A method and system for virtual power plant participating in optimal dispatch of super large-scale power grid

Abstract

The present application relates to a kind of virtual power plant participates in the optimization scheduling method and system of super large scale power grid, the method includes the following steps: obtaining the relevant data of virtual power plant and super large scale power grid, wherein multiple units are provided in super large scale power grid;Based on the relevant data, construct all time period security constrained unit commitment model, and using trained neural network model to solve the initial unit start-stop state, the initial output distribution of virtual power plant and unit;Based on the initial unit start-stop state, the output distribution of virtual power plant and unit, according to similar time period is merged;Based on the merging result, construct security constrained economic dispatch model and solve, obtain the optimal output of virtual power plant and unit, complete scheduling process.Compared with prior art, the present application has the advantages of improving scheduling flexibility and stability.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of power grid optimal scheduling, in particular to a virtual power plant participating in the optimal scheduling of a super large-scale power grid and a system thereof. BACKGROUND

[0002] With the acceleration of global energy transformation and the widespread application of renewable energy, virtual power plants (VPP) as an innovative energy aggregation and management technology have received widespread attention. Virtual power plants integrate various resources together to form a virtual entity that can be flexibly scheduled, providing auxiliary services and power support for the power grid. In a super large-scale power grid, the optimal scheduling method of virtual power plants is particularly important.

[0003] The complexity and dynamics of super large-scale power grids pose a huge challenge to traditional scheduling methods. Due to the intermittency and uncertainty of distributed energy and the volatility of power demand, traditional centralized scheduling methods often fail to meet the requirements of real-time and accuracy. The emergence of virtual power plants provides a new way to solve these problems. By integrating multiple resources, virtual power plants can achieve complementary and optimal allocation of resources, improve energy utilization efficiency, and enhance the stability of the power grid. However, due to the characteristics of super large-scale power grids, multiple factors need to be considered when performing optimal scheduling, resulting in very complex existing scheduling methods, and sometimes the scheduling scheme is not suitable, leading to insufficient stability and reliability of the power grid, which is usually not applicable to actual application scenarios.

[0004] Therefore, there is an urgent need for a scheduling method to solve the above problems and ensure the safe and stable operation of the power grid. SUMMARY

[0005] The purpose of the present application is to provide a virtual power plant participating in the optimal scheduling of a super large-scale power grid that improves scheduling accuracy.

[0006] The purpose of the present application can be achieved by the following technical solutions:

[0007] A virtual power plant participating in the optimal scheduling of a super large-scale power grid, comprising:

[0008] Obtaining relevant data of the virtual power plant and the super large-scale power grid, wherein the super large-scale power grid has multiple units;

[0009] Based on the relevant data, a full-period safety constraint unit commitment model is constructed, and a trained neural network model is used to solve the initial unit start-stop state, the initial output distribution of the virtual power plant and the units;

[0010] Based on the initial unit start-stop state, the output distribution of the virtual power plant and the units, and according to similar time periods, merging is performed;

[0011] Based on the merging result, a security-constrained economic dispatch model is constructed and solved to obtain optimal outputs of the virtual power plant and the units, and the dispatch process is completed.

[0012] Further, the full-period security-constrained unit commitment model takes the minimum total operation cost as the target, and includes a target function and corresponding constraint conditions, wherein the target function is:

[0013] ,

[0014] Wherein:

[0015] The generation cost of the conventional unit:

[0016] ,

[0017] The start-up cost of the conventional unit:

[0018] ,

[0019] The cost of the virtual power plant:

[0020] ,

[0021] In the formula, is to minimize the total cost, is a set of time periods, is a set of units, is the total cost of the conventional unit j , is the generation cost curve of the conventional unit j , is the active power output of the unit j in the time period t , , indicates the start-stop state of the unit j in the time period t , is the start-stop cost of the unit j , , indicates the start-up action of the unit j in the time period t , is the overall cost of the VPP, is the net output of the VPP in the time period t , , , is a weight coefficient, , , …, is a weight coefficient, , …, is a set value, , the minimum, maximum net output of VPP;

[0022] The constraints include:

[0023] Power balance constraint:

[0024] ,

[0025] where, is the renewable energy i the output of renewable energy in time period t , is the set of renewable energy, is the total load of the system in time period t ;

[0026] Traditional unit constraints:

[0027] 1) Upper and lower output constraints:

[0028] ,

[0029] where, , is the minimum, maximum output of unit j ;

[0030] 2) Ramp constraints:

[0031] ,

[0032] where, is the ramp rate of unit j , is the length of time period;

[0033] 3) Minimum up time constraints:

