Power Load Scheduling Method, Device, Electronic Device and Storage Medium

Abstract

This application belongs to the technical field of power dispatching, and discloses a power load dispatching method, device, electronic device and storage medium. The method includes: obtaining the load information of the loads in the area to be measured, where the load information includes the number of all loads and the operating conditions of all loads within a preset period; according to the number and operating conditions, using the spectral clustering algorithm to divide the loads to obtain multiple load groups; based on the operating conditions of the multiple load groups, constructing a likelihood function based on the Markov random field to calculate, through Kalman filtering, the planned operating conditions of the multiple load groups; and according to the planned operating conditions, performing operating dispatching on the corresponding load groups. By using the likelihood function based on the Markov random field and Kalman filtering to dispatch the load groups divided by the spectral clustering algorithm, the dispatching efficiency of the power system is improved.

Description

Technical Field

[0001] The present application relates to the technical field of power dispatching. Specifically, it relates to a power load dispatching method, device, electronic device and storage medium. Background Art

[0002] In recent years, the continuous growth of power demand and the diversification of power loads have brought new challenges to the stable operation and efficient dispatching of power systems. However, traditional power dispatching methods often rely on empirical judgments or simple models, making it difficult to accurately capture the temporal variations and spatial distribution characteristics of loads, resulting in problems such as uneven power distribution and insufficient efficiency.

[0003] Therefore, in order to solve the technical problems that the existing power dispatching methods rely on empirical judgments or simple models and are difficult to accurately capture the temporal variations and spatial distribution characteristics of loads, leading to uneven power distribution and insufficient power dispatching efficiency, there is an urgent need for a power load dispatching method, device, electronic device and storage medium. Summary of the Invention

[0004] The purpose of the present application is to provide a power load dispatching method, device, electronic device and storage medium. By using the likelihood function based on the Markov random field and applying the Kalman filter, the operation scheduling of the load groups divided by the spectral clustering algorithm is carried out, solving the problems that the existing power dispatching methods rely on empirical judgments or simple models and are difficult to accurately capture the temporal variations and spatial distribution characteristics of loads, resulting in uneven power distribution and insufficient power dispatching efficiency. By grouping the loads to achieve the optimal dispatching of the power system, it can dispatch the power system from the perspectives of time and space, reduce the transmission loss of the power system, and improve the dispatching efficiency of the power system.

[0005] In a first aspect, the present application provides a power load dispatching method, including:

[0006] Obtain the load information of the loads in the area to be measured; the load information includes the number of all loads and the operation conditions of all loads within a preset period;

[0007] According to the number and the operation conditions, use the spectral clustering algorithm to divide the loads to obtain multiple load groups;

[0008] Based on the operation conditions of the multiple load groups, construct a likelihood function based on the Markov random field to calculate the planned operation conditions of the multiple load groups through the Kalman filter;

[0009] According to the planned operation conditions, perform operation scheduling on the corresponding load groups.

[0010] The power load scheduling method provided by this application can achieve optimized scheduling of power loads. By using the likelihood function based on the Markov random field and applying the Kalman filter, the operation scheduling of the load groups divided by the spectral clustering algorithm is carried out, solving the problems that the existing power scheduling methods rely on empirical judgments or simple models and are difficult to accurately capture the temporal changes and spatial distribution characteristics of the loads, resulting in uneven power distribution and insufficient power scheduling efficiency. By grouping the loads to achieve optimized scheduling of the power system, the power system can be scheduled from the perspectives of time and space, reducing the transmission loss of the power system and improving the scheduling efficiency of the power system.

[0011] Optionally, according to the quantity and the operating conditions, using the spectral clustering algorithm, the loads are divided to obtain multiple load groups, including:

[0012] According to the quantity, an adjacency matrix of the loads with respect to the operating conditions is constructed;

[0013] According to the Laplacian matrix corresponding to the adjacency matrix, a feature vector space is constructed;

[0014] Through the K-means clustering algorithm, based on the feature vectors in the feature vector space, the loads are divided to obtain multiple load groups.

[0015] The power load scheduling method provided by this application can achieve optimized scheduling of power loads. Through the feature vector space constructed based on the Laplacian matrix corresponding to the operating conditions and applying the K-means clustering, the loads are divided to obtain multiple load groups, and scheduling the loads through the load groups is beneficial to improving the scheduling efficiency of the power system.

[0016] Optionally, according to the quantity, constructing the adjacency matrix of the loads with respect to the operating conditions includes:

[0017] Construct an operation matrix of the quantity of the loads with respect to the operating conditions;

[0018] Using the difference method, a difference operation matrix corresponding to the operation matrix is established;

[0019] According to the specific values of the elements in the difference operation matrix, an adjacency matrix of the loads with respect to the operating conditions is constructed.

