Planning method and system for a posteriori extreme scenario re-identification considering energy storage system

Abstract

The present invention belongs to the field of integrated energy dispatch technology, and in particular, relates to a planning method and system that considers a posteriori extreme scenario re-identification of energy storage systems. The method includes: using an improved clustering and scenario reduction method to cluster and reduce scenarios of new energy output and load data to obtain typical scenarios; inputting typical scenarios into a multi-objective two-layer capacity optimization configuration model and solving them to obtain the optimal configuration capacity of each device in the electric-hydrogen hybrid energy storage system; performing a full-year optimization operation simulation on the obtained optimal configuration capacity to verify whether there are extreme scenarios in all scenarios of the electric-hydrogen hybrid energy storage system; if so, adding extreme scenarios to typical scenarios and re-optimizing the configuration, and iterating and looping to obtain the final optimal configuration capacity. The method proposed in the present invention can give full play to the daytime energy transfer potential of hydrogen energy storage and simultaneously consider normal scenarios and extreme scenarios of continuous low output of new energy, thereby reducing the dependence of the electric-hydrogen coupling system on the upper power grid.

Description

Technical Field

[0001] The present invention belongs to the technical field of integrated energy scheduling, and in particular relates to a planning method and system considering a posteriori extreme scenario re-identification of an energy storage system. Background Art

[0002] The statements in this section merely provide background information related to the present invention and do not necessarily constitute prior art.

[0003] Due to their intermittent and unstable nature, wind and solar power generation have a continuous low output scenario, which has a certain impact on the reliable operation of microgrids. At the same time, when wind and solar power generation have a continuous low output, microgrids need to purchase a large amount of electricity from the upper grid to meet the power demand during peak hours, which affects the economic efficiency of microgrid operation.

[0004] In order to improve the operational reliability and economy of microgrids, long- and short-cycle electric hydrogen energy storage systems play a vital role. Short-term electric energy storage equipment can quickly respond to changes in grid demand and enhance the instantaneous stability of the grid, while long-term hydrogen energy storage systems can achieve cross-day energy scheduling and provide more lasting and stable power support for microgrids.

[0005] However, most of the current planning scenarios for the electric-hydrogen coupling system use the typical day method. Although the amount of calculation is small, it cannot fully tap the potential of hydrogen storage in daily energy transfer. In addition, the typical day obtained by clustering and scenario reduction methods is easy to screen out extreme scenarios such as low-probability continuous low output of new energy, and cannot assess the risk of imbalance between energy supply and demand in the system in the long term, which has a certain impact on the reliability and economy of microgrid operation. In the face of this challenge, if the full-year data is used for optimization and configuration, although the system operation can be considered more comprehensively, the difficulty of solving the problem increases significantly, and the time cost of solving the problem is high. Summary of the invention

[0006] In order to address the deficiencies in the prior art, the present invention provides a planning method, system, electronic device, computer-readable storage medium and computer program product that consider the a posteriori extreme scenario re-identification of the energy storage system, which can give full play to the daytime energy transfer potential of hydrogen energy storage and simultaneously consider normal scenarios and extreme scenarios of continuous low output of new energy, reduce the dependence of the electric-hydrogen coupling system on the upper power grid, reduce the risk of medium- and long-term energy supply and demand imbalance, and thus improve the economy, environmental protection and reliability of the planning scheme.

[0007] To achieve the above objectives, one or more embodiments of the present invention provide the following technical solutions:

[0008] In a first aspect, the present invention provides a planning method for considering a posteriori extreme scenario re-identification of an energy storage system, comprising: clustering and reducing scenarios of new energy output and load data using an improved clustering and scenario reduction method to obtain a typical scenario;

[0009] Input typical scenarios into the constructed multi-objective dual-layer capacity optimization configuration model of the electric-hydrogen coupling system and solve it to obtain the optimal configuration capacity of each device in the electric-hydrogen hybrid energy storage system;

[0010] The optimal configuration capacity obtained is subjected to a full-year optimization operation simulation to verify whether there is an extreme scenario in which the electric-hydrogen hybrid energy storage system has excessive power purchases from the upper power grid due to continued low output of renewable energy under all scenarios; if so, the extreme scenario is added to the typical scenario and re-optimized, and the iterative cycle is carried out until the power purchase in the extreme scenario is reduced to the set range to obtain the final optimal configuration capacity.

[0011] In a second aspect, the present invention provides a planning system considering a posteriori extreme scenario re-identification of an energy storage system, comprising:

[0012] The energy storage system scenario clustering and reduction module is configured to: use an improved clustering and scenario reduction method to cluster and reduce the renewable energy output and load data to obtain a typical scenario;

[0013] The energy storage system optimal configuration capacity solving module is configured to: input typical scenarios into the constructed multi-objective double-layer capacity optimization configuration model of the electric-hydrogen hybrid energy storage system, and solve it to obtain the optimal configuration capacity of each device in the electric-hydrogen hybrid energy storage system;

[0014] The a posteriori extreme scenario re-identification module is configured to: conduct a full-year optimization operation simulation on the obtained optimal configuration capacity to verify whether there is an extreme scenario in which the electric-hydrogen hybrid energy storage system has excessive power purchases from the upper power grid due to continuous low output of renewable energy under all scenarios; if so, the extreme scenario is added to the typical scenario and re-optimized, and the iterative cycle is carried out until the power purchase in the extreme scenario is reduced to the set range to obtain the final optimal configuration capacity.

