus flowers, and perform data preprocessing on the historical data.

[0071] Preferably, the data preprocessing includes eliminating outliers and filling in missing values.

[0072] Step 2: Construct a two-layer planning model for energy storage configuration and solve the model based on the NSGA-II algorithm;

[0073] The energy storage configuration two-layer planning model includes an upper-layer planning model and a lower-layer operation model. The upper-layer planning model aims to minimize the total system cost, while the lower-layer operation model aims to minimize the system operating cost. Simultaneously, it clarifies the configuration constraints such as energy storage capacity and power in the upper layer of the planning model, as well as the constraints such as power balance and energy storage operation in the lower layer, and determines the transmission relationship of key decision variables such as energy storage configuration schemes and system operating status between the upper and lower layers.

[0074] Preferably, in this embodiment, the energy storage configuration two-layer planning model takes minimizing the sum of the investment cost of new energy storage equipment, the system operation cost under a unified time scale, and the system stable operation cost as the optimization objective. The layer model is a sub-problem of system operation simulation under each time series scenario, with the goal of minimizing the sum of comprehensive operation costs. The constraints include power flow constraints and steady-state operation simulation constraints.

[0075] 1) Construct a two-layer planning model for energy storage configuration;

[0076] Based on the characteristics of the power grid planning model, the energy storage capacity C of each node is taken as the decision variable, and considering the relevant constraints of the installed capacity, the bi-level programming model can be expressed as:

[0077] (1)

[0078] In the formula , and These are constraints between upper and lower levels; , This refers to the annualized investment cost and annual operating cost; The parameters are uncertain. The upper-level constraints are investment constraints, including energy storage investment constraints and the number of installations, while the lower-level constraints are operational simulation constraints, including power balance, current constraints, and voltage constraints. As shown in equation (1), this model is a unidirectional two-layer model, meaning the upper-level model transmits state variables to the lower-level model, and the lower-level variable decisions and influences the upper-level model are reflected by adding operating costs to the upper-level constraints. The framework diagram of the two-layer planning model for energy storage configuration is as follows: Figure 2 As shown.

[0079] 2) Objective function

[0080] The energy storage configuration planning model aims to optimize the overall investment in new equipment and the system operating cost. As shown in Equation (1), the model adopts annualized investment and annual operating cost. The annualized investment in new energy storage equipment can be modeled by considering the annualized investment over the entire service life of the equipment, as shown in Equation (2).

[0081] (2)

[0082] In the formula: The discount rate during the planning period; For the lifespan of energy storage equipment; This refers to the set of locations for energy storage devices. For energy storage equipment in site selection The unit capacity investment cost; For energy storage equipment in site selection The newly installed capacity.

[0083] The overall annual operating cost of the system mainly includes the operating cost of thermal power units and environmental protection, load shedding penalties and curtailment penalties, which correspond to the operating costs of thermal power units, loads and wind and solar power units respectively. In this invention, only the energy storage configuration is considered, so this item can be ignored.

[0084] 3) Planning Level Constraints

[0085] The main planning constraints are the total capacity of energy storage equipment construction, i.e., the total installed capacity of newly built equipment; as shown in equation (3):

[0086] (3)

[0087] In the formula: The total capacity to be built for energy storage is [to be determined]. This represents the maximum permissible error in the installed capacity of energy storage equipment.

[0088] 4) Runtime layer constraints

[0089] Operational constraints mainly include power generation-load balance constraints, reserve demand constraints, line transmission capacity constraints, and equipment characteristic constraints, including output and flexibility constraints of conventional units (such as ramping constraints and minimum start-up and shutdown time constraints), output constraints of wind and solar units, and power constraints of energy storage devices.

[0090] a) Power balance constraint:

[0091] (4)

[0092] In the formula: For wind and solar turbine units At the node Those who have made meritorious contributions; , These are the time intervals for energy storage devices. At the node The charging and discharging power; For load time At the node The active power.