[0020] Optionally, through the K-means clustering algorithm, based on the feature vectors in the feature vector space, the loads are divided to obtain multiple load groups, including:

[0021] Based on the feature vectors in the feature vector space, the operating conditions of the loads are represented in a reduced dimension to obtain the load information after dimensionality reduction;

[0022] Select multiple loads from all the loads as multiple initial clustering centers according to the dimension-reduced load information;

[0023] Based on the initial clustering centers, group all the loads through the K-means clustering algorithm to obtain multiple load groups.

[0024] Optionally, based on the initial clustering centers, group all the loads through the K-means clustering algorithm to obtain multiple load groups, including:

[0025] Step A1, calculate the distances between other loads except the initial clustering centers among all the loads and the multiple initial clustering centers;

[0026] Step A2, assign the other loads to the nearest initial clustering center to form primary load groups;

[0027] Step A3, recalculate the clustering centers in the primary load groups, and reassign all the other loads except the clustering centers according to the distances between the clustering centers and all the other loads except the clustering centers;

[0028] Step A4, repeatedly execute Step A3 until the clustering centers no longer change, and obtain multiple load groups.

[0029] Optionally, based on the operating conditions of the multiple load groups, construct a likelihood function based on the Markov random field to calculate the planned operating conditions of the multiple load groups through Kalman filtering, including:

[0030] Based on the operating conditions of the multiple load groups, construct a corresponding Markov random function;

[0031] Add a correction term to the Markov random function to construct a likelihood function based on the Markov random field;

[0032] Through Kalman filtering, iterate the likelihood function based on the Markov random field to calculate the planned operating conditions of the multiple load groups.

[0033] The power load scheduling method provided by this application can realize the optimal scheduling of power loads. By adding a correction term to the Markov random function constructed based on the operating conditions to construct a likelihood function based on the Markov random field, and through Kalman filtering, iterating the likelihood function based on the Markov random field to calculate the planned operating conditions of multiple load groups, and scheduling the operation of multiple load groups through the planned operating conditions, which is beneficial to improving the scheduling efficiency of the power system.

[0034] Optionally, the correction terms include an observation matrix correction term and a noise correction term; adding correction terms to the Markov random function to construct a likelihood function based on the Markov random field, including:

[0035] Adding the noise correction term to the Markov random function to obtain an optimized Markov random function;

[0036] According to the optimized Markov random function and the actual operating conditions of the corresponding load groups, combining with the observation matrix correction term, constructing an operating condition function;

[0037] Based on the noise correction term, the optimized Markov random function, the observation matrix correction term, and the operating condition function, constructing a likelihood function based on the Markov random field.

[0038] In a second aspect, the present application provides a power load scheduling device, including:

[0039] An acquisition module, configured to acquire load information of loads in a to-be-tested area; the load information includes the number of all loads and the operating conditions of all loads within a preset period;

[0040] A partitioning module, configured to partition the loads by using a spectral clustering algorithm according to the number and the operating conditions to obtain a plurality of load groups;

[0041] A calculation module, configured to construct a likelihood function based on the Markov random field based on the operating conditions of the plurality of load groups, and calculate the planned operating conditions of the plurality of load groups through Kalman filtering;

[0042] A scheduling module, configured to perform operating scheduling on the corresponding load groups according to the planned operating conditions.

[0043] The power load scheduling device uses the likelihood function based on the Markov random field and applies Kalman filtering to perform operating scheduling on the load groups obtained by partitioning based on the spectral clustering algorithm, solving the problem that the existing power scheduling method depends on empirical judgment or simple models and is difficult to accurately capture the temporal changes and spatial distribution characteristics of loads, resulting in uneven power distribution and insufficient power scheduling efficiency. By grouping the loads to achieve optimized scheduling of the power system, it can schedule the power system from the perspectives of time and space, reduce the transmission loss of the power system, and improve the scheduling efficiency of the power system.

[0044] In a third aspect, the present application provides an electronic device, including a processor and a memory, where the memory stores a computer program executable by the processor, and when the processor executes the computer program, it runs the steps in the power load scheduling method described above.

[0045] In a fourth aspect, the present application provides a computer-readable storage medium having a computer program stored thereon, and when the computer program is executed by a processor, it runs the steps in the power load scheduling method described above.

[0046] Beneficial effects: The power load scheduling method, device, electronic device and storage medium provided by the present application use Kalman filtering through the likelihood function of the Markov random field to perform operation scheduling on the load groups divided by the spectral clustering algorithm, solving the problem that the existing power scheduling methods rely on empirical judgