A virtual power plant resource classification and aggregation method, system, device and medium

Abstract

The application belongs to the technical field of power grid optimal operation, and discloses a virtual power plant resource classification and aggregation method, system, device and medium, the method comprising the following steps: obtaining virtual power plant distributed resources, obtaining distributed resource features by performing feature extraction on the virtual power plant distributed resources; and constructing a numerical representation model of the distributed resource features according to the distributed resource features; performing normalization processing on the distributed resource features by using a normalization method; classifying the distributed resources by using a clustering algorithm and a second-order Minkowski distance; performing Minkowski summation on distributed resources of the same type to obtain feasible regions of multiple types of distributed resources, and determining a feasible region of the total power of the virtual power plant according to the feasible regions of the multiple types of distributed resources. The application can equivalently convert a large amount of resources in the virtual power plant into a few different types of resources, thereby reducing the difficulty of solving the aggregation method based on vertex search.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of optimal operation of power grid, and particularly relates to a virtual power plant resource classification and aggregation method, system, device and medium. BACKGROUND

[0002] At present, new energy such as wind and light is booming, and the installed capacity is increasing year by year, but the random, intermittent and fluctuation characteristics of new energy pose a challenge to the safety of power grid. The uncertainty of the power system is gradually increasing, while the flexibility resource is gradually decreasing. With the construction of new-type power system, the number and types of distributed flexible resources in the distribution network are gradually increasing. Through the virtual power plant technology, the unified management of distributed resources can be realized, and the available regulation capacity can be provided for the system.

[0003] Among them, the research on the distributed resource feasible region aggregation method in the virtual power plant mainly focuses on the description of the active and reactive power feasible region, the feasible region description under uncertain environment, the feasible region aggregation of multi-energy and the virtual power plant dispatching based on the feasible region aggregation. For the aggregation method, including the Minkowski summation and the feasible region projection method. However, the above research does not consider the influence of the difference of the heterogeneous resources in the virtual power plant, and the different types of distributed resources are uniformly modeled, and the massive heterogeneous distributed resources are mechanically aggregated to obtain the external feasible region, which has the defects of low aggregation precision and long solving time.

[0004] Therefore, how to provide a virtual power plant resource classification and aggregation method, system, device and medium is a problem to be solved at present. SUMMARY

[0005] The embodiments of the present application provide a virtual power plant resource classification and aggregation method, system, device and medium to solve the problems in the prior art.

[0006] To have a basic understanding of some aspects of the disclosed embodiments, a brief summary is given below. This part is not a general review, nor is it intended to determine the key / important elements or delineate the protection scope of these embodiments. Its only purpose is to present some concepts in a simple form as a prelude to the detailed description that follows.

[0007] According to a first aspect of the embodiments of the present application, a virtual power plant resource classification and aggregation method is provided.

[0008] In one embodiment, the virtual power plant resource classification and aggregation method comprises the following steps:

[0009] The virtual power plant distributed resources are obtained, the distributed resource features are obtained by feature extraction on the virtual power plant distributed resources, and the numerical representation model of the distributed resource features is constructed according to the distributed resource features;

[0010] The numerical characterization model based on the distributed resource features normalizes the distributed resource features by using a normalization method, and classifies the distributed resources by using a clustering algorithm and a second-order Minkowski distance according to the normalization result, to obtain different types of distributed resources.

[0011] The Minkowski summation is performed on the distributed resources of the same type by using an aggregation method of vertex search, to obtain the feasible region of the multiple types of distributed resources, and the feasible region of the aggregated total power of the virtual power plant is determined according to the feasible region of the multiple types of distributed resources.

[0012] In one embodiment, the distributed resource features include a resource type, a rated capacity, a production and sales characteristic, a response type, an advance notification time, a response adjustment time, a response duration, a response capacity, and a response period.

[0013] In one embodiment, the numerical characterization model based on the distributed resource features normalizes the distributed resource features by using a normalization method, and classifies the distributed resources by using a clustering algorithm and a second-order Minkowski distance according to the normalization result, to obtain different types of distributed resources, including the following steps:

[0014] The distributed resource features are normalized by using a min-max normalization method according to the numerical characterization model of the distributed resource features, to obtain a normalization result.

[0015] The second-order Minkowski distance between different distributed resources is calculated by using a second-order Minkowski distance according to the normalization result.

[0016] The distributed resources are classified by using a K-means clustering algorithm according to the second-order Minkowski distance between the distributed resources, to obtain different types of distributed resources.

[0017] In one embodiment, the calculation formula of the second-order Minkowski distance between different distributed resources calculated by using a second-order Minkowski distance according to the normalization result is as follows:

[0018]

[0019] In the formula, di,j represents the second-order Minkowski distance between the ith distributed resource and the jth distributed resource, Cn represents the total number of the distributed resource features, i,j wherein, di,j represents the second-order Minkowski distance between the ith distributed resource and the jth distributed resource, Cn represents the total number of the distributed resource features, wherein, di,j represents the second-order Minkowski distance between the ith distributed resource and the jth distributed resource, Cn represents the total number of the distributed resource features, wherein, di,j represents the second-order Minkowski distance between the ith distributed resource and the jth distributed resource, Cn represents the total number of the distributed resource features,