[0015] One or more of the above technical solutions have the following beneficial effects:

[0016] (1) The method proposed in the present invention can give full play to the daytime energy transfer potential of hydrogen energy storage and take into account both normal scenarios and extreme scenarios of continuous low output of new energy, reduce the dependence of the electric-hydrogen coupling system on the upper power grid, and reduce the risk of medium- and long-term imbalance in energy supply and demand, thereby improving the economy, environmental protection and reliability of the planning scheme; it is expected to further promote the application of the electric-hydrogen coupling system in the energy field and provide strong support for the further development of renewable energy technology.

[0017] (2) The present invention uses an improved neighbor propagation clustering algorithm based on a semi-supervised mechanism to explore the intrinsic grouping structure of the data, obtain the number of clusters and clustering scenarios, and further uses the synchronous back-elimination method based on Kantorovich distance to extract representative subsets in the clustering scenarios. The present invention uses a semi-supervised mechanism to update the similarity matrix in the neighbor propagation clustering algorithm, which can combine the intrinsic structure of a large amount of unlabeled data and the guiding information of a small amount of labeled data, thereby improving the generalization ability and accuracy of the model.

[0018] Advantages of additional aspects of the present invention will be given in part in the following description, and in part will become obvious from the following description, or will be learned through practice of the present invention. BRIEF DESCRIPTION OF THE DRAWINGS

[0019] The accompanying drawings in the specification, which constitute a part of the present invention, are used to provide a further understanding of the present invention. The exemplary embodiments of the present invention and their descriptions are used to explain the present invention and do not constitute improper limitations on the present invention.

[0020] Figure 1 This is a flow chart of a planning method for considering a posteriori extreme scenario re-identification of an energy storage system according to a first embodiment.

[0021] Figure 2 This is a flowchart of scene clustering and reduction in the first embodiment.

[0022] Figure 3 This is a flowchart of the a posteriori extreme scene re-identification of the first embodiment. DETAILED DESCRIPTION

[0023] It should be noted that the following detailed descriptions are exemplary and are intended to provide further explanation of the present invention. Unless otherwise specified, all technical and scientific terms used in the present invention have the same meanings as those commonly understood by those skilled in the art to which the present invention belongs.

[0024] It should be noted that the terms used herein are only for describing specific embodiments, and are not intended to limit exemplary embodiments according to the present invention. As used herein, unless the context clearly indicates otherwise, the singular form is also intended to include the plural form. In addition, it should be understood that the terms "include" and "have" and any of their variations are intended to cover non-exclusive inclusions, for example, a process, method, system, product or device comprising a series of steps or units is not necessarily limited to those steps or units clearly listed, but may include other steps or units that are not clearly listed or inherent to these processes, methods, products or devices.

[0025] In the absence of conflict, the embodiments of the present invention and the features of the embodiments may be combined with each other.

[0026] Embodiment 1

[0027] like Figure 1 As shown, this embodiment discloses a planning method considering a posteriori extreme scenario re-identification of an energy storage system, including:

[0028] Step 1: Use the improved clustering and scenario reduction method to cluster and reduce the renewable energy output and load data to obtain typical scenarios;

[0029] Step 2: Input the typical scenario into the constructed multi-objective double-layer capacity optimization configuration model of the electric-hydrogen coupling system, and use the solution algorithm to solve the multi-objective double-layer capacity optimization configuration model to obtain the optimal configuration capacity of each device in the electric-hydrogen hybrid energy storage system;

[0030] Step 3: The obtained optimal configuration capacity is simulated for a full year of optimization operation to verify whether there is an extreme scenario in which the electric-hydrogen hybrid energy storage system has excessive power purchases from the upper power grid due to continued low output of renewable energy under all scenarios; if so, the extreme scenario is added to the typical scenario and re-optimized, and the iterative cycle is repeated until the power purchase in the extreme scenario is reduced to within the set range to obtain the final optimal configuration capacity.

[0031] In step 1, the improved clustering and scene reduction methods include an improved neighbor propagation clustering algorithm based on a semi-supervised mechanism and a synchronous back-elimination method based on Kantorovich distance;

[0032] like Figure 2 As shown in the figure, the improved neighbor propagation clustering algorithm based on the semi-supervised mechanism is first used to group the data to obtain the number of clusters and clustering scenarios; then the synchronous back-elimination method based on Kantorovich distance is used to extract representative subsets, namely typical scenarios, from the clustering scenarios.

[0033] Among them, the improved neighbor propagation clustering algorithm based on semi-supervised mechanism is improved by integrating semi-supervised mechanism on the basis of the neighbor propagation clustering algorithm.

[0034] The neighbor propagation clustering algorithm does not require the number of clusters to be set in advance. Compared with traditional clustering algorithms such as K-means and K-medoids, it can reduce the impact of manually setting the number of clusters on clustering accuracy; the neighbor propagation clustering algorithm will automatically converge after summarizing the similarity score between each point and other points, thereby reducing the workload of human intervention. However, the accuracy of the neighbor propagation clustering algorithm is closely related to the similarity matrix, which contains the similarity score of each data point with other points, and the similarity score can be defined as a measurement method for measuring how similar or dissimilar two data points are.

[0035] Semi-supervised mechanism is a machine learning paradigm, which is between supervised learning and unsupervised learning. It assumes that the class label of the data is unknown and obtains certain prior knowledge through the pairwise constraint relationship between data points. Semi-supervised mechanism attempts to learn from unlabeled data and identify the potential structure and pattern of the data through the algorithm's own inference. The present invention uses a semi-supervised mechanism to update the similarity matrix, which can combine the intrinsic structure of a large amount of unlabeled data and the guiding information of a small amount of labeled data, thereby improving the generalization ability and accuracy of the model.

[0036] Step 101: grouping the data using an improved neighbor propagation clustering algorithm based on a semi-supervised mechanism to obtain the number of clusters and clustering scenarios; specifically including:

[0037] Step 1011: Data initialization

[0038] Step 1012: construct a similarity matrix of the required clustering data. The similarity matrix is ​​calculated as follows:

[0039] (1)

[0040] In the formula, Represents data points and The distance metric between and Represent two data points respectively.

[0041] Step 1013: Add pairwise constraints, adjust the similarity matrix, and simplify the similarity matrix:

[0042] if and If the Must-link constraint is satisfied, the similarity value is set to the maximum (10000), that is:

[0043] (2)

[0044] if and If the Cannot-link constraint is satisfied, the similarity value is set to the minimum (-10000), that is:

[0045] (3)

[0046] if and Neither the Must-link constraint nor the Cannot-link constraint is satisfied, but exists in the data set With data points If they are connected separately, the similarity of this pair of data points is adjusted to a larger similarity (take 1000), that is:

[0047] (4)

[0048] It should be noted that the maximum value of similarity can usually be set to a very high positive value to make the algorithm think that these points are highly similar, thus tending to classify them into the same category, while the minimum value of similarity is set to a very small negative value to prevent the algorithm from classifying them into the same category; in this embodiment, the maximum value is set to 10 5 , the minimum value is set to -10 5 , the smaller value is set to 1000.

[0049] if With data points If one data point satisfies the Must-link constraint but satisfies the Cannot-link constraint with another data point, then the pair of data points is considered to satisfy the Cannot-link constraint, and the similarity of the data points is adjusted to the minimum, that is:

[0050] (5)

[0051] Step 1014: Calculate the attraction matrix and the attribution matrix between the sample points of the required clustering data; wherein the sample points refer to the cluster centers automatically selected through calculation and optimization during the algorithm iteration process.

[0052] (6)

[0053] (7)

[0054] Step 1015: Update the attraction matrix and the attribution matrix, and determine which points can be used as the center points of the cluster by continuously updating the attraction matrix and the attribution matrix.

[0055] (8)

[0056] (9)

[0057] Step 1016: Iterate the loop until the stopping criterion is reached; if the selection of the cluster center does not change after multiple iterations, it can be considered that the algorithm has converged and the iteration is stopped.

[0058] Step 1017: Output cluster centers and sample groups.

[0059] After obtaining the clustering results, the simultaneous back-elimination method based on Kantorovich distance is further applied to reduce the number of original scenes to reduce the computational complexity while trying to maintain the statistical characteristics of the scene set.

[0060] Step 102: The implementation steps of the synchronous back-substitution elimination method based on Kantorovich distance specifically include:

[0061] Step 1021: Input each group of samples after clustering and initialize the scene set.

[0062] Step 1022: Calculate the Kantorovich distance between scenes. The Kantorovich distance is calculated as follows:

[0063] (10)

[0064] In the formula, Represents two probability distributions and Kantorovich distance between and Respectively represent the definitions in space and space The probability distribution on ; It means to find the minimum value. represents all possible joint distributions , so that their marginal distributions are and .

[0065] Step 1023: Select a scene pair.

[0066] Select the two scene pairs that are closest in distance and closest in probability distribution.

[0067] Step 1024: Synchronous back generation elimination.

[0068] The selected scene pairs are merged or eliminated to reduce the number of scenes. This step needs to ensure that the merging or elimination operation has the least impact on the statistical characteristics of the overall scene collection.

[0069] Step 1025: Update the scene set.

[0070] The scene collection is updated based on the elimination results, removing the merged or eliminated scenes, and new scenes may need to be added to maintain the representativeness of the scene collection.

[0071] Step 1026: Iterate the loop until the number of scenarios is reduced to a predetermined target number or other stopping conditions are met.

[0072] Step 1027: Output the reduced scenario set to obtain typical weekly data of net output.

[0073] In step 2, a multi-objective dual-layer capacity optimization configuration model for the electric-hydrogen coupling system is constructed, including:

[0074] Step 201: Establishing a life cycle investment cost model for a long- and short-cycle electric-hydrogen hybrid energy storage system that can be used to optimize configuration problems;

[0075] The full life cycle investment cost model of the long- and short-cycle electric-hydrogen hybrid energy storage system fully considers the costs incurred by electric energy storage and hydrogen energy storage in the initial construction stage, operation and maintenance stage, replacement stage and recovery stage, which correspond to the investment cost, operation and maintenance cost, replacement cost and recovery cost of electric energy storage and hydrogen energy storage respectively.

[0076] Among them, the investment cost of the hybrid energy storage system takes into account the hardware costs of battery packs, electrolyzers, hydrogen storage equipment, inverters, etc.; the operation and maintenance cost takes into account the daily operation and maintenance costs incurred during the operation of the energy storage system; as time goes by, some components in the energy storage system (such as battery cells, electrolyzer membranes, catalysts, etc.) may need to be replaced due to wear, aging or technological updates, so replacement costs will be incurred; when the energy storage system reaches the end of its life, it needs to be properly handled, including dismantling, recycling and waste disposal, which is the recovery cost.

[0077] Among them, the full life cycle investment cost model of the long- and short-cycle electric-hydrogen hybrid energy storage system is:

[0078] (11)

[0079] In the formula: D represents electrical energy storage, H represents hydrogen energy storage; , , , , They are the total life cycle cost of the energy storage system, investment cost, operation and maintenance cost, replacement cost, and recovery cost.

[0080] (1) Investment cost of energy storage system:

[0081] (12)

[0082] In the formula, , They are the initial installation cost coefficients of the configured capacity and configured power, in units of RMB / kWh; Configure capacity for electric and hydrogen energy storage; The rated power of electric energy storage and hydrogen energy storage; is the discount rate; Plan for the lifetime of your hybrid energy storage system.

[0083] (2) Operation and maintenance costs of energy storage systems:

[0084] (13)

[0085] In the formula, The real-time available electricity price. is the charge and discharge time.

[0086] (3) Replacement cost of energy storage system:

[0087] (14)

[0088] In the formula, is the number of times the energy storage device is replaced; , They are the replacement cost coefficients of the configured capacity and configured power, unit: yuan / kWh;

[0089] (4) Recovery cost of energy storage system:

[0090] (15)

[0091] In the formula, , are the recovery cost coefficients of configured capacity and configured power, RMB / kWh;

[0092] At the same time, based on the energy storage characteristics of electric energy storage and hydrogen energy storage, long-term hydrogen energy storage and short-term electric energy storage equipment models are constructed, including power models, start-stop models, and output models, to more flexibly respond to different energy needs.

[0093] (5) Electric energy storage model:

[0094] (16)

[0095] In the formula, , and Respectively represent the minimum and maximum values ​​of the energy storage operating power and the Actual operating power; Indicates the amount of energy stored in the energy storage at the beginning of a new day; , , and Respectively The charging and discharging power and efficiency of energy storage at all times; and They respectively represent the flags for charging and discharging.

[0096] (6) Hydrogen energy storage model:

[0097] (17)

[0098] (18)

[0099] In the formula, and are the state variables representing hydrogen injection and release, respectively; , and Respectively indicate the minimum and maximum pressures allowed in the hydrogen cylinder and the actual pressure; Indicates the amount of hydrogen stored in the hydrogen tank for the new week; Indicates the initial hydrogen storage capacity in the hydrogen cylinder.

[0100] In the hydrogen energy storage system model constructed in the present invention, the hydrogen storage tank is charged and discharged on a weekly basis. Compared with electric energy storage, which needs to restore the initial energy storage amount every day, hydrogen energy storage only needs to restore the initial energy storage amount every week, which can better balance the daytime energy within the week and thus give full play to the daytime energy transfer potential of hydrogen energy storage.

[0101] Step 202: construct a multi-objective dual-layer capacity optimization configuration model for an electric-hydrogen coupling system that takes into account both economy and environmental protection, including a capacity configuration layer and an optimization operation layer;

[0102] Among them, the electric-hydrogen coupling system includes wind and solar renewable energy power generation devices, electric-hydrogen coupling equipment, short-term electric energy storage equipment and long-term hydrogen energy storage equipment; at the same time, the system has electricity purchase and sales transactions with the upper-level power grid to meet the electricity load demand.

[0103] It should be noted that the electric-hydrogen coupling system in this embodiment includes not only the electric-hydrogen hybrid energy storage system but also the wind turbine photovoltaic power generation device and the electric load; although the equipment that needs to be configured with capacity is only the electric-hydrogen hybrid energy storage system, in the two-layer capacity optimization configuration model, both the wind turbine photovoltaic power generation and the electric load need to meet certain constraints, so the present invention constructs a multi-objective two-layer capacity optimization configuration model for the electric-hydrogen coupling system.

[0104] Step 2021: The capacity configuration layer takes planning economy and environmental protection as optimization goals, the capacity of the electric-hydrogen hybrid energy storage system as the decision variable, and the upper and lower limits of the equipment configuration capacity as constraints; specifically:

[0105] (1) The capacity configuration layer objective functions include:

[0106] Economic objective function:

[0107] (19)

[0108] (20)

[0109] in, and They represent the upper layer capacity configuration target and the lower layer operation optimization target respectively; It represents the investment conversion coefficient, which is taken as 0.0192 in this paper; , , They are the investment cost, replacement cost and recovery cost of the hybrid energy storage system.

[0110] Environmental protection objective function:

[0111] (twenty one)

[0112] in, represents the penalty coefficient for carbon dioxide emissions; represents the electricity-to-carbon conversion coefficient; Indicates time The amount of electricity purchased from the superior power grid.

[0113] (2) The capacity configuration layer constraints are:

[0114] (twenty two)

[0115] In the formula, , They respectively represent the upper and lower limits of the equipment configuration capacity; the equipment that requires capacity configuration includes hydrogen energy storage, electrical energy storage, hydrogen fuel cells, and electrolyzers.

[0116] Step 222: The optimization operation layer takes the economic efficiency of typical scenario operation as the optimization target, the operating power of the electric-hydrogen hybrid energy storage system equipment and the power of the power purchase and sale transaction between the system and the upper-level power grid as the decision variables, and the constraints include power balance constraints, power purchase and sale constraints, and electric-hydrogen coupling equipment constraints; specifically:

[0117] (1) The objective function of the optimized operation layer is:

[0118] (twenty three)

[0119] (twenty four)

[0120] In the formula, represents the non-utilization penalty coefficient of wind turbine power generation and photovoltaic power generation; and Represents the unused electricity from wind power and photovoltaic power generation; , , and They represent the purchase and sales coefficient and electricity quantity respectively.

[0121] (2) Optimize operating layer constraints:

[0122] Electric power balance constraints:

[0123] (25)

[0124] In the formula, Indicates the net output of the system, that is, the difference between the wind and solar power generation and the power load demand; Indicates the abandoned wind and solar power; and They represent the discharge power and charging power of the electric energy storage respectively; and Respectively represent the power purchased and sold by the system and the upper-level power grid; and They represent the hydrogen production power of the electrolyzer and the power generation power of the fuel cell respectively.

[0125] Restrictions on electricity purchase and sale:

[0126] (26)

[0127] In the formula, and Respectively represent the purchased power and sold power of the system and the upper power grid; 0-1 variable and Action flags that indicate whether the system is purchasing or selling electricity from the upper-level power grid.

[0128] Electricity-hydrogen coupling equipment constraints:

[0129] (27)

[0130] (28)

[0131] (29)

[0132] In the formula, , and Respectively represent the upper limit, lower limit and actual operating power of the electric-hydrogen coupling device during normal operation; Indicates the switching status of the device; Indicates the power consumed by the device when it is started; and They represent the hydrogen production efficiency and heat production efficiency of the electrolyzer respectively; and Indicates the quality and thermal power of hydrogen produced by the electrolyzer; and They represent the electricity generation efficiency and heat generation efficiency of hydrogen fuel cells respectively. and Represents the electrical power and thermal power generated by a hydrogen fuel cell.

[0133] In step 3, the a posteriori extreme scenario re-identification stage refers to the simulation of the optimal configuration capacity obtained after the two-layer capacity optimization configuration is performed under the typical scenario obtained by the improved clustering and scenario reduction method. The whole year optimization operation is simulated to check whether there is an extreme scenario under the configured capacity where the renewable energy continues to have low output and causes excessive power purchase from the upper power grid. If so, it is added to the typical scenario and re-optimized. The cycle is iterated until the power purchase in the extreme scenario is reduced to a reasonable range. The cycle ends and the optimal configuration plan is obtained.

[0134] like Figure 3 As shown in the figure, the integrated process of optimized configuration including a posteriori extreme scene re-identification is:

[0135] Step 301: using an improved clustering and scene reduction method to obtain typical scenes.

[0136] A clustering algorithm (an improved neighbor propagation clustering algorithm based on a semi-supervised mechanism) is used to group the scenes from the original data set. A scene reduction method (a synchronous back-elimination method based on Kantorovich distance) is applied to further select representative typical scenes in each group to reduce computational complexity.

[0137] Step 302: calling a multi-objective particle swarm algorithm to initialize the configuration capacity; calling an optimization solver (in this embodiment, the Gurobi solver is selected) to solve the operating power of the system equipment.

[0138] With the goals of economy and environmental protection, a multi-objective particle swarm algorithm is used for capacity configuration. After that, the a posteriori extreme scenario re-identification phase is entered, and the Gurobi solver is used to solve the underlying optimization operation model.

[0139] Input the typical week into the two-layer capacity optimization configuration model (using the multi-objective particle swarm algorithm to generate the capacity configuration plan and using gurobi to solve the objective function of the lower optimization model as the particle fitness value, iterative cycle) to obtain the capacity configuration plan;

[0140] Under this capacity configuration scheme, the optimized operation of 52 weeks throughout the year is tested. Matlab is still used to call the gurobi solver for solution. However, compared with the previous two-layer capacity optimization configuration process, the equipment capacity is fixed at this time, that is, to test whether the configuration scheme is applicable to all weeks of the year.

[0141] Step 303: Check the number of iterations.

[0142] Make sure the optimization process does not exceed a preset number of iterations and stop iterating if the limit is reached.

[0143] Step 304: Full cycle optimization operation inspection.

[0144] The configured initial capacity is simulated for a full cycle, i.e. 52 weeks a year, and it is checked whether there are extreme situations in which the amount of electricity purchased from the upper-level power grid is large under all scenarios.

[0145] Step 305: Identify extreme scenarios.

[0146] The power purchase amount of the upper power grid in all scenarios is normalized. If the normalized value is greater than 0.8, it is considered an extreme week, that is, the top 20% scenarios with the largest power purchase amount are selected as extreme scenarios.

[0147] Step 306: Determine the optimal capacity configuration.

[0148] The sum of electricity purchases for all extreme scenarios is used as the criterion for judging the optimal capacity configuration. If the sum of electricity purchases for all extreme scenarios of the current capacity configuration is less than the historical optimal value, the optimal capacity configuration is updated.

[0149] Step 307: Add extreme scenarios to typical scenarios, and re-call the multi-objective particle swarm algorithm to update the capacity configuration to ensure that the optimization model can fully consider these situations that have a significant impact on performance.

[0150] Step 308: Output the final optimal capacity configuration result.