[0034] ,

[0035] where, is the minimum up time of unit j in time period t , is the operation state of unit j in time period , is the start-up variable of unit j in time period t ;

[0036] 4) Minimum down time constraints:

[0037] ,

[0038] where, is the minimum down time of unitj During the period t Minimum downtime, For the unit j During the period t The shutdown variable;

[0039] 5) Start / Stop logic constraints:

[0040] ,

[0041] Virtual power plant constraints:

[0042] 1) Net output range constraint:

[0043] ,

[0044] In the formula, , These are the minimum and maximum net outputs of VPP;

[0045] 2) Climbing constraint:

[0046] ,

[0047] In the formula, The gradeability of VPP;

[0048] 3) Energy conservation constraint:

[0049] ,

[0050] ,

[0051] In the formula, For time period t energy, , For charging and discharging efficiency, The duration is the length of the time period. , These are the upper and lower limits of energy.

[0052] Power grid security constraints:

[0053] 1) Current constraints:

[0054] ,

[0055] In the formula, For line flow, For the set of power grid bus nodes, The line-node power transfer distribution factor. For connecting busbars n A collection of units, For VPP on the busn component output, load value of the bus n in the time period t;

[0056] 2) line capacity limit constraint:

[0057] ,

[0058] wherein, is the transmission line capacity limit value, is the transmission line set.

[0059] Further, in the training process, the neural network model is constructed as a graph model by constructing a full-time period safety constraint unit combination model, based on the structure of the graph model, a neural network model is selected and trained to obtain a trained neural network model.

[0060] Further, the graph model is represented as G=(V,E,W), wherein V is a node set, represented as:

[0061] ,

[0062] wherein, is a graph node, representing the operating state of the unit j in the time period t , is a graph node, representing the net output of the VPP in the time period t , is a graph node, as a system-level coupling point, representing power balance, is a unit set, is a time period set;

[0063] E represents an edge set, including time edges, space edges and constraint edges, wherein the time edges include unit ramping constraint edges ( ) and virtual power plant ramping constraint edges ( ), the space edges include the coefficient of unit output in power balance ( ), the coefficient of virtual power plant output in power balance ( ) and power flow constraints ( ), , the constraint edges include minimum operating time constraint edges and minimum shutdown time constraints;

[0064] W represents a weight set, including node weights and edge weights, wherein the node weights include the total cost j of the traditional unit and the overall cost of the VPP, the edge weights include time edge weights, space edge weights and constraint edge weights.

[0065] Furthermore, the neural network model is a CNN model or an RNN model.

[0066] Furthermore, the step of merging according to similar time periods includes:

[0067] a) Feature vector construction:

[0068] Build each time period t eigenvectors , represented as:

[0069] ,

[0070] In the formula, For the time period t The total system load, For the number of online units, For the assembly of generator units, For the unit j During the period t The initial start-stop state, The proportion of traditional power supplies, For the unit j During the period t The initial output distribution;

[0071] b) Distance calculation:

[0072] Calculate the current time period t distance :

[0073] ,

[0074] In the formula, , In order to be in t During the period a Dot and b The system load at the point, , for t Units during the period j exist a Dot and b Start / stop status of the point. , These are adjustable weighting coefficients. + ;

[0075] Based on the distance To obtain the similarity score , represented as:

[0076] ,

[0077] In the formula, is a scaling factor;

[0078] c) Typical scenario clustering:

[0079] Initialization of cluster centers:

[0080] ,

[0081] where K is the number of clusters, i.e. the number of typical scenarios, is the center of the k th cluster, is the similarity between the time period t and the time period ;

[0082] Assignment of time periods:

[0083] ,

[0084] where is the set of time periods of the k th cluster, is the feature vector of the th cluster center;

[0085] Virtual power plant engagement correction:

[0086] If the virtual power plant engagement t of time period , then it is forced to be an independent cluster and the is updated, otherwise no operation is performed, where is the initial net output distribution of the virtual power plant at time period t ;

[0087] Update of cluster centers:

[0088] ,

[0089] where is the feature vector of the k th cluster center;

[0090] Iteration until convergence: repeat the steps of assignment of time periods to update of cluster centers until the cluster assignment no longer changes, resulting in a merged time period set containing k K typical time periods and k typical load values for each typical time period .