[0093] b) Power flow balance constraint:

[0094] This model adopts AC power flow. The safe and stable operation of the system's network line requires satisfying AC power flow balance, with the following constraints:

[0095] (5)

[0096] (6)

[0097] In the formula: and These represent the active power and reactive power output by the generator set at node i, respectively. It is the active load at node i. It is the reactive load at node i; The phase angle difference between node i and node j; and Let be the conductance and susceptance of the element in the i-th row and j-th column of the admittance matrix, respectively. This represents the power output of the energy storage device at node i in the discharge state.

[0098] c) Node voltage magnitude and phase angle constraints

[0099] After obtaining the power flow calculation results in b), the voltages of each node in the system can be constrained under the conditions of simultaneous parallel operation and stable operation of the system:

[0100] (7)

[0101] (8)

[0102] In the formula: Let be the voltage magnitude at node i. and Its upper and lower limits; Let i be the voltage phase angle at node i. Let J be the voltage phase angle at node j. This is the upper limit threshold of the difference between the two.

[0103] d) Line transmission capacity constraints

[0104] Based on the power flow calculation results in b), the actual power flow of the line should not exceed the transmission capacity of the line in the network structure; as shown in equation (9):

[0105] (9)

[0106] In the formula: For the first Timetable for each line The transmission capacity; For the first The maximum forward transmission capacity of each line; For the first The upper limit of the reverse transmission capacity of each line.

[0107] e) Operational characteristic constraints

[0108] The model includes constraints on the charging and discharging power of energy storage devices, state of charge constraints, daily clearing constraints, and energy balance constraints.

[0109] (10)

[0110] (11)

[0111] (12)

[0112] (13)

[0113] (14)

[0114] In the formula: This refers to the charging power. , These are the lower and upper limits of the charging power, respectively. This refers to the discharge power. , These represent the lower and upper limits of the discharge power, respectively. The upper limit of the charge and discharge power is generally determined by physical factors, while the lower limit is usually set to 0. Sometimes, a threshold value can be set to prevent small power fluctuations. , The variables are 0 and 1, representing the battery's discharge and charging states, respectively. State of charge of the energy storage device; , These are the maximum and minimum energy storage ratios of the energy storage device, respectively. , Each day and State of charge of energy storage; For nodes The rated capacity of the energy storage; The time step for energy storage charging and discharging.

[0115] In the upper-level constraints of the two-layer planning model for energy storage configuration, power flow constraints can not only screen the safety and stability of the scheme, but also ensure the operational scheduling flexibility of the scheme. The aforementioned... The algorithm solves the model using steps S100-S300, each of which is described in detail below:

[0116] Step S100: Initialize the population and randomly generate M configuration schemes, where M is the population size for the genetic algorithm.

[0117] Step S200: Based on the capacity ratio of each node in each configuration scheme, generate massive scenarios and perform power flow calculations on the preprocessed wind-solar-load historical data through Monte Carlo simulation.

[0118] In this embodiment, the 8760h historical data of wind, solar and load are allocated and organized according to the capacity ratio of each node in different configuration schemes. Time is used as the unified interface input. Based on the law of large numbers, Monte Carlo sampling method is used to generate massive scenes, and the 8760h dataset of wind, solar and load can be obtained.

[0119] Preferably, the power flow calculation method is as follows:

[0120] Step S210: Perform cluster analysis on the massive number of scenes to obtain the optimal number of clusters, obtain typical scenes after clustering based on the optimal number of clusters, and verify the typical scenes;

[0121] The sheer volume of scenarios imposes a heavy workload on practical applications, leading to low efficiency and making them unsuitable for power system planning, operation, and control. Therefore, scenario reduction is necessary. In this embodiment, the K-means clustering algorithm is used to cluster the generated massive number of scenarios. The optimal number of clusters is determined based on the adaptability and efficiency of HEM power flow calculations, resulting in representative typical scenarios.

[0122] Furthermore, clustering algorithms are used to cluster massive amounts of data, outputting the cluster centers that minimize the objective function for each type of scene. k To ensure the optimal number of clusters that simultaneously guarantees HEM adaptability and overall computational efficiency, the scenarios corresponding to each cluster center are taken as typical scenarios; the clustering includes the K-means clustering algorithm.

[0123] In this field, as an alternative approach, the K-means clustering algorithm performs the following calculations:

[0124] 1) Random selection k One initial cluster center;

[0125] 2) Based on the principle of minimum distance, assign all individuals to be clustered to the cluster with the closest distance to them;

[0126] The key to clustering methods is how to measure the similarity between each individual in a set. This is addressed by using a distance function, which is used to measure the similarity of a dataset to be clustered. X = {x1, x2, …, xn} In this context, the similarity between any two individuals is a function; given two individuals, two... P dimensional vector and To illustrate, according to the definition of Euclidean distance:

[0127] (15)

[0128] This formula is used to calculate the distance from each sample to all cluster centers, and the individual is assigned to the cluster with the closest distance. C i inside;

[0129] 3) Resolve for the centroid of each cluster, and use it as the new cluster center for the next iteration;

[0130] 4) Repeat steps 2) and 3) to minimize the objective function;

[0131] The objective function is the minimum variance function, which is the sum of the squared Euclidean distances between all individuals within a cluster and the cluster center. The function is defined as follows:

[0132] (16)

[0133] In the formula, E is the sum of squared errors between all individuals in the dataset to be clustered and their corresponding cluster centers. x It belongs to the cluster in the set to be clustered C i individuals, c i It is a cluster C i Cluster centers.

[0134] 5) Output the cluster centers that minimize the objective function for each type of cluster. The corresponding optimization scheme generates k A typical scenario. Then in the... i Cluster C i In the middle, cluster center c i In its typical scenario, this cluster contains, in addition to c i Other than these are typical scenarios. c i Corresponding similar scenarios.

[0135] Step S220: Perform HEM power flow calculations on each typical scenario to obtain the power flow analytical expression corresponding to the typical scenario. The power flow analytical expression includes the power series coefficient matrix of state variables and the complex variables of each node.

[0136] Preferably, the HEM power flow calculation method is as follows:

[0137] Step S221: Construct the power balance equation in fully embedded form and recursively solve for the power series coefficients of the state variables in each typical scenario;

[0138] Step S222: Calculate PA based on the power series coefficients of the state variables, and let the complex variables of each node... s Setting it to 1 yields an approximate solution to the power flow equations;

[0139] Step S223: Determine whether the approximate solution satisfies the termination condition;

[0140] If satisfied, the approximate solution is the power flow steady-state solution;

[0141] If the condition is not met, the number of terms in the power series and the order of PA are increased, and the process returns to step S221 to perform the next order of recursion until the termination condition is met.

[0142] Preferably, when applying the HEM method for power flow calculation in the energy storage configuration planning model used in this invention, the convergence accuracy of the HEM power flow calculation needs to be preset. Based on the above HEM power flow calculation steps, the obtained... k For a typical scenario, power flow calculations are performed, and the coefficients of the power series of higher-order state variables need to be recursively solved before the convergence accuracy requirements are met.

[0143] The termination condition is whether the convergence accuracy is met:

[0144] When the power flow calculation results reach the aforementioned convergence accuracy, an N×n order state variable power series coefficient matrix is ​​constructed based on the solved state variable power series coefficients. ; (17)

[0145] in, u , v Here, PQ nodes and PV nodes represent the number of system nodes, respectively, and N is the total number of system nodes. n The number of iterations required for HEM to converge;

[0146] Otherwise, if determining the number of recursions n Once the preset maximum number of attempts is reached, a filtering signal will be issued to filter out the typical scenario and its corresponding similar scenarios.

[0147] Step S230: Based on the scenario difference analysis, the complex variable correction mechanism corrects the complex variables of each node in the massive number of similar scenarios corresponding to the typical scenario. The corrected complex variables are substituted into the power flow analysis expression to obtain the power flow calculation result, thereby realizing the rapid power flow calculation of massive scenarios.