[0151] Embodiment 2

[0152] This embodiment discloses a planning system that considers a posteriori extreme scenario re-identification of an energy storage system, including:

[0153] The energy storage system scenario clustering and reduction module is configured to: use an improved clustering and scenario reduction method to cluster and reduce the renewable energy output and load data to obtain a typical scenario;

[0154] The energy storage system optimal configuration capacity solving module is configured to: input typical scenarios into the constructed multi-objective double-layer capacity optimization configuration model of the electric-hydrogen hybrid energy storage system, and solve the multi-objective double-layer capacity optimization configuration model based on the improved multi-objective particle swarm solving algorithm and the optimization solver to obtain the optimal configuration capacity of each device in the electric-hydrogen hybrid energy storage system;

[0155] The a posteriori extreme scenario re-identification module is configured to: conduct a full-year optimization operation simulation on the obtained optimal configuration capacity to verify whether there is an extreme scenario in which the electric-hydrogen hybrid energy storage system has excessive power purchases from the upper power grid due to continuous low output of renewable energy under all scenarios; if so, the extreme scenario is added to the typical scenario and re-optimized, and the iterative cycle is carried out until the power purchase in the extreme scenario is reduced to the set range to obtain the final optimal configuration capacity.

[0156] In the specific implementation process, in the energy storage system scenario clustering and reduction module, the improved clustering and scenario reduction method includes an improved neighbor propagation clustering algorithm based on a semi-supervised mechanism and a synchronous back-generation elimination method based on Kantorovich distance;

[0157] Firstly, an improved neighbor propagation clustering algorithm based on semi-supervised mechanism is used to group the data to obtain the number of clusters and clustering scenarios. Then, the synchronous back-elimination method based on Kantorovich distance is used to extract representative subsets, namely typical scenarios, from the clustering scenarios.

[0158] The improved neighbor propagation clustering algorithm based on the semi-supervised mechanism is used to group the data to obtain the number of clusters and clustering scenarios, including:

[0159] Construct the similarity matrix of the required clustering data, add pairwise constraints to the data set, and adjust the similarity matrix, including:

[0160] If the data point pair satisfies the Must-link constraint, the similarity is adjusted to the maximum; if the data point pair satisfies the Cannot-link constraint, the similarity is adjusted to the minimum;

[0161] If the data point pair does not satisfy the Must-link constraint and the Cannot-link constraint, and there are data points in the data set that are connected to two data point pairs in the data points, the similarity of the data point pair is adjusted to the maximum;

[0162] If there is a data point in the data set that satisfies the Must-link constraint with one data point in the data point pair and satisfies the Cannot-link constraint with the other data point in the data point pair, the similarity of the data point pair is adjusted to the minimum;

[0163] Calculate the attraction matrix and belonging matrix between the sample points of the required clustering data and update them;

[0164] After loop iteration, the number of clusters and clustering scenarios are obtained.

[0165] In the energy storage system optimal configuration capacity solving module, the multi-objective double-layer capacity optimization configuration model of the electric-hydrogen coupling system includes a capacity configuration layer and an optimization operation layer;

[0166] The capacity configuration layer takes planning economy and environmental protection as optimization goals, takes the capacity of the electric-hydrogen hybrid energy storage system equipment as a decision variable, and takes the upper and lower limits of the equipment configuration capacity as constraints;

[0167] The optimization operation layer takes the economic efficiency of typical scenario operation as the optimization target, the operating power of the electric-hydrogen hybrid energy storage system equipment and the power of the power purchase and sale transaction between the system and the upper-level power grid as the decision variables, and the constraints include power balance constraints, power purchase and sale constraints, and electric-hydrogen coupling equipment constraints;

[0168] The capacity configuration layer is solved using an improved multi-objective particle swarm solving algorithm, and the optimization operation layer is solved using an optimization solver.

[0169] It should be noted that the above modules correspond to the steps in Embodiment 1, and the examples and application scenarios implemented by the above modules and the corresponding steps are the same, but are not limited to the contents disclosed in the above Embodiment 1. It should be noted that the above modules, as part of the system, can be executed in a computer system such as a set of computer executable instructions.

[0170] Embodiment 3

[0171] Embodiment 3 of the present invention provides an electronic device, including a memory and a processor, and computer instructions stored in the memory and executed on the processor. When the computer instructions are executed by the processor, the steps of the above-mentioned planning method for considering a posteriori extreme scenario re-identification of the energy storage system are completed.

[0172] Embodiment 4

[0173] Embodiment 4 of the present invention provides a computer-readable storage medium for storing computer instructions. When the computer instructions are executed by a processor, the steps of the above-mentioned planning method for considering a posteriori extreme scenario re-identification of an energy storage system are completed.

[0174] Embodiment 5

[0175] Embodiment 5 of the present invention provides a computer program product, including a computer program / instruction, which, when executed by a processor, implements the steps of the above-mentioned planning method for considering a posteriori extreme scenario re-identification of an energy storage system.

[0176] The present invention is described with reference to flowcharts and / or block diagrams of methods, devices (systems), and computer program products according to embodiments of the present invention. It should be understood that each process and / or block in the flowchart and / or block diagram, as well as the combination of processes and / or blocks in the flowchart and / or block diagram, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, a special-purpose computer, an embedded processor, or other programmable data processing device to produce a machine, so that the instructions executed by the processor of the computer or other programmable data processing device generate instructions for implementing the processes in the flowchart and / or block diagram. Figure 1 A process or multiple processes and / or boxes Figure 1 A device that provides the functions specified in a block or multiple blocks.

[0177] These computer program instructions may also be stored in a computer-readable memory capable of directing a computer or other programmable data processing device to operate in a specific manner, so that the instructions stored in the computer-readable memory produce an article of manufacture comprising an instruction device, which implements the process Figure 1 A process or multiple processes and / or boxes Figure 1 A function specified in one or more boxes.

[0178] These computer program instructions can also be loaded onto a computer or other programmable data processing device, and a series of operating steps are executed on the computer or other programmable device to produce a computer-implemented process, so that the instructions executed on the computer or other programmable device provide for implementing the process. Figure 1 A process or multiple processes and / or boxes Figure 1 The steps for the functions specified in one or more boxes.

[0179] The description of each embodiment in the above embodiments has different emphases. For parts not described in detail in a certain embodiment, reference can be made to the relevant descriptions of other embodiments.

[0180] The above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. For those skilled in the art, the present invention may have various modifications and variations. Any modification, equivalent replacement, improvement, etc. made within the spirit and principle of the present invention shall be included in the protection scope of the present invention.

Claims

1. A planning method considering a posteriori extreme scenario re-identification of energy storage systems, characterized in that: include: Improved clustering and scenario reduction methods are used to cluster and reduce the renewable energy output and load data to obtain typical scenarios. The improved clustering and scene reduction method includes an improved neighbor propagation clustering algorithm based on a semi-supervised mechanism and a synchronous back-generation elimination method based on Kantorovich distance; Firstly, an improved neighbor propagation clustering algorithm based on semi-supervised mechanism is used to group the data to obtain the number of clusters and clustering scenarios; Then, the simultaneous back-elimination method based on Kantorovich distance is used to extract representative subsets, i.e., typical scenes, from the clustering scenes; Input typical scenarios into the constructed multi-objective dual-layer capacity optimization configuration model of the electric-hydrogen hybrid energy storage system and solve it to obtain the optimal configuration capacity of each device in the electric-hydrogen hybrid energy storage system; The multi-objective double-layer capacity optimization configuration model for the electric-hydrogen hybrid energy storage system is constructed, including: Establish a full life cycle investment cost model for a long- and short-cycle electric-hydrogen hybrid energy storage system that can be used to optimize configuration problems; the full life cycle investment cost model for a long- and short-cycle electric-hydrogen hybrid energy storage system includes the costs incurred by electric energy storage and hydrogen energy storage in the initial construction stage, operation and maintenance stage, replacement stage, and recovery stage, which correspond to the investment cost, operation and maintenance cost, replacement cost, and recovery cost of electric energy storage and hydrogen energy storage, respectively; The multi-objective double-layer capacity optimization configuration model of the electric-hydrogen hybrid energy storage system includes a capacity configuration layer and an optimization operation layer; The capacity configuration layer takes planning economy and environmental protection as optimization goals, takes the capacity of the electric-hydrogen hybrid energy storage system equipment as a decision variable, and takes the upper and lower limits of the equipment configuration capacity as constraints; The optimization operation layer takes the economic efficiency of typical scenario operation as the optimization target, the operating power of the electric-hydrogen hybrid energy storage system equipment and the power of the power purchase and sale transaction between the system and the upper-level power grid as the decision variables, and the constraints include power balance constraints, power purchase and sale constraints, and electric-hydrogen hybrid energy storage equipment constraints; The capacity configuration layer is solved using an improved multi-objective particle swarm solving algorithm, and the optimized operation layer is solved using an optimization solver. The optimal configuration capacity is subjected to a full-year optimization operation simulation to verify whether there is an extreme scenario in which the electric-hydrogen hybrid energy storage system has excessive power purchases from the upper power grid due to continued low output of renewable energy under all scenarios. If so, the extreme scenario is added to the typical scenario and re-optimized, and the iterative cycle is repeated until the power purchases in the extreme scenario are reduced to a set range to obtain the final optimal configuration capacity.

2. A planning method for considering a posteriori extreme scenario re-identification of an energy storage system as claimed in claim 1, characterized in that: The improved neighbor propagation clustering algorithm based on the semi-supervised mechanism is used to group the data to obtain the number of clusters and clustering scenarios, including: Construct the similarity matrix of the required clustering data, add pairwise constraints to the data set, and adjust the similarity matrix, including: If the data point pair satisfies the Must-link constraint, the similarity is adjusted to the maximum; if the data point pair satisfies the Cannot-link constraint, the similarity is adjusted to the minimum; If the data point pair does not satisfy the Must-link constraint and the Cannot-link constraint, and there are data points in the data set that are connected to two data point pairs in the data points, the similarity of the data point pair is adjusted to the maximum; If there is a data point in the data set that satisfies the Must-link constraint with one data point in the data point pair and satisfies the Cannot-link constraint with the other data point in the data point pair, the similarity of the data point pair is adjusted to the minimum; Calculate the attraction matrix and belonging matrix between the sample points of the required clustering data and update them; After loop iteration, the number of clusters and clustering scenarios are obtained.

3. A planning method for considering a posteriori extreme scenario re-identification of energy storage systems as claimed in claim 1, characterized in that: The electric-hydrogen hybrid energy storage system includes wind and solar renewable energy power generation devices, electric-hydrogen hybrid energy storage equipment, short-term electric energy storage equipment and long-term hydrogen energy storage equipment; at the same time, the electric-hydrogen hybrid energy storage system has electricity purchase and sales transactions with the upper-level power grid.

4. A planning method for considering a posteriori extreme scenario re-identification of an energy storage system as claimed in claim 1, characterized in that: Calculate the normalized value of the power purchased by the upper-level power grid in all scenarios. If the normalized value in a scenario is greater than the set threshold, it is an extreme scenario.