[0091] Further, the step of merging according to similar time periods further comprises: outputting a constraint activation mapping table by performing constraint validity judgment to avoid ignoring key line constraints after merging, wherein the specific steps of outputting the constraint activation mapping table comprise:

[0092] For each time period t and each line l Calculate the original power flow distribution:

[0093] ,

[0094] In the formula, For the original tidal current distribution, The line-node power transfer distribution factor. For connecting busbars n A collection of generator units, For VPP on the bus n The initial component of the output force, For time period t Internal busbar n The load;

[0095] Identify key constraint periods:

[0096] ,

[0097] In the formula, For the critical constraint period, For transmission line capacity limits;

[0098] Using constraint activation logic, a constraint activation mapping table is output to avoid ignoring critical path constraints after merging. The constraint activation mapping is represented as follows:

[0099] .

[0100] Furthermore, the safety-constrained economic scheduling model includes an objective function and corresponding constraints, wherein the objective function is based on a typical time period. k The objective is to minimize operating cost, expressed as:

[0101] ,

[0102] In the formula, Typical period k Minimum operating cost For traditional units j The electricity generation cost curve, For traditional units j During typical periods k Those who have made contributions This is the overall quote for VPP. For VPP during typical periods k Net output power, For unit assembly;

[0103] The constraints include:

[0104] Power balance constraints:

[0105] ,

[0106] In the formula, For renewable energy i The predicted output For renewable energy collection, Typical period k Typical load values;

[0107] Unit output constraints:

[0108] ,

[0109] In the formula, For the unit j The initial start-stop state, , For the unit j Minimum and maximum output;

[0110] Virtual power plant constraints:

[0111] (1) Net output range constraint:

[0112] ,

[0113] In the formula, , These are the minimum and maximum net outputs of VPP;

[0114] (2) Energy conservation constraint:

[0115] ,

[0116] ,

[0117] In the formula, Typical period k energy, , For charging and discharging efficiency, The duration is the length of the time period. , These are the upper and lower limits of energy.

[0118] Power grid security constraints:

[0119] (1) Current constraint:

[0120] ,

[0121] In the formula, For line flow, is a set of bus nodes of the power grid, is a line-node power transfer distribution factor, is a set of buses connected to the power grid, n is a set of generating units connected to the bus, is a component output of the VPP at the bus, n is a load value of the bus at a typical time period, is a load value of the bus at a typical time period, k is a load value of the bus at a typical time period, n is a load value of the bus at a typical time period.

[0122] (2) Line capacity limit constraint:

[0123] ,

[0124] wherein, is a transmission line capacity limit value, is a set of transmission lines.

[0125] Further, the security-constrained economic dispatch model is solved by calling a pandapower solver.

[0126] The application also provides an optimal dispatch system for a virtual power plant participating in a super large-scale power grid, comprising:

[0127] a data acquisition module configured to acquire relevant data of the virtual power plant and the super large-scale power grid, wherein the super large-scale power grid comprises a plurality of generating units;

[0128] an initial state and output acquisition module configured to output initial generating unit start-stop states, initial output distributions of the virtual power plant and the generating units based on the relevant data and by using a trained neural network model;

[0129] a merging module configured to merge the initial generating unit start-stop states and the output distributions of the virtual power plant and the generating units according to similar time periods;

[0130] a dispatch module configured to construct a security-constrained economic dispatch model and solve the model based on the merging result to obtain optimal outputs of the virtual power plant and the generating units and complete the dispatch process.

[0131] Compared with the prior art, the application has the following beneficial effects:

[0132] (1) The application constructs a full-period security-constrained generating unit combination model and solves the model to obtain initial start-stop states and output distributions, mines similarities to perform time period merging, reduces the problem solving scale, and solves the technical problem of low dispatching precision caused by high dispatching calculation complexity when the virtual power plant participates in the super large-scale power grid.

[0133] (2) The application can provide a learnable topological structure for a neural network by converting a complex SCUC model into a graph structured expression form, thereby accelerating the solution of the SCUC model without losing the optimization accuracy.

[0134] (3) The application reduces the number of optimization time periods by the similar time period merging method, reduces the problem size while ensuring the rationality and safety of the scheduling optimization result, thereby improving the technical problems of the virtual power plant in participating in the scheduling of the super large scale power grid, such as complexity, low precision and low safety.

[0135] (4) The application constructs a model with the minimum operation cost as the target, realizes the dual optimization of economy and scheduling precision, and has important significance for real application scenarios. BRIEF DESCRIPTION OF DRAWINGS

[0136] Figure 1 The figure is a schematic diagram of the method of the application. DETAILED DESCRIPTION

[0137] The application will be described in detail below in combination with the drawings and specific embodiments. The embodiments are implemented on the premise of the technical solution of the application, and detailed implementation modes and specific operation processes are given, but the protection scope of the application is not limited to the following embodiments.

[0138] Embodiment 1

[0139] The embodiment provides a virtual power plant participating in super large scale power grid optimization scheduling method, as shown in the figure, the method comprises the following steps: Figure 1 The method comprises the following steps:

[0140] S1, obtaining the related data of the virtual power plant and the super large scale power grid, wherein the super large scale power grid is provided with a plurality of units.

[0141] The related data of the virtual power plant and the super large scale power grid obtained in the embodiment includes the total load of the system composed of the virtual power plant and the super large scale power grid , power, net output offer of the virtual power plant (VPP), state, power generation cost, start-stop cost, etc. of each traditional unit in the super large scale power grid.

[0142] S2, based on the related data, constructing a full time period safety constraint unit combination model, and using a trained neural network model to solve the initial unit start-stop state, the initial output distribution of the virtual power plant and the unit.

[0143] S21, constructing a full time period safety constraint unit combination (SCUC) model

[0144] The embodiment takes the minimum total operation cost in the whole period (including 12 hours, one day, one week, etc.) as the target, and constructs a whole period security constrained unit commitment model. The objective function is:

[0145]

[0146] Among them:

[0147] The traditional unit generation cost:

[0148]

[0149] The traditional unit start-up cost:

[0150]

[0151] The virtual power plant cost:

[0152]

[0153] In the formula, is the minimum total cost, is the time period set, is the unit set, is the total cost of the traditional unit j , is the generation cost curve of the traditional unit j , is the active power output of the unit j in the time period t , , j indicates the start-stop state of the unit t in the time period , j is the start-stop cost of the unit , j , t indicates the start-up action of the unit in the time period , t is the overall cost of the VPP, , , is the net output of the VPP in the time period , , is the weight coefficient, , is the set value, , is the minimum and maximum net output of the VPP;

[0154] The key constraint conditions are:

[0155] ​​​​Power balance constraint:

[0156] ,

[0157] where, is the renewable energy i output in time period t , is the set of renewable energy sources, is the total system load in time period t ;

[0158] Conventional unit constraints:

[0159] 1) Upper and lower output constraints:

[0160] ,

[0161] where, , is the minimum and maximum output of unit j ;

[0162] 2) Ramp constraints:

[0163] ,

[0164] where, is the ramp rate of unit j , is the time period length;

[0165] 3) Minimum up time constraints:

[0166] ,

[0167] where, is the minimum up time of unit j in time period t , is the operation status of unit j in time period , is the start-up variable of unit j in time period t ;

[0168] 4) Minimum down time constraints:

[0169] ,

[0170] where, is the minimum down time of unit j in time period t , is the operation status of unit j in time period t ;parking variable;

[0171] 5) start-stop logic constraints:

[0172] ,

[0173] virtual power plant constraints:

[0174] 1) net power range constraints:

[0175] ,

[0176] where, , Pmin, Pmaxare the minimum and maximum net power of the VPP;

[0177] 2) ramp constraints:

[0178] ,

[0179] where, RampVPPis the ramp rate of the VPP;

[0180] 3) energy conservation constraints:

[0181] ,

[0182] ,

[0183] where, E(t) is the energy of the time period t , , η is the charge-discharge efficiency, T is the time period length, , Emin, Emmaxare the energy lower and upper bounds;

[0184] grid security constraints:

[0185] 1) power flow constraints:

[0186] ,

[0187] where, P is the line power flow, B is the set of grid bus nodes, B is the line-to-node power transfer distribution factor, G is the set of generators connected to bus n , Pbusis the component power of the VPP at bus n , Pload(t) is the load value at bus n during time period t;

[0188] 2) Line capacity constraints:

[0189] ,

[0190] In the formula, For transmission line capacity limits, For power transmission line collection.

[0191] S22, Constructing a graph model

[0192] This embodiment constructs a graph model based on the aforementioned SCUC model to intuitively represent the complex constraints of the SCUC model graphically, facilitating understanding and analysis. By transforming the SCUC model into a structured representation, a learnable topological structure can be provided for the neural network, accelerating the solution of the SCUC model without sacrificing optimization accuracy.

[0193] Graph structures can be used to identify the independence between constraints, thereby decomposing the problem and simplifying the solution.

[0194] The graph structure is defined as G=(V,E,W):

[0195] Node set :

[0196] Graph Node :unit j During the period t The running status;

[0197] Graph Node VPP during the time period t Net output;

[0198] Graph Node : As a system-level coupling point, it represents power balance.

[0199] Edge set E:

[0200] Time edge: including unit ramp-up constraint edge ( ) and virtual power plant ramp constraint edge ( );

[0201] Spatial edge: includes the coefficient of unit output in power balance ( ), the coefficient of virtual power plant output in power balance ( ) and trend constraints ( ), ;

[0202] Constraint edges: including minimum running time constraint edges and minimum downtime constraint edges.

[0203] Weight set W:

[0204] Node weights: including traditional unitsTotal cost of j and VPP overall cost and VPP overall cost ;

[0205] Edge weight: including time edge weight, space edge weight and constraint edge weight.

[0206] S23, neural network model solves graph model

[0207] The embodiment adopts CNN neural network model or RNN neural network model to solve the historical solution, and predicts the initial solution of the current problem.

[0208] Training neural network: input a large number of historical scene features X (load curve , real-time electricity price of historical data, etc.), output label Y: initial unit start-stop state, initial output distribution of virtual power plant and unit.

[0209] Through the trained neural network, the graph model is solved and predicted to obtain the initial unit start-stop state , initial output distribution of virtual power plant , initial output distribution of unit .

[0210] S3, based on the initial unit start-stop state, output distribution of virtual power plant and unit, merging is performed according to similar time periods.

[0211] The embodiment reduces the number of optimization time periods by merging similar time periods.

[0212] S31, feature vector construction

[0213] The feature vector of each time period t is constructed , which is represented as:

[0214] ,

[0215] In the formula, is the total load of the system in the time period t , is the number of online units, is the unit set, is the initial start-stop state of the unit in the time period j , is the proportion of traditional power supply, t is the initial output distribution of the unit in the time period ; j t S32, distance calculation

[0216]

[0217] ​Calculate the current time period t distance :

[0218] ,

[0219] In the formula, , In order to be in t During the period a Point and b The point system conforms, , for t Units during the period j exist a Point and b Start / stop status of the point. , These are adjustable weighting coefficients. + ;

[0220] Based on the distance To obtain the similarity score , represented as:

[0221] ,

[0222] In the formula, This is the scaling factor;

[0223] S33, Typical Context Clustering

[0224] Initialize cluster centers:

[0225] ,

[0226] In the formula, K represents the number of clusters, i.e., the number of typical scenarios. For the first k The center of each cluster, For time period t and Similarity of time periods;

[0227] Time slot allocation:

[0228] ,

[0229] In the formula, For the first k Clustering time set, For the first The feature vectors of the cluster centers;

[0230] Virtual power plant participation correction:

[0231] If time period t Virtual power plant participation , then force to be independent cluster , and update , otherwise do nothing, where is the initial net output profile of the VPP in time period t ;

[0232] Update cluster centers:

[0233] ,

[0234] where is the feature vector of the k th cluster center;

[0235] Iterate until convergence: repeat the steps of assigning time periods to updating cluster centers until the cluster assignment no longer changes, resulting in a merged time period set containing k typical time periods and k typical load values for each typical time period .

[0236] S34, constraint validity judgment

[0237] For each time period t and each line l , calculate the original power flow profile:

[0238] ,

[0239] where is the original power flow profile, is the line-node power transfer profile factor, is the set of units connected to bus n , is the initial component output of the VPP at bus n , is the load at bus t in time period n ;

[0240] Identify critical constraint time periods:

[0241] ,

[0242] where is the critical constraint time period, is the transmission line capacity limit;

[0243] Use constraint activation logic to output a constraint activation mapping table to avoid ignoring key line constraints after merging, where the constraint activation mapping is represented as:

[0244] .

[0245] , indicating the line l During typical periods k Constraints need to be activated. By judging the validity of constraints, a constraint activation mapping table is output to identify potentially blocked lines, so as to avoid ignoring critical line constraints after merging.

[0246] This step reduces the problem size while ensuring the rationality and security of the scheduling optimization results by merging similar time periods.

[0247] S4. Based on the merging results, construct and solve the safety-constrained economic dispatch model to obtain the optimal output of the virtual power plant and units, and complete the dispatching process.