[0020] In one embodiment, the K-means clustering algorithm is used to classify the distributed resources according to the second-order Minkowski distance between the distributed resources, and different types of distributed resources are obtained, including the following steps:

[0021] Randomly select several data points from the virtual power plant distributed resources as cluster centers, and according to the Minkowski distance between different distributed resources, each distributed resource is assigned to the resource category represented by the nearest cluster center to form a preliminary clustering result;

[0022] For each resource category, calculate the average value of all distributed resources in the resource category, and take the average value of the distributed resources as a new cluster center;

[0023] Compare the change value of the cluster center. If the change value of the cluster center is less than a set threshold, output the clustering result of the distributed resources to obtain different types of distributed resources, otherwise, recalculate the Minkowski distance between different distributed resources, and update the cluster center according to the recalculated result until the change value of the cluster center is less than the set threshold, stop updating, and output the clustering result of the distributed resources to obtain different types of distributed resources.

[0024] In one embodiment, the aggregation method of vertex search is used to sum the same type of distributed resources by Minkowski, obtain the feasible region of multiple types of distributed resources, and determine the feasible region of the virtual power plant aggregation total power according to the feasible region of multiple types of distributed resources, including the following steps:

[0025] The feasible region of different types of distributed resources is determined by the chino polyhedron method, and the total feasible region of the same type of distributed resources is calculated by the Minkowski summation method;

[0026] Based on the vertex search method, the feasible regions of multiple types of distributed resources are aggregated, and the vertices of the feasible regions of the distributed resources are iteratively searched. According to the vertices of the feasible regions of multiple types of distributed resources, the feasible region of the virtual power plant aggregation total power is calculated by the feasible region calculation formula of the virtual power plant aggregation total power.

[0027] In one embodiment, the aggregation method of vertex search is used to aggregate the feasible regions of multiple types of distributed resources, and the vertices of the feasible regions of the distributed resources are iteratively searched. According to the vertices of the feasible regions of multiple types of distributed resources, the feasible region of the virtual power plant aggregation total power is calculated by the feasible region calculation formula of the virtual power plant aggregation total power, including the following steps:

[0028] Randomly select a type of distributed resource as a starting point, and take the total feasible region of the type of distributed resource as an initial convex hull;

[0029] The outer normal vector of the hyperplane of the initial convex hull is taken as a search direction, and the overall feasible region vertex of the distributed resource is searched through iteration based on a vertex search model;

[0030] When the overall feasible region vertex of the vertex distributed resource reaches a preset threshold, the feasible region of the aggregated total power of the virtual power plant is calculated through a feasible region calculation formula of the aggregated total power of the virtual power plant.

[0031] In one embodiment, the expression of the vertex search model is:

[0032] maxη Tran P total

[0033]

[0034] In the formula, max represents the maximum value, η represents the search direction, P total represents the vertex of the feasible region of the aggregated total power of the virtual power plant, represents the aggregated power of the kth type of distributed resource, and K represents the number of distributed resource types, represents the feasible region of the aggregated power of the kth type of distributed resource.

[0035] In one embodiment, the feasible region calculation formula of the aggregated total power of the virtual power plant is:

[0036]

[0037] In the formula, F VPP represents the feasible region of the aggregated total power of the virtual power plant, represents the feasible region vertex, and α q represents the convex combination coefficient, N vpp represents the number of vertices of the virtual power plant feasible region, and q represents the qth vertex.

[0038] According to a second aspect of the embodiment of the present application, a virtual power plant resource classification and aggregation system is provided.

[0039] In one embodiment, a virtual power plant resource classification and aggregation system comprises:

[0040] A data processing module is configured to obtain virtual power plant distributed resources, extract features of the virtual power plant distributed resources to obtain distributed resource features, and construct a numerical representation model of the distributed resource features according to the distributed resource features.

[0041] The resource classification module is configured to normalize the distributed resource features by using a normalization method based on the numerical representation model of the distributed resource features, and classify the distributed resources by using a clustering algorithm and a second-order Minkowski distance based on the normalized results, so as to obtain distributed resources of different types.

[0042] The feasible region determination module is configured to perform Minkowski summation on the distributed resources of the same type by using an aggregation method of vertex search, obtain the feasible regions of the distributed resources of multiple types, and determine the feasible region of the aggregated total power of the virtual power plant based on the feasible regions of the distributed resources of the multiple types.

[0043] In an embodiment, the distributed resource features include resource type, rated capacity, production and consumption characteristics, response type, advance notification time, response adjustment time, response duration, response capacity, and response period.

[0044] In an embodiment, when the resource classification module normalizes the distributed resource features by using a normalization method based on the numerical representation model of the distributed resource features, and classifies the distributed resources by using a clustering algorithm and a second-order Minkowski distance based on the normalized results, so as to obtain distributed resources of different types, the resource classification module can normalize the distributed resource features by using a min-max normalization method based on the numerical representation model of the distributed resource features, so as to obtain the normalized results; the resource classification module can calculate the second-order Minkowski distance between different distributed resources based on the normalized results; and the resource classification module can classify the distributed resources based on a K-means clustering algorithm and the second-order Minkowski distance between the distributed resources, so as to obtain distributed resources of different types.

[0045] In an embodiment, when the resource classification module classifies the distributed resources based on the K-means clustering algorithm and the second-order Minkowski distance between the distributed resources, so as to obtain distributed resources of different types, the resource classification module can randomly select a plurality of data points from the distributed resources of the virtual power plant as clustering centers, and assign each distributed resource to a resource category represented by the clustering center closest to the distributed resource based on the Minkowski distance between the distributed resources, so as to form a preliminary clustering result.

[0046] In an embodiment, when the resource classification module classifies the distributed resources based on the K-means clustering algorithm and the second-order Minkowski distance between the distributed resources, so as to obtain distributed resources of different types, the resource classification module can randomly select a plurality of data points from the distributed resources of the virtual power plant as clustering centers, and assign each distributed resource to a resource category represented by the clustering center closest to the distributed resource based on the Minkowski distance between the distributed resources, so as to form a preliminary clustering result.

[0047] In an embodiment, when the resource classification module classifies the distributed resources based on the K-means clustering algorithm and the second-order Minkowski distance between the distributed resources, so as to obtain distributed resources of different types, the resource classification module can randomly select a plurality of data points from the distributed resources of the virtual power plant as clustering centers, and assign each distributed resource to a resource category represented by the clustering center closest to the distributed resource based on the Minkowski distance between the distributed resources, so as to form a preliminary clustering result.

[0048] The change value of the clustering center is compared, if the change value of the clustering center is less than a set threshold value, the clustering result of the distributed resources is output, different types of distributed resources are obtained, otherwise, the Minkowski distance between different distributed resources is recalculated, and the clustering center is updated according to the recalculated result, until the change value of the clustering center is less than the set threshold value, the updating is stopped, and the clustering result of the distributed resources is output, different types of distributed resources are obtained.

[0049] In one embodiment, when the feasible region determination module determines the feasible region of the virtual power plant aggregation total power through the aggregation method of vertex search, Minkowski summation is performed on the distributed resources of the same type, the feasible region of the distributed resources of multiple types is obtained, and the feasible region of the virtual power plant aggregation total power is determined according to the feasible region of the distributed resources of multiple types, the feasible region of the distributed resources of different types can be determined through the Zonotope method, and the total feasible region of the distributed resources of the same type can be calculated through the Minkowski summation method; the vertex search method is used to aggregate the feasible regions of the distributed resources of multiple types, the vertex of the feasible region of the distributed resources is searched through iteration, and the feasible region of the virtual power plant aggregation total power is calculated through the feasible region calculation formula of the virtual power plant aggregation total power according to the vertex of the feasible region of the distributed resources of multiple types.

[0050] According to a third aspect of the embodiments of the present application, a computer device is provided.

[0051] In some embodiments, the computer device comprises a memory and a processor, the memory stores a computer program, and the processor implements the steps of the above method when executing the computer program.

[0052] According to a fourth aspect of the embodiments of the present application, a computer readable storage medium is provided.

[0053] In one embodiment, the computer readable storage medium stores a computer program, and the computer program is executed by a processor to implement the steps of the above method.

[0054] The technical solutions provided by the embodiments of the present application can include the following beneficial effects:

[0055] The present application extracts features of distributed resources, classifies the distributed resources based on K-meas and Minkowski distance, realizes efficient aggregation of the feasible region of different types of distributed resources, adopts the aggregation method based on Zonotope for the same type of distributed resources to reduce the solving time and improve the solving accuracy, can equivalent a large amount of resources in the virtual power plant to a few different types of resources, and reduces the solving difficulty for the aggregation method based on vertex search.

[0056] It should be understood that the above general description and the following detailed description are only exemplary and explanatory, and cannot limit the present application. BRIEF DESCRIPTION OF DRAWINGS

[0057] The accompanying drawings, which are incorporated herein and constitute part of this specification, illustrate embodiments consistent with the application and, together with the description, further serve to explain the principles of the application.

[0058] Figure 1 is a flow chart of a virtual power plant resource classification and aggregation method according to an example embodiment;

[0059] Figure 2 is a principle block diagram of a virtual power plant resource classification and aggregation system according to an example embodiment;

[0060] Figure 3 is a flow chart of distributed resource clustering in a virtual power plant resource classification and aggregation method according to an example embodiment;

[0061] Figure 4 is a schematic diagram of a chino polyhedron based distributed resource feasible region in a virtual power plant resource classification and aggregation method according to an example embodiment;

[0062] Figure 5 is a structural schematic diagram of a computer device according to an example embodiment. DETAILED DESCRIPTION

[0063] The following description and drawings are illustrative of specific embodiments of the present document and are not intended to be limiting of the present document. Particular embodiments of the present document can include any or all of the features discussed herein. Particular embodiments of the present document can include any or all of the features discussed herein. The description herein can be used to explain principles of various embodiments of the present document. Particular embodiments of the present document can be implemented in the following modes:

[0064] The terms "longitudinal", "lateral", "upper", "lower", "front", "rear", "left", "right", "vertical", "horizontal", "top", "bottom", "inner", "outer" and the like as used herein to indicate orientation or positional relationships based on the orientations or positional relationships shown in the drawings, are for purposes of this description only, and are not intended to indicate or imply that the device or element referred to must have a particular orientation, be constructed and operated in a particular orientation, and therefore should not be interpreted in that way unless specifically stated and limited herein.

[0065] In this document, the term "plurality" means two or more, unless otherwise specified and limited.

[0066] In this document, the character " / " means that the front and rear objects are in an "or" relationship. For example, A / B means: A or B.

[0067] In this document, the term "and / or" is a description of the association relationship of the object, which means that there can be three relationships. For example, A and / or B means: A or B, or, A and B, three relationships.

[0068] It should be understood that although each step in the flowchart is shown in sequence according to the direction of the arrow, these steps are not necessarily executed in sequence according to the direction of the arrow. Unless otherwise specified herein, the execution of these steps is not strictly limited in sequence, and these steps can be executed in other order. Moreover, at least part of the steps in the figure can include multiple sub-steps or multiple stages, which are not necessarily executed at the same time, but can be executed at different times, and the execution order of these sub-steps or stages is not necessarily sequential, but can be executed in rotation or alternation with other steps or sub-steps or stages of other steps.

[0069] Each module in the device or system of the present application can be realized by software, hardware and their combination in whole or in part. The above modules can be embedded in or independent of the processor in the computer device in hardware form, or stored in the memory in the computer device in software form, so as to call and execute the operations corresponding to the above modules by the processor.

[0070] In the case of no conflict, the embodiments in the present application and the features in the embodiments can be combined with each other.

[0071] Figure 1An embodiment of the virtual power plant resource classification and aggregation method is shown.

[0072] Specifically, by proposing a classification method for distributed resources in a virtual power plant, clustering is performed according to the feasible region characteristics; secondly, based on the feasible region characteristics, a high-precision Zonotope is determined to approximate the feasible region; finally, based on the mixed method of Minkowski sum and vertex search, the feasible regions of multiple distributed resources are added to obtain the feasible region of the aggregated virtual power plant.

[0073] In this optional embodiment, the virtual power plant resource classification and aggregation method comprises the following steps:

[0074] In step S101, the virtual power plant distributed resources are obtained, the distributed resource features are obtained by feature extraction of the virtual power plant distributed resources, and a numerical representation model of the distributed resource features is constructed according to the distributed resource features;

[0075] In step S103, the distributed resource features are normalized by using a normalization method based on the numerical representation model of the distributed resource features, and the distributed resources are classified by using a clustering algorithm and a second-order Minkowski distance according to the normalization processing result, to obtain distributed resources of different types;

[0076] In step S105, the Minkowski sum of the distributed resources of the same type is obtained by using the aggregation method of vertex search, the feasible regions of multiple types of distributed resources are obtained, and the feasible region of the total power of the virtual power plant is determined according to the feasible regions of the multiple types of distributed resources.

[0077] In this optional embodiment, the distributed resource features include resource type, rated capacity, production and sales characteristics, response type, advance notification time, response adjustment time, response duration, response capacity, and response period.

[0078] The resource type includes photovoltaic, energy storage, electric vehicles, industrial loads, commercial loads, and residential loads. The rated capacity is the maximum power of the resource. The production and sales characteristics include three types of electricity, power generation, and producers and sellers. The response type includes direct response and market response. The advance notification time is the minimum advance time for the distributed resource to make corresponding preparations. The response adjustment time is the time for the distributed resource to complete the response after the virtual power plant issues an instruction. The response duration is the time for the distributed resource to maintain the response to the instruction. The response capacity is the maximum power and energy response that the distributed resource can achieve. The response period is the period during which the distributed resource can provide a response.

[0079] Specifically, the features of the i-th distributed resource are as follows:

[0080]

[0081] wherein, respectively represent resource type, rated capacity, production and consumption characteristics, response type, advance notice time, response adjustment time, response duration, response capacity, response period, and I represents the total number of distributed resources.

[0082] In addition, it should be noted that a numerical representation model of distributed resource characteristics is constructed, and the above characteristics are represented in numerical form to facilitate subsequent classification processing.

[0083] For resource type, photovoltaic is converted to 1, energy storage is converted to 2, electric vehicle is converted to 3, industrial load is converted to 4, commercial load is converted to 5, and residential load is converted to 6. For production and consumption characteristics, producers are converted to -1, consumers are converted to 1, and producers and consumers are converted to 0. For response type, market response is converted to 0, and direct response is converted to 1.

[0084] For rated capacity, advance notice time, response adjustment time, response duration, response capacity, and response period, their original values are kept.

[0085] In this optional embodiment, based on the numerical representation model of distributed resource characteristics, the distributed resource characteristics are normalized by using a normalization method; and according to the normalization processing result, the distributed resources are classified by using a clustering algorithm and a second-order Minkowski distance, so that different types of distributed resources are obtained. According to the numerical representation model of distributed resource characteristics, the distributed resource characteristics are normalized by using a min-max normalization method to obtain a normalization processing result; according to the normalization processing result, the second-order Minkowski distance between different distributed resources is calculated by using a second-order Minkowski distance; based on a K-means clustering algorithm, the distributed resources are classified according to the second-order Minkowski distance between the distributed resources, so that different types of distributed resources are obtained.

[0086] Specifically, the above characteristics are normalized by using a min-max method. Based on a second-order Minkowski distance, the similarity between different distributed resources is defined.

[0087] The expression for normalization by using a min-max method is:

[0088]

[0089] wherein, represents the jth feature of the ith distributed resource, represents the normalized value of the jth feature of the ith distributed resource, represents the minimum value of the jth feature, represents the maximum value of the jth feature.

[0090] The K-means method clusters the different distributed resources in the virtual power plant by analyzing the similarity between the different distributed resources, and the distributed resources with high similarity are classified into one category, thereby realizing automatic classification of the distributed resources based on the response characteristics. The distributed resources are classified into K categories based on the above method. The flowchart is shown in FIG. 2. Figure 3

[0091] In the optional embodiment, the calculation formula of the second-order Minkowski distance between the different distributed resources is as follows according to the normalized processing result:

[0092]

[0093] wherein di,j represents the second-order Minkowski distance between the ith distributed resource and the jth distributed resource, Cn represents the total number of the distributed resource characteristics, i,j Cic represents the cth distributed resource characteristic value of the ith distributed resource, Cjc represents the cth distributed resource characteristic of the jth distributed resource.

[0094] In the optional embodiment, when the K-means clustering algorithm is used to classify the distributed resources based on the second-order Minkowski distance between the distributed resources, a plurality of data points are randomly selected as the clustering centers from the distributed resources in the virtual power plant, and each distributed resource is allocated to the resource category represented by the nearest clustering center based on the Minkowski distance between the different distributed resources, thereby forming a preliminary clustering result. The average value of all the distributed resources in each resource category is calculated, and the average value of the distributed resources is used as the new clustering center. The change value of the clustering center is compared, and if the change value of the clustering center is less than the set threshold value, the clustering result of the distributed resources is outputted, thereby obtaining the different types of distributed resources. Otherwise, the Minkowski distance between the different distributed resources is recalculated, and the clustering center is updated based on the recalculated result until the change value of the clustering center is less than the set threshold value, and the updating is stopped. The clustering result of the distributed resources is outputted, thereby obtaining the different types of distributed resources.

[0095] ​​In the optional embodiment, in the aggregation method through vertex search, the same type of distributed resources are Minkowski summed to obtain the feasible region of multiple types of distributed resources, and the feasible region of the virtual power plant aggregation total power is determined according to the feasible region of multiple types of distributed resources. The feasible region of different types of distributed resources can be determined through the zonotope method, and the total feasible region of the same type of distributed resources can be calculated through the Minkowski summation method. Based on the vertex search method, the feasible regions of multiple types of distributed resources are aggregated, and the vertices of the feasible regions of distributed resources are searched through iteration. The feasible region of the virtual power plant aggregation total power is calculated through the feasible region calculation formula of the virtual power plant aggregation total power according to the vertices of the feasible regions of multiple types of distributed resources.

[0096] In the optional embodiment, in the aggregation method through vertex search, the feasible regions of multiple types of distributed resources are aggregated based on the vertex search method, the vertices of the feasible regions of distributed resources are searched through iteration, and the feasible region of the virtual power plant aggregation total power is calculated through the feasible region calculation formula of the virtual power plant aggregation total power according to the vertices of the feasible regions of multiple types of distributed resources. A type of distributed resource can be randomly selected as a starting point, and the total feasible region of the distributed resource of this type can be taken as an initial convex hull. The outer normal vector of the hyperplane of the initial convex hull is taken as a search direction, and the vertices of the total feasible region of the distributed resource are searched through iteration based on the vertex search model. When the vertices of the total feasible region of the distributed resource reach a pre-set threshold, the feasible region of the virtual power plant aggregation total power is calculated through the feasible region calculation formula of the virtual power plant aggregation total power.

[0097] In addition, it should be noted that the zonotope-based distributed resource feasible region approximation method can reduce the irregularity of the distributed resource feasible region, thereby reducing the difficulty of subsequent feasible region aggregation. The zonotope-based distributed resource feasible region diagram is shown in Figure 4 , and the representation method is as follows:

[0098]

[0099] P i G = [P i G(1) , P i G(2) , …, P i G(M) ] ∈ R T×M ;

[0100] ||P i G(m) ||2=1m=1,2,…,M;

[0101] λ i = [λ i,1 , λi,2 ,…,λ i,M ] Tran ;

[0102] F i represents the active power feasible region of the ith distributed resource in T time periods, T represents the number of time periods, P i c represents the center point of the feasible region of the ith distributed resource, P i G represents the direction matrix of the feasible region extending in various directions from the center point, which is composed of M direction vectors P i G(m) , the modulus of the direction vector is kept as 1, and λ represents the extension length of the distributed resource feasible region in various directions, represents the maximum value of the extension length, P i represents the active power of the ith distributed resource, R T represents a T-dimensional real number space, R T×M represents a T*M-dimensional real number space, λ i,M represents the extension length of the ith distributed resource in M directions, m represents the extension vector count, and Tran represents the matrix transpose.

[0103] Different types of distributed resources have different characteristics of feasible regions, corresponding to different Zino polyhedrons. The selection method of the Zino polyhedron is as follows:

[0104] For resources without time period coupling characteristics, the characteristics mainly manifest as single time period upper and lower power constraints. Therefore, the direction vector is selected as follows:

[0105]

[0106] , wherein represents the extension vector of the mth direction of the resource without time period coupling characteristics, T represents the maximum number of m values, and Tran represents the matrix transpose.

[0107] For energy storage type distributed resources, the feasible region also has multi-time period power sum upper and lower limit constraints. Therefore, the direction vector is selected as follows:

[0108]

[0109] , wherein represents the extension vector of the mth direction of the energy storage type resource, represents the extension vector of the T+mth direction of the energy storage type resource, Tran represents the matrix transpose, and T represents the maximum number of m values.

[0110] For the resources of the ramp constraint type, the feasible region also has upper and lower constraints on the difference between the power in different time periods. Therefore, the direction vector is selected as follows:

[0111]

[0112] In the formula, denotes the extension vector of the mth direction of the ramp resource, denotes the extension vector of the T+mth direction of the ramp resource, Tran denotes the matrix transpose, and T denotes the maximum value of m.

[0113] The Minkowski sum of each type of distributed resource is obtained, that is, the total feasible region of each type of distributed resource, and the formula is as follows:

[0114]

[0115] In the formula, denotes the center point of the aggregated feasible region, P i c denotes the center point of the ith distributed resource feasible region, denotes the extension length in each direction after aggregation, λ i denotes the extension length of the ith distributed resource in each direction, I k denotes the total number of the kth type of distributed resource.

[0116] The essence of the algorithm is to approach the virtual power plant feasible region from the inside to the outside, search for a part of the vertices to obtain a conservative approximation of the convex hull, use the outer normal vector of the hyperplane of the convex hull as the search direction, and enumerate new vertices in each iteration until the accuracy requirement is met. When the number of vertices reaches the set threshold, the feasible region of the total power of the virtual power plant is obtained.

[0117] In the optional embodiment, the expression of the vertex search model is as follows:

[0118] maxη Tran P total

[0119]

[0120] In the formula, max denotes the maximum value, η denotes the search direction, P total denotes the vertex of the aggregated total power feasible region of the virtual power plant, denotes the aggregated power of the kth type of distributed resource, and K denotes the number of distributed resource types, denotes the feasible region of the aggregated power of the kth type of distributed resource.

[0121] In the optional embodiment, the feasible region calculation formula of the virtual power plant aggregated total power is:

[0122]

[0123] In the formula, F VPP represents the feasible region of the virtual power plant aggregated total power, represents the vertex of the feasible region, and q represents the convex combination coefficient, N vpp represents the number of vertices of the virtual power plant feasible region, and q represents the qth vertex.

[0124] In summary, the method has the advantages of fast solving speed and high solving precision for the feasible region aggregation of a large number of distributed resources in the virtual power plant.

[0125] Figure 2 An embodiment of a virtual power plant resource classification aggregation system of the application is shown.

[0126] In the optional embodiment, the virtual power plant resource classification aggregation system comprises:

[0127] The data processing module 201 is configured to acquire the virtual power plant distributed resources, extract features of the virtual power plant distributed resources to obtain distributed resource features, and construct a numerical representation model of the distributed resource features according to the distributed resource features.

[0128] The resource classification module 203 is configured to normalize the distributed resource features by using a normalization method based on the numerical representation model of the distributed resource features, and classify the distributed resources by using a clustering algorithm and a second-order Minkowski distance according to the normalization processing result to obtain distributed resources of different types.

[0129] The feasible region determination module 205 is configured to perform Minkowski summation on the distributed resources of the same type by using an aggregation method of vertex search to obtain feasible regions of the distributed resources of multiple types, and determine the feasible region of the virtual power plant aggregated total power according to the feasible regions of the distributed resources of multiple types.

[0130] In the optional embodiment, the distributed resource features comprise resource type, rated capacity, production and sales characteristics, response type, advance notice time, response adjustment time, response duration, response capacity, and response period.

[0131] In the optional embodiment, the resource classification module normalizes the distributed resource features based on the numerical representation model of the distributed resource features, and classifies the distributed resources based on the normalized results, the clustering algorithm and the second-order Minkowski distance, to obtain different types of distributed resources. When the numerical representation model of the distributed resource features is obtained, the min-max normalization method is used to normalize the distributed resource features, to obtain the normalized results. The second-order Minkowski distance between different distributed resources is calculated based on the normalized results. The K-means clustering algorithm is used to classify the distributed resources based on the second-order Minkowski distance between the distributed resources, to obtain different types of distributed resources.

[0132] In the optional embodiment, the K-means clustering algorithm is used to classify the distributed resources based on the second-order Minkowski distance between the distributed resources, to obtain different types of distributed resources, which includes: randomly selecting a plurality of data points as clustering centers from the virtual power plant distributed resources, and distributing each distributed resource to the resource category represented by the nearest clustering center based on the Minkowski distance between different distributed resources, to form a preliminary clustering result. The average value of all distributed resources in each resource category is calculated, and the average value of the distributed resources is used as a new clustering center. The change value of the clustering center is compared. If the change value of the clustering center is less than a set threshold, the clustering result of the distributed resources is output, to obtain different types of distributed resources. Otherwise, the Minkowski distance between different distributed resources is recalculated, and the clustering center is updated based on the recalculated results, until the change value of the clustering center is less than the set threshold, the updating is stopped, and the clustering result of the distributed resources is output, to obtain different types of distributed resources.