[0148] The power flow calculation results are obtained as follows:

[0149] Step S231, for the first i A typical scenario The corresponding massive number of similar scenarios change the complex variables of the corresponding nodes. s The values ​​are selected to perform power flow calculations on the massive number of similar scenarios;

[0150] by As a benchmark, when the complex variables of each node s When both are 1, the corresponding The steady-state solution of the current trend; when for When performing power flow calculations on a large number of similar scenarios, the complex variables of the corresponding nodes can be changed. s The values ​​are selected to obtain the power flow calculation results for a large number of scenarios:

[0151] The expression for calculating the state variables of power flow in massive scenarios using HEM is as follows:

[0152] (18)

[0153] (19)

[0154] In the formula, Indicates the first i The voltage holomorphic function of node j in a typical scenario; Represents a node in the i-th typical scenario j A reactive holomorphic function; For nodes j Embedded complex variables.

[0155] Step S232: Correct the complex variables of the massive number of similar scenes, using typical scenes. The complex variable takes the value 1 as the baseline. It varies with the generator output and load at each node. For example, when nodes are in a similar scenario... j When the load increases compared to a typical scenario, the complex variables corresponding to this node... It should also be increased The specific values ​​of the single-input complex variables at each node correspond proportionally to the net load of each node in the current scenario and the net load of each node in a typical scenario.

[0156] (20)

[0157] In the formula, For nodes in similar scenarios j The complex variable to be determined; Nodes in typical scenarios j A known complex variable with a value of 1; , Nodes in similar scenarios j The load and generator power, where the generator power is the system load at node j The total power generation capacity includes the power generation capacity of thermal power, wind power, and photovoltaic units; , Nodes in typical scenarios j The load and generator power, where the generator power is the system load at node j The total power generation capacity includes the power generation capacity of thermal power, wind power and photovoltaic units.

[0158] Step S233: Substitute the corrected complex variables of the massive number of similar scenes into the corresponding typical clustering scenarios. The power flow calculation results are obtained from the power series coefficient matrix of the state variables.

[0159] Compared to traditional methods, the HEM power flow calculation method of this invention can avoid repeated iterative calculations when processing massive similar scenarios. By converting the differences between massive similar scenarios and typical scenarios into differences in the values ​​of complex variables, the complex traditional power flow calculation is transformed into simple algebraic operations, which greatly improves the solution efficiency and realizes fast power flow calculation for massive similar scenarios based on HEM.

[0160] Furthermore, the fast power flow calculation results obtained through the above steps for a large number of similar scenarios show some deviation compared to the power flow calculation results of traditional iterative methods. This deviation is caused by the approximation of the transformation in equation (20). When the selected general scenario has a high similarity to the typical scenario, the deviation of the power flow calculation results is small; when the selected general scenario has a low similarity to the typical scenario, the deviation of the power flow calculation results is large. Therefore, HEM fast power flow calculation itself has a certain scope of application, but it can be improved by increasing the number of clusters. k This enhances the similarity between general and typical scenarios within each scenario category, thereby addressing the issue of insufficient applicability of the method and ensuring the authenticity and reliability of the output results.

[0161] Step S300: Based on the power flow analysis expression and power flow calculation results, perform multi-level optimization and screening on each configuration scheme, and output the configuration scheme that meets the screening conditions as a reasonable energy storage configuration scheme; the multi-level optimization and screening includes scenario optimization screening and scheme optimization screening.

[0162] The method for optimizing and filtering the scenarios is as follows:

[0163] For each scenario under each configuration scheme, a penalty term is obtained by constraining the node voltage, and this term is added to the target value of the upper-level planning model.

[0164] Preferably, the method for constraining the node voltage is as follows:

[0165] Step S301: Based on the power flow calculation results, determine the voltage limit of each node, identify and record the voltage limit-over-limit nodes, and add penalty items to the scenarios corresponding to the voltage limit-over-limit nodes;

[0166] Step S302: Adjust the complex variable value of the voltage over-limit node until the target voltage returns to the allowable range, thereby correcting the voltage over-limit node;

[0167] Step S303: After all voltage over-limit nodes have been corrected, perform a new fast power flow calculation for the massive scenario based on the adjusted values ​​of the complex variables of each node, and update the power flow calculation results.

[0168] Step S304: Perform another voltage limit judgment on the updated power flow calculation results. If there are still voltage limit nodes in the updated scenario or a scenario still has a large limit (such as fluctuations of 10% or more), then add a large value penalty item to the scenario or filter it out directly.

[0169] In this field, as an option, upper and lower limits are set for the voltage of each node, with an allowable fluctuation range of 10%.

[0170] The optimization and screening method for the proposed scheme is as follows:

[0171] For each configuration scheme after scenario optimization and screening, the scenarios of each scheme after correction and screening in the above steps are statistically analyzed. If the penalty item corresponding to each configuration scheme is less than the penalty item threshold and the number of remaining scenarios is greater than the set scenario number threshold, then the configuration scheme is output as a reasonable energy storage configuration scheme; otherwise, based on sensitivity analysis, key nodes are determined, and the energy storage capacity configured at the key nodes is adjusted to optimize the configuration schemes that do not meet the requirements. After optimization, steps S200-S300 are executed again to delete unreasonable energy storage configuration schemes.

[0172] Preferably, configuration schemes that do not meet the requirements are optimized:

[0173] For configuration schemes that do not meet the screening criteria, key nodes are identified based on sensitivity analysis, and the energy storage capacity of the key nodes is adjusted to optimize the schemes that do not meet the requirements. Then, the process returns to step S200 to perform another round of massive scenario generation and HEM power flow calculation.

[0174] Furthermore, the method for identifying key nodes based on sensitivity analysis is as follows:

[0175] Statistically analyze the nodes where voltage exceeds limits under this scheme. When node p When a voltage over-limit occurs, the focus should be on the node. p Adjustments were made to the newly configured energy storage nodes in and around the nodes. Based on the power flow analytical expression composed of the power series coefficient matrix of the state variables obtained from HEM power flow calculation and the complex variables of each node, methods such as differentiation were used to determine the appropriate nodes. p The most significant impact of voltage on new energy storage nodes q This is used to adjust the energy storage configuration capacity in the plan. Assuming all configured energy storage devices are in a discharging state, if the nodes... p If the voltage exceeds the upper limit, a new energy storage node should be configured accordingly. q The configured capacity should be reduced. In this case, to meet the overall capacity requirements of the scheme, it is necessary to simultaneously select the node where the voltage exceeds the lower limit. r Similarly, apply the same logic to the nodes. r The new energy storage node configuration that has the most significant impact on voltage t The configuration capacity is increased, and the increase in capacity is proportional to the node's capacity. q The reduction in capacity at each location should be consistent.

[0176] Furthermore, the above energy storage configuration optimization methods must be adjusted according to the actual charging and discharging conditions of the energy storage equipment during implementation.

[0177] Although the steps in the above embodiments are described in the above order, those skilled in the art will understand that in order to achieve the effect of this embodiment, different steps do not need to be executed in such an order. They can be executed simultaneously (in parallel) or in a reverse order. These simple variations are all within the protection scope of this invention.

[0178] The second embodiment of the present invention provides a grid energy storage configuration optimization system based on a fully embedded method, the system comprising:

[0179] The data module is configured to acquire historical wind, solar and celestial load data and perform data preprocessing on the historical wind, solar and celestial load data.

[0180] The model building module is configured to build a two-layer planning model for energy storage configuration.

[0181] The energy storage configuration module is configured to generate a random configuration scheme.

[0182] The scene generation module is configured to generate massive scenes from preprocessed historical wind, solar and load data through Monte Carlo simulation and perform power flow calculations, cluster the generated massive scenes, and generate typical scenes after clustering.

[0183] The power flow calculation module is configured to perform power flow calculations on various typical scenarios to obtain preliminary calculation results; based on the complex variable correction mechanism of scenario difference analysis, the complex variables of each node in the massive number of similar scenarios corresponding to the typical scenarios are corrected, and the corrected complex variables are substituted into the power flow analytical expression to obtain the power flow calculation results.

[0184] The optimization and filtering module is configured to perform multi-level optimization and filtering of each configuration scheme based on the power flow analysis expression and power flow calculation results, and output a reasonable energy storage configuration.

[0185] Those skilled in the art will understand that, for the sake of convenience and brevity, the specific working process and related descriptions of the system described above can be found in the corresponding processes in the foregoing method embodiments, and will not be repeated here.

[0186] It should be noted that the grid energy storage configuration optimization system based on the fully embedded method provided in the above embodiments is only an example of the division of the above functional modules. In practical applications, the above functions can be assigned to different functional modules as needed, that is, the modules or steps in the embodiments of the present invention can be further decomposed or combined. For example, the modules in the above embodiments can be merged into one module, or further divided into multiple sub-modules to complete all or part of the functions described above. The names of the modules and steps involved in the embodiments of the present invention are only for distinguishing the various modules or steps and are not considered as an improper limitation of the present invention.

[0187] An electronic device according to a third embodiment of the present invention includes:

[0188] At least one processor; and

[0189] A memory communicatively connected to at least one of the processors; wherein,

[0190] The memory stores instructions that can be executed by the processor to implement the above-described grid energy storage configuration optimization method based on the fully embedded method.

[0191] A computer-readable storage medium according to a fourth embodiment of the present invention stores computer instructions, which are executed by the computer to implement the above-described grid energy storage configuration optimization method based on the fully embedded method.

[0192] Those skilled in the art will understand that, for the sake of convenience and brevity, the specific working process and related descriptions of the electronic device and computer-readable storage medium described above can be referred to the corresponding processes in the foregoing method embodiments, and will not be repeated here.

[0193] Those skilled in the art will recognize that the modules and method steps of the various examples described in conjunction with the embodiments disclosed herein can be implemented in electronic hardware, computer software, or a combination of both. The programs corresponding to the software modules and method steps can be placed in random access memory (RAM), main memory, read-only memory (ROM), electrically programmable ROM, electrically erasable programmable ROM, registers, hard disks, removable disks, CD-ROMs, or any other form of storage medium known in the art. To clearly illustrate the interchangeability of electronic hardware and software, the components and steps of the various examples have been generally described in terms of functionality in the foregoing description. Whether these functions are implemented in electronic hardware or software depends on the specific application and design constraints of the technical solution. Those skilled in the art can use different methods to implement the described functions for each specific application, but such implementation should not be considered beyond the scope of the invention.

[0194] Computer program code for performing the operations of this application can be written in one or more programming languages ​​or a combination thereof, including object-oriented programming languages ​​such as Java, Smalltalk, and C++, and conventional procedural programming languages ​​such as the "C" language or similar programming languages. The program code can be executed entirely on the user's computer, partially on the user's computer, as a standalone software package, partially on the user's computer and partially on a remote computer, or entirely on a remote computer or server. In cases involving remote computers, the remote computer can be connected to the user's computer via any type of network—including a local area network (LAN) or a wide area network (WAN)—or can be connected to an external computer (e.g., via the Internet using an Internet service provider).

[0195] The flowcharts and block diagrams in the accompanying drawings illustrate the architecture, functionality, and operation of possible implementations of systems, methods, and computer program products according to various embodiments of this application. In this regard, each block in a flowchart or block diagram may represent a module, segment, or portion of code containing one or more executable instructions for implementing a specified logical function. It should also be noted that in some alternative implementations, the functions indicated in the blocks may occur in a different order than those indicated in the drawings. For example, two consecutively indicated blocks may actually be executed substantially in parallel, and they may sometimes be executed in reverse order, depending on the functions involved. It should also be noted that each block in the block diagrams and / or flowcharts, and combinations of blocks in the block diagrams and / or flowcharts, can be implemented using a dedicated hardware-based system that performs the specified function or operation, or using a combination of dedicated hardware and computer instructions.

[0196] The terms “first”, “second”, etc., are used to distinguish similar objects, not to describe or indicate a specific order or sequence.

[0197] The term "comprising" or any other similar term is intended to cover non-exclusive inclusion, such that a process, method, article, or apparatus / device that comprises a list of elements includes not only those elements but also other elements not expressly listed, or elements inherent in such process, method, article, or apparatus / device.

[0198] The technical solution of the present invention has been described above with reference to the preferred embodiments shown in the accompanying drawings. However, it will be readily understood by those skilled in the art that the scope of protection of the present invention is obviously not limited to these specific embodiments. Without departing from the principles of the present invention, those skilled in the art can make equivalent changes or substitutions to the relevant technical features, and the technical solutions after these changes or substitutions will all fall within the scope of protection of the present invention.

Claims

1. A grid energy storage configuration optimization method based on the fully embedded method, characterized in that, Includes the following steps: Acquire historical data on wind, solar and lotus flowers, and perform data preprocessing on the historical data on wind, solar and lotus flowers; Construct a two-level planning model for energy storage configuration and solve the model based on the NSGA-II algorithm: Step S100: Initialize the population and randomly generate M configuration schemes, where M is the population size for the genetic algorithm; Step S200: Based on the capacity ratio of each node in each configuration scheme, and combined with the preprocessed historical wind-solar-load data, generate massive scenarios corresponding to each configuration scheme and perform parallel power flow calculations. The power flow calculation method for massive scenarios under each configuration scheme is as follows: Step S210: Perform cluster analysis on the massive number of scenes to obtain the optimal number of clusters, and obtain the clustered typical scenes based on the optimal number of clusters; Step S220: Perform HEM power flow calculation on the typical scenarios to obtain the power flow analytical expression corresponding to each typical scenario. The power flow analytical expression includes the power series coefficient matrix of state variables and the complex variables of each node. The method for calculating the HEM power flow is as follows: Step S221: Construct the holomorphic embedded form of the power balance equation and recursively solve for the power series coefficients of the state variables in each typical scenario; Step S222: Calculate PA based on the power series coefficients of the state variables, set the complex variable s of the node to 1, and obtain an approximate solution to the power flow equation; Step S223: Determine whether the approximate solution satisfies the termination condition; If satisfied, the approximate solution is the power flow steady-state solution; If not satisfied, increase the number of terms in the power series and the order of PA, and return to step S221 to perform the next order of recursion until the termination condition is met. Step S230: Based on the scenario difference analysis, the complex variable correction mechanism corrects the complex variables of each node in the massive number of similar scenarios corresponding to the typical scenario. The corrected complex variables are substituted into the power flow analysis expression to obtain the power flow calculation result. Step S300: Based on the power flow analytical expression and power flow calculation results, perform multi-level optimization and screening of each configuration scheme, and output the configuration scheme that meets the screening conditions as a reasonable energy storage configuration scheme.

2. The grid energy storage configuration optimization method based on the fully embedded method according to claim 1, characterized in that, The energy storage configuration two-layer planning model includes an upper-layer planning model and a lower-layer operation model. The upper-layer planning model takes minimizing the total system cost as the objective function, while the lower-layer operation model takes minimizing the system operating cost as the objective.

3. The grid energy storage configuration optimization method based on the fully embedded method according to claim 1, characterized in that, The method for obtaining the typical scenarios after clustering is as follows: Clustering algorithms are used to cluster massive amounts of data, and the cluster centers that minimize the objective function are output. k is the number of clusters that simultaneously ensures HEM adaptability and overall computational efficiency, and the scenario corresponding to each cluster center is taken as a typical scenario; the clustering includes the K-means clustering algorithm.

4. The grid energy storage configuration optimization method based on the fully embedded method according to claim 3, characterized in that, The termination condition is the convergence accuracy: When the approximate solution reaches the convergence accuracy, a power series is constructed based on the coefficients of the state variables. N×n Power series coefficient matrix of state variables ;in N The number of system nodes. n The number of iterations required for HEM to converge; Otherwise, if the number of iterations n reaches the set maximum limit, a filtering signal will be issued to filter out the typical scenario and its corresponding similar scenarios.