[0248] The Safety-Constrained Economic Scheduling (SCED) model includes typical time periods. k The objective function and corresponding constraints are defined as minimizing the operating cost. The objective function is:

[0249] ,

[0250] In the formula, for k Minimum operating cost for a given period of time. For traditional units j The electricity generation cost curve, For traditional units j During typical periods k Those who have made contributions This is the overall quote for VPP. For VPP during typical periods k Net output power, For unit assembly;

[0251] The constraints include:

[0252] Power balance constraints:

[0253] ,

[0254] In the formula, For renewable energy i The predicted output For renewable energy collection, Typical period k Typical load values;

[0255] Unit output constraints:

[0256] ,

[0257] In the formula, For the unit jinitial start-stop state, , minimum and maximum power output of the units j ;

[0258] Virtual power plant constraints:

[0259] (1) Net power range constraint:

[0260] ,

[0261] where, , minimum and maximum net power output of the VPP;

[0262] (2) Energy conservation constraint:

[0263] ,

[0264] ,

[0265] where, energy of the typical time period k , , charge and discharge efficiency, time period length, , upper and lower energy limits;

[0266] Grid security constraints:

[0267] (1) Power flow constraint:

[0268] ,

[0269] where, line power flow, set of grid bus nodes, line-node power transfer distribution factor, set of units connected to bus n , VPP component output at bus n , load value of bus k in the typical time period n ;

[0270] (2) Line capacity limit constraint:

[0271] ,

[0272] where, transmission line capacity limit, set of transmission lines.

[0273] After obtaining a good initial solution, the optimal virtual power plant and 0 / 1 set of unit start-stop plan are obtained by accelerating the speed of calling the pandapower solver to solve the SCED model, so that the virtual power plant participates in the dispatch optimization of the super large-scale power grid.

[0274] Embodiment 2

[0275] The embodiment provides a virtual power plant participating in optimization dispatch system of super large-scale power grid, comprising:

[0276] The data acquisition module is used for acquiring related data of the virtual power plant and the super large-scale power grid, wherein a plurality of units are arranged in the super large-scale power grid.

[0277] The initial state and output acquisition module is used for outputting initial unit start-stop state, initial output distribution of the virtual power plant and the units based on the related data and by using a trained neural network model.

[0278] The merging module is used for merging according to similar time periods based on the initial unit start-stop state and the output distribution of the virtual power plant and the units.

[0279] The dispatch module is used for constructing a security constrained economic dispatch model and solving based on the merging result, obtaining optimal output of the virtual power plant and the units, and completing the dispatch process.

[0280] The rest is as in embodiment 1.

[0281] If the above functions are realized in the form of software function units and sold or used as independent products, they can be stored in a computer readable storage medium. Based on this understanding, the technical solutions of the present application or the parts of the prior art that essentially contribute or the parts of the technical solutions can be embodied in the form of a software product, which is stored in a storage medium and includes a plurality of instructions for causing a computer device (which can be a personal computer, a server, or a network device) to execute all or part of the steps of the method described in the embodiments of the present application. The foregoing storage medium includes a U disk, a mobile hard disk, a read-only memory (ROM, Read-Only Memory), a random access memory (RAM, Random Access Memory), a magnetic disk or an optical disk and various program code storage media.

[0282] Those skilled in the art will appreciate that embodiments of the present application can be readily used as software, hardware, or a combination of software and hardware. In a software embodiment, the methods can be tangibly embodied in a machine-readable storage medium having stored thereon instructions that can be used to program a computer to perform any of the methods. The software implementation can be initialized by loading and executing a set of instructions arranged to perform one of the methods into the computer's memory. Alternatively, hard-wired circuitry can be used in place of, or in combination with, software instructions. Thus, the

[0283] The present application is described in reference to the flowchart and / or block diagrams of the methods, apparatus (systems) and computer program products according to embodiments of the application. It will be understood that each block of the flowchart and / or block diagrams, and combinations of blocks in the flowchart and / or block diagrams, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general purpose computer, special purpose computer, embedded processor, or other programmable data processing apparatus to produce a machine, such that the instructions, which execute via the processor of the computer or other programmable data processing apparatus, create means for implementing the functions specified in the flowchart and / or block diagram block or blocks. Figure 1 one or more functions specified in one or more of the flowchart or block diagram block or blocks. Figure 1 means for performing the function specified by the flowchart or block diagram block or blocks.

[0284] These computer program instructions can also be stored in a computer- readable memory that can direct a computer or other programmable data processing apparatus to function in a particular manner, such that the instructions stored in the computer-readable memory produce an article of manufacture including instructions which implement the flowchart or block diagram block or blocks. Figure 1 one or more functions specified in one or more of the flowchart or block diagram block or blocks. Figure 1 means for performing the function specified by the flowchart or block diagram block or blocks.

[0285] These computer program instructions can also be loaded onto a computer or other programmable data processing apparatus to cause a series of operational steps to be performed on the computer or other programmable apparatus to produce a computer implemented process such that the instructions which execute on the computer or other programmable apparatus provide steps for implementing the flowchart or block diagram block or blocks. Figure 1 one or more functions specified in one or more of the flowchart or block diagram block or blocks. Figure 1 means for performing the function specified by the flowchart or block diagram block or blocks.