[0133] In the optional embodiment, the feasible region determination module uses the vertex search aggregation method to perform Minkowski summation on the same type of distributed resources, to obtain the feasible regions of a plurality of types of distributed resources, and determines the feasible region of the virtual power plant aggregated total power based on the feasible regions of the plurality of types of distributed resources. The feasible regions of different types of distributed resources can be determined by the chino polyhedron method, and the total feasible region of the same type of distributed resources can be calculated by the Minkowski summation method. The vertex search method is used to aggregate the feasible regions of a plurality of types of distributed resources, and the vertex of the distributed resource feasible region is searched by iteration. The vertex of the feasible region of the plurality of types of distributed resources is used to calculate the feasible region of the virtual power plant aggregated total power by the formula of the virtual power plant aggregated total power.

[0134] In one embodiment, a computer device is provided, which may be a server, and its internal structure diagram may be as follows: Figure 5 As shown, the computer device includes a processor, memory, and a network interface connected via a system bus. The processor provides computing and control capabilities. The memory includes a non-volatile storage medium and internal memory. The non-volatile storage medium stores an operating system, computer programs, and a database. The internal memory provides an environment for the operation of the operating system and computer programs in the non-volatile storage medium. The database stores static and dynamic information data. The network interface communicates with external terminals via a network connection. When the computer program is executed by the processor, it implements the steps in the above method embodiments.

[0135] Those skilled in the art will understand that Figure 5 The structure shown is merely a block diagram of a portion of the structure related to the present invention and does not constitute a limitation on the computer device to which the present invention is applied. A specific computer device may include more or fewer components than those shown in the figure, or combine certain components, or have different component arrangements.

[0136] In addition, the present invention also provides a computer device, including a memory and a processor, wherein the memory stores a computer program, and the processor executes the computer program to implement the steps in the above method embodiments.

[0137] In addition, the present invention also provides a computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, implements the steps in the above method embodiments.

[0138] Those skilled in the art can understand that all or part of the processes in the above-mentioned embodiment methods can be completed by instructing relevant hardware through a computer program. The computer program can be stored in a non-volatile computer readable storage medium, and when executed, can include the processes of the above-mentioned embodiments. Any reference to memory, storage, database or other medium used in each embodiment of the present application can include at least one of non-volatile and volatile memory. Non-volatile memory can include read-only memory (ROM), magnetic tape, floppy disk, flash memory or optical memory, etc. Volatile memory can include random access memory (RAM) or external cache memory. As an illustration but not limitation, RAM can be in various forms, such as static random access memory (SRAM) or dynamic random access memory (DRAM), etc.

[0139] The present application is not limited to the structures described above and shown in the drawings, and various modifications and changes can be made without departing from the scope thereof. The scope of the present application is only limited by the appended claims.

Claims

1. A virtual power plant resource classification and aggregation method, characterized in that, The method comprises the following steps: The virtual power plant distributed resources are obtained, and the distributed resource features are obtained by feature extraction on the virtual power plant distributed resources; A numerical representation model of the distributed resource features is constructed according to the distributed resource features; The distributed resource features are normalized by using a normalization method based on the numerical representation model of the distributed resource features; Different types of distributed resources are obtained by classifying the distributed resources by using a clustering algorithm and a second-order Minkowski distance according to the normalization result; The feasible region of the different types of distributed resources is determined by using a Zonotope method, and the overall feasible region of the same type of distributed resources is calculated by using a Minkowski summation method; A distributed resource type is randomly selected as a starting point, and the overall feasible region of the distributed resources of the type is taken as an initial convex hull; an outer normal vector of a hyperplane of the initial convex hull is taken as a search direction, and the overall feasible region vertex of the distributed resources is searched by using an iterative method based on a vertex search model; when the overall feasible region vertex of the vertex distributed resources reaches a preset threshold, the feasible region of the virtual power plant aggregated total power is calculated by using a feasible region calculation formula of the virtual power plant aggregated total power. The expression of the vertex search model is: maxη Tran P total where max denotes the maximum value, η denotes the search direction, P total denotes the vertex of the feasible region of the aggregated total power of the virtual power plant, denotes the aggregated power of the kth type of distributed resource, K denotes the number of distributed resource types, denotes the feasible region of the aggregated power of the kth type of distributed resource.

2. The virtual power plant resource classification and aggregation method of claim 1, wherein, The distributed resource features include a resource type, a rated capacity, a production and sales characteristic, a response type, an advance notification time, a response adjustment time, a response duration, a response capacity and a response period. 3.The virtual power plant resource classification and aggregation method according to claim 1, characterized in that, The numerical representation model of the distributed resource features is normalized by using a normalization method, and different types of distributed resources are obtained by classifying the distributed resources by using a clustering algorithm and a second-order Minkowski distance according to the normalization result. The distributed resource features are normalized by using a min-max normalization method according to the numerical representation model of the distributed resource features, and a normalization result is obtained. The second-order Minkowski distance between different distributed resources is calculated by using a second-order Minkowski distance according to the normalization result. The distributed resources are classified by using a K-means clustering algorithm according to the second-order Minkowski distance between the distributed resources, and different types of distributed resources are obtained.

4. The virtual power plant resource classification and aggregation method of claim 3, wherein, The calculation formula of the second-order Minkowski distance between different distributed resources is: wherein d i,j denotes the second order Minkowski distance between the ith distributed resource and the jth distributed resource, Cn denotes the total number of distributed resource features, denotes the cth distributed resource feature value of the ith distributed resource, denotes the cth distributed resource feature of the jth distributed resource.