5. A planning system considering a posteriori extreme scenario re-identification of energy storage systems, characterized in that: include: The energy storage system scenario clustering and reduction module is configured to: use an improved clustering and scenario reduction method to cluster and reduce the renewable energy output and load data to obtain a typical scenario; the improved clustering and scenario reduction method includes an improved neighbor propagation clustering algorithm based on a semi-supervised mechanism and a synchronous back-generation elimination method based on Kantorovich distance; Firstly, an improved neighbor propagation clustering algorithm based on semi-supervised mechanism is used to group the data to obtain the number of clusters and clustering scenarios; Then, the simultaneous back-elimination method based on Kantorovich distance is used to extract representative subsets, i.e., typical scenes, from the clustering scenes; The energy storage system optimal configuration capacity solving module is configured to: input typical scenarios into the constructed multi-objective double-layer capacity optimization configuration model of the electric-hydrogen hybrid energy storage system and solve it to obtain the optimal configuration capacity of each device in the electric-hydrogen hybrid energy storage system; The multi-objective double-layer capacity optimization configuration model for the electric-hydrogen hybrid energy storage system is constructed, including: Establish a full life cycle investment cost model for a long- and short-cycle electric-hydrogen hybrid energy storage system that can be used to optimize configuration problems; the full life cycle investment cost model for a long- and short-cycle electric-hydrogen hybrid energy storage system includes the costs incurred by electric energy storage and hydrogen energy storage in the initial construction stage, operation and maintenance stage, replacement stage, and recovery stage, which correspond to the investment cost, operation and maintenance cost, replacement cost, and recovery cost of electric energy storage and hydrogen energy storage, respectively; The multi-objective double-layer capacity optimization configuration model of the electric-hydrogen hybrid energy storage system includes a capacity configuration layer and an optimization operation layer; The capacity configuration layer takes planning economy and environmental protection as optimization goals, takes the capacity of the electric-hydrogen hybrid energy storage system equipment as a decision variable, and takes the upper and lower limits of the equipment configuration capacity as constraints; The optimization operation layer takes the economic efficiency of typical scenario operation as the optimization target, the operating power of the electric-hydrogen hybrid energy storage system equipment and the power of the power purchase and sale transaction between the system and the upper-level power grid as the decision variables, and the constraints include power balance constraints, power purchase and sale constraints, and electric-hydrogen hybrid energy storage equipment constraints; The capacity configuration layer is solved by using an improved multi-objective particle swarm solving algorithm, and the optimization operation layer is solved by using an optimization solver; The a posteriori extreme scenario re-identification module is configured to: conduct a full-year optimization operation simulation on the obtained optimal configuration capacity to verify whether there is an extreme scenario in which the electric-hydrogen hybrid energy storage system has excessive power purchases from the upper power grid due to continuous low output of renewable energy under all scenarios; if so, the extreme scenario is added to the typical scenario and re-optimized, and the iterative cycle is carried out until the power purchase in the extreme scenario is reduced to the set range to obtain the final optimal configuration capacity.

6. A planning system considering a posteriori extreme scenario re-identification of energy storage systems as claimed in claim 5, characterized in that: In the energy storage system scenario clustering and reduction module, the improved clustering and scenario reduction method includes an improved neighbor propagation clustering algorithm based on a semi-supervised mechanism and a synchronous back-generation elimination method based on Kantorovich distance; Firstly, an improved neighbor propagation clustering algorithm based on semi-supervised mechanism is used to group the data to obtain the number of clusters and clustering scenarios; Then, the simultaneous back-elimination method based on Kantorovich distance is used to extract representative subsets, namely typical scenes, from the clustering scenes.

7. A planning system considering a posteriori extreme scenario re-identification of energy storage systems as claimed in claim 6, characterized in that: The improved neighbor propagation clustering algorithm based on the semi-supervised mechanism is used to group the data to obtain the number of clusters and clustering scenarios, including: Construct the similarity matrix of the required clustering data, add pairwise constraints to the data set, and adjust the similarity matrix, including: If the data point pair satisfies the Must-link constraint, the similarity is adjusted to the maximum; if the data point pair satisfies the Cannot-link constraint, the similarity is adjusted to the minimum; If the data point pair does not satisfy the Must-link constraint and the Cannot-link constraint, and there are data points in the data set that are connected to two data point pairs in the data points, the similarity of the data point pair is adjusted to the maximum; If there is a data point in the data set that satisfies the Must-link constraint with one data point in the data point pair and satisfies the Cannot-link constraint with the other data point in the data point pair, the similarity of the data point pair is adjusted to the minimum; Calculate the attraction matrix and belonging matrix between the sample points of the required clustering data and update them; After loop iteration, the number of clusters and clustering scenarios are obtained.

8. A planning system considering a posteriori extreme scenario re-identification of energy storage systems as claimed in claim 5, characterized in that: In the energy storage system optimal configuration capacity solving module, the multi-objective double-layer capacity optimization configuration model of the electric-hydrogen hybrid energy storage system includes a capacity configuration layer and an optimization operation layer; The capacity configuration layer takes planning economy and environmental protection as optimization goals, takes the capacity of the electric-hydrogen hybrid energy storage system equipment as a decision variable, and takes the upper and lower limits of the equipment configuration capacity as constraints; The optimization operation layer takes the economic efficiency of typical scenario operation as the optimization target, the operating power of the electric-hydrogen hybrid energy storage system equipment and the power of the power purchase and sale transaction between the system and the upper-level power grid as the decision variables, and the constraints include power balance constraints, power purchase and sale constraints, and electric-hydrogen hybrid energy storage equipment constraints; The capacity configuration layer is solved using an improved multi-objective particle swarm solving algorithm, and the optimization operation layer is solved using an optimization solver.

Images