5. The grid energy storage configuration optimization method based on the fully embedded method according to claim 4, characterized in that, The power flow calculation results are obtained as follows: Step S231, for the first i A typical scenario The corresponding massive number of similar scenarios change the complex variables of the corresponding nodes. s Values ​​are selected to perform power flow calculations on the massive number of similar scenarios: ； ； In the formula, Indicates the first i Nodes in similar scenarios j The voltage holomorphic function; Indicates the first i Nodes in similar scenarios j A reactive holomorphic function; For nodes j Embedded complex variables; Step S232: Correct the complex variables of the massive number of similar scenes, using typical scenes. The complex variable takes the value 1 as the baseline. It varies with the generator output and load of each node; the specific value of the single-input complex variable of each node corresponds proportionally to the net load of each node in the current scenario and the net load of each node in a typical scenario: ； In the formula, For nodes in similar scenarios j The complex variable to be determined; Nodes in typical scenarios j A known complex variable with a value of 1; , Nodes in similar scenarios j The load and generator power; , Nodes in typical scenarios j The load and generator power; Step S233: Substitute the corrected complex variables of the massive number of similar scenarios into the power series coefficient matrix of the state variables of the corresponding clustering typical scenario to obtain the power flow calculation results.

6. The grid energy storage configuration optimization method based on the fully embedded method according to claim 2, characterized in that, The multi-level optimization screening includes scenario optimization screening and solution optimization screening: The method for optimizing and filtering the scenarios is as follows: For each scenario under each configuration scheme, a penalty term is obtained by constraining the node voltage, and this term is added to the target value of the upper-level planning model. The optimization and screening method for the proposed scheme is as follows: For each configuration scheme after the scenario optimization and screening, if the penalty item corresponding to each configuration scheme is less than the penalty item threshold and the number of remaining scenarios is greater than the set scenario number threshold, then the configuration scheme is output as a reasonable energy storage configuration scheme; otherwise, based on sensitivity analysis, key nodes are determined, and the energy storage capacity configured at the key nodes is adjusted to optimize the configuration schemes that do not meet the requirements. After optimization, steps S200-S300 are executed again to delete the unreasonable energy storage configuration schemes. The penalty item in the configuration scheme is the sum of the penalty items corresponding to the remaining scenarios under the configuration scheme.

7. The grid energy storage configuration optimization method based on the fully embedded method according to claim 6, characterized in that, The method for constraining node voltages is as follows: Step S301: Based on the power flow calculation results, determine the voltage limit of each node, identify and record the voltage limit-over-limit nodes, and add penalty items to the scenarios corresponding to the voltage limit-over-limit nodes; Step S302: Adjust the complex variable value of the voltage over-limit node until the target voltage returns to the allowable range, thereby correcting the voltage over-limit node; Step S303: After all voltage over-limit nodes have been corrected, return to step S230 based on the adjusted complex variable values ​​of each node, recalculate and update the power flow calculation results; Step S304: Perform another voltage limit judgment on the updated power flow calculation results. If there are still voltage limit exceeding nodes in the updated scenario, add the set large value penalty item to the scenario or filter them out directly.

8. The grid energy storage configuration optimization method based on the fully embedded method according to claim 6, characterized in that, The method for identifying key nodes based on sensitivity analysis is as follows: Differentiating the power flow analytical expression determines the relationship between nodes. p The new energy storage node with the greatest voltage impact q As a key node.

9. A grid energy storage configuration optimization system based on a fully embedded method, comprising the grid energy storage configuration optimization method based on a fully embedded method according to any one of claims 1-8, characterized in that, The system includes: The data processing module is configured to acquire historical wind, solar and load data and perform data preprocessing on the historical wind, solar and load data. The model building module is configured to build a two-layer planning model for energy storage configuration. The energy storage configuration module is configured to generate a random configuration scheme. The scene generation module is configured to generate massive scenes from preprocessed historical wind, solar and load data through Monte Carlo simulation and perform power flow calculations, cluster the generated massive scenes, and generate typical scenes after clustering. The power flow calculation module is configured to perform power flow calculations on various typical scenarios to obtain power flow analytical expressions. Based on the complex variable correction mechanism of scenario difference analysis, the complex variables of each node in the massive number of similar scenarios corresponding to the typical scenarios are corrected. The corrected complex variables are substituted into the power flow analytical expressions to obtain the power flow calculation results. The optimization and filtering module is configured to perform multi-level optimization and filtering of each configuration scheme based on the power flow analysis expression and power flow calculation results, and output a reasonable energy storage configuration scheme.

Images