5. The virtual power plant resource classification and aggregation method of claim 3, wherein, The distributed resources are classified by using a K-means clustering algorithm according to the second-order Minkowski distance between the distributed resources, and different types of distributed resources are obtained. A plurality of data points are randomly selected as clustering centers from the virtual power plant distributed resources, and each distributed resource is allocated to a resource category represented by a clustering center closest to the distributed resource according to the Minkowski distance between the different distributed resources, and a preliminary clustering result is formed; The average value of all the distributed resources in each resource category is calculated, and the average value of the distributed resources is taken as a new clustering center. The change value of the clustering center is compared, if the change value of the clustering center is less than a set threshold, the clustering result of the distributed resources is output, different types of distributed resources are obtained, otherwise, the Minkowski distance between different distributed resources is recalculated, and the clustering center is updated according to the recalculated result, until the change value of the clustering center is less than the set threshold, the updating is stopped, and the clustering result of the distributed resources is output, different types of distributed resources are obtained. 6.The virtual power plant resource classification and aggregation method according to claim 1, characterized in that, The feasible region calculation formula of the virtual power plant aggregation total power is: In the formula, F VPP represents the feasible region of the virtual power plant aggregate total power, represents the feasible region vertex, alpha q represents the convex combination coefficient, N vpp represents the number of vertices of the virtual power plant feasible region, q represents the qth vertex.

7. A virtual power plant resource classification and aggregation system, characterized by, It comprises: The data processing module is configured to obtain the distributed resources of the virtual power plant, extract features of the distributed resources of the virtual power plant, and obtain distributed resource features; and construct a numerical representation model of the distributed resource features according to the distributed resource features; The resource classification module is configured to normalize the distributed resource features by using a normalization method based on the numerical representation model of the distributed resource features; and classify the distributed resources by using a clustering algorithm and a second-order Minkowski distance according to the normalization result, to obtain different types of distributed resources; The feasible region determination module is configured to determine the feasible region of different types of distributed resources by using a Zonotope method, and calculate the overall feasible region of the same type of distributed resources by using a Minkowski summation method. Randomly select a type of distributed resources as a starting point, and take the overall feasible region of the type of distributed resources as an initial convex hull; take the outward normal vector of the hyperplane of the initial convex hull as a search direction, search the overall feasible region vertex of the distributed resources by using an iterative method based on a vertex search model; when the overall feasible region vertex of the vertex distributed resources reaches a pre-set threshold, calculate the feasible region of the virtual power plant aggregation total power by using the feasible region calculation formula of the virtual power plant aggregation total power. The expression of the vertex search model is: maxη Tran P total where max denotes the maximum value, η denotes the search direction, P total denotes the vertex of the feasible region of the aggregated total power of the virtual power plant, denotes the aggregated power of the kth type of distributed resource, K denotes the number of distributed resource types, denotes the feasible region of the aggregated power of the kth type of distributed resource.

8. The virtual power plant resource classification and aggregation system of claim 7, wherein, The distributed resource features include: resource type, rated capacity, production and sales characteristics, response type, advance notification time, response adjustment time, response duration, response capacity, and response period. 9.The virtual power plant resource classification and aggregation system of claim 8, wherein, When the resource classification module normalizes the distributed resource features by using a normalization method based on the numerical representation model of the distributed resource features, and classifies the distributed resources by using a clustering algorithm and a second-order Minkowski distance according to the normalization result, to obtain different types of distributed resources, the distributed resource features are normalized by using a min-max normalization method according to the numerical representation model of the distributed resource features, to obtain a normalization result; the second-order Minkowski distance between different distributed resources is calculated according to the normalization result by using a second-order Minkowski distance; and the distributed resources are classified by using a K-means clustering algorithm according to the second-order Minkowski distance between the distributed resources, to obtain different types of distributed resources. 10.The virtual power plant resource classification and aggregation system of claim 9, wherein, The K-means clustering algorithm is used to classify the distributed resources according to the second-order Minkowski distance between the distributed resources, to obtain different types of distributed resources, which comprises: Randomly select several data points from the virtual power plant distributed resources as clustering centers, and according to the Minkowski distance between different distributed resources, distribute each distributed resource to the resource category represented by the nearest clustering center to form a preliminary clustering result; For each resource category, calculate the average value of all distributed resources in the resource category, and take the average value of the distributed resources as a new clustering center; Compare the change value of the clustering center, if the change value of the clustering center is less than the set threshold value, output the clustering result of the distributed resources, and obtain different types of distributed resources, otherwise, recalculate the Minkowski distance between different distributed resources, and update the clustering center according to the recalculated result, until the change value of the clustering center is less than the set threshold value, stop updating, and output the clustering result of the distributed resources, and obtain different types of distributed resources. 11.A computer device, comprising a memory and a processor, wherein the memory stores a computer program, and the computer device is configured to perform the method according to any one of claims 1-10. The processor executes the computer program to realize the steps of the method in any one of claims 1 to 6.

12. A computer readable storage medium having stored thereon a computer program, characterized in that, The computer program is executed by the processor to realize the steps of the method in any one of claims 1 to 6.

Images