[0286] While preferred embodiments of the application have been described, modifications and variations can be apparent to those skilled in the art once aware of the general underlying concepts. Accordingly, the appended claims are intended to embrace all such modifications and variations as fall within the scope of the application.

[0287] It will be apparent to those skilled in the art that various modifications and variations can be made to the present application without departing from the spirit or scope of the application. Thus, it is intended that the present application cover modifications and variations of this application provided they come within the scope of the appended claims and their equivalents.

Claims

1. A method for participating in optimal dispatching of a super large-scale power grid by a virtual power plant, characterized in that, The method comprises the following steps: acquiring relevant data of a virtual power plant and a super large-scale power grid, wherein the super large-scale power grid is provided with multiple units; based on the relevant data, constructing a full-period security-constrained unit commitment model, and solving an initial unit start-stop state, an initial output distribution of the virtual power plant and the units by using a trained neural network model; based on the initial unit start-stop state and the output distribution of the virtual power plant and the units, merging according to similar time periods; based on the merging result, constructing a security-constrained economic dispatch model and solving the same to obtain optimal output of the virtual power plant and the units, and completing the dispatch process; in the training process of the neural network model, the full-period security-constrained unit commitment model is constructed into a graph model, a neural network model is selected and trained based on the structure of the graph model, and a trained neural network model is obtained; the graph model is represented as G=(V,E,W), wherein V is a node set and is represented as: ， In the formula, As a node in the graph, it represents a generator unit. j During the period t The running status, As a graph node, it represents the VPP during the time period. t Net output, As a graph node, serving as a system-level coupling point, it represents power balance. For the assembly of generator units, For time periods; E represents an edge set, including time edges, space edges and constraint edges, wherein the time edges include unit climbing constraint edges ( ) and virtual power plant climbing constraint edges ( ), the space edges include coefficients of unit output in power balance ( ), coefficients of virtual power plant output in power balance ( ) and power flow constraints ( ), , and the constraint edges include minimum running time constraint edges and minimum downtime constraints; W represents a weight set, including node weights and edge weights, wherein the node weights include total costs of conventional units j and overall costs of VPPs , and the edge weights include time edge weights, space edge weights and constraint edge weights; ​ the step of merging according to similar time periods comprises: a) feature vector construction: constructing a feature vector for each time period t is represented as: is represented as: ， In the formula, is the total load of the system in the time period t , is the number of online units, is the unit set, is the unit j in the initial start-stop state in the time period t , is the proportion of traditional power sources, is the unit j initial output distribution in the time period t ; b) distance calculation: computing a current time period t distance of : ， In the formula, , is the system load at the point of time t and the point of time a , b is the system load at the point of time , is the system load at the point of time t , j is the start-stop state of the unit at the point of time a and the point of time b , , is the adjustable weight coefficient + ; based on the distance , obtaining a similarity degree , expressed as: ， In the formula, is a scaling factor; c) typical scenario clustering: initializing a cluster center: ， where K is the number of clusters, i.e. the number of typical scenarios, is the center of the k th cluster, is the similarity between the t period and the th period. allocating a time period: ， In the formula, is the k set of time periods of the cluster, is the feature vector of the k-th cluster center; virtual power plant participation correction: If the time period t of virtual power plant engagement , then force to independent cluster , and update , otherwise do nothing, where is the initial net output distribution of the virtual power plant for the time period t . updating the cluster center: ， In the formula, is the feature vector of the k th cluster center; Iterate until convergence: Repeat the step of assigning time periods to update the cluster centers until the cluster assignments no longer change, resulting in a set of merged time periods k containing k typical time periods and k typical load values for each typical time period .

2. The method of claim 1, wherein, the full-period security-constrained unit commitment model takes minimum total operation cost as an objective, comprises an objective function and corresponding constraint conditions, wherein the objective function is: ， wherein: traditional unit generation cost: ， traditional unit start-up cost: ， virtual power plant cost: ， wherein, is the total cost, is the set of time periods, is the set of units, is the conventional unit j is the total cost, is the generation cost curve of the conventional unit j is the active power output of the unit in the time period j is the start-up status of the unit t in the time period is the start-up cost of the unit j is the start-up action of the unit t in the time period is the overall cost of the VPP, j is the net power output of the VPP in the time period j , t , is the weight coefficient, , t , , , is the set value, , , is the weight coefficient, , is the set value, , is the minimum and maximum net power output of the VPP; the constraint conditions comprise: power balance constraint: ， wherein is a renewable energy source i in a time period t of the output, is a set of renewable energy sources, is in a time period t total system load; traditional unit constraint: 1) output upper and lower limit constraint: ， wherein , are the minimum and maximum power output of the unit j ; 2) ramping constraint: ， In the formula, For the unit j The rate of ascent, The duration of the time period; 3) minimum operation time constraint: ， wherein is the set of units j is the operating state of the set of units t is the minimum operating time of the set of units is the set of units j is the operating state of the set of units is the start-up variable of the set of units is the set of units j is the start-up variable of the set of units t is the start-up variable of the set of units 4) minimum shutdown time constraint: ， In the formula, For the unit j During the period t Minimum downtime, For the unit j During the period t The shutdown variable; 5) start-stop logic constraint: ， virtual power plant constraint: 1) net output range constraint: ， wherein , Pmin, max are the minimum and maximum net power output of the VPP; 2) ramping constraint: ， In the formula, is the ramp rate for VPP; 3) energy conservation constraint: ， ， In the formula, is the time period t is the energy, , is the charge and discharge efficiency, is the time period length, , is the energy upper and lower limit; grid security constraint: 1) power flow constraint: ， In the formula, For line flow, For the set of power grid bus nodes, The line-node power transfer distribution factor. For connecting busbars n A collection of generator units, For VPP on the bus n The weight of the output, For time period t, within the bus n The load value; 2) line capacity limit constraint: ， wherein is a transmission line capacity limit, is a transmission line set. 3.The method of claim 1, wherein, the neural network model is a CNN model or an RNN model.

4. The method of claim 1, wherein, The step of merging according to similar time periods further comprises: by performing constraint effectiveness judgment, outputting a constraint activation mapping table to avoid ignoring key line constraints after merging, wherein the specific steps of outputting the constraint activation mapping table comprise: For each time interval t and each line l the original power flow distribution is calculated: ， In the formula, For the original tidal current distribution, The line-node power transfer distribution factor. For connecting busbars n A collection of generator units, For VPP on the bus n The initial component of the output force, For time period t Internal busbar n The load; identifying a key constraint time period: ， In the formula, is a critical constraint period, is a transmission line capacity limit; by using constraint activation logic, outputting a constraint activation mapping table to avoid ignoring key line constraints after merging, wherein the constraint activation mapping table is represented as: 。 5. The method of claim 1, wherein, The security-constrained economic dispatch model includes an objective function and corresponding constraint conditions, wherein the objective function is expressed as a typical time period k The minimum operating cost is targeted, expressed as: ， wherein is the minimum operating cost for a typical period k , is the generation cost curve for a conventional unit j , is the active power output of a conventional unit j for a typical period k , is the overall offer of the VPP is the net power output of the VPP for a typical period k , is the set of units; constraint conditions comprise: power balance constraint: ， wherein the predicted output of a renewable energy source i the predicted output of a renewable energy source the set of renewable energy sources the typical load value of a typical time period k the typical load value of a typical time period unit output constraint: ， wherein is the initial start-stop state of the unit j , , is the minimum, maximum power of the unit j ; virtual power plant constraint: (1) net output range constraint: ， wherein , Pmin, Pmax are the minimum, maximum net power output of the VPP; (2) energy conservation constraint: ， ， In the formula, Energy for a typical period k , , Charge-discharge efficiency Period length , Upper and lower limits of energy grid security constraint: (1) power flow constraint: ， wherein, is the line flow, is the set of grid bus nodes, is the line-to-node power transfer distribution factor, is the set of generators connected to bus n is the component output of the VPP at bus is the load value at bus n for a typical time period, is the load value at bus k for a typical time period, n is the load value at bus (2) line capacity limit constraint: ， In the formula, is a transmission line capacity limit, is a set of transmission lines.

6. The method of claim 1, wherein, the security-constrained economic dispatch model is solved by calling a pandapower solver.

7. The scheduling system for the method of optimal scheduling of a virtual power plant participating in a super large-scale power grid according to any one of claims 1-6, characterized in that, The method comprises the following steps: a data acquisition module is used to acquire relevant data of a virtual power plant and a super large-scale power grid, wherein the super large-scale power grid is provided with multiple units; an initial state and output acquisition module is used to output an initial unit start-stop state, an initial output distribution of the virtual power plant and the units by using a trained neural network model based on the relevant data; a merging module is used to merge according to similar time periods based on the initial unit start-stop state and the output distribution of the virtual power plant and the units; The scheduling module is used for constructing a security-constrained economic dispatch model and solving the model based on the merging result to obtain optimal outputs of the virtual power plant and the units and complete the scheduling process.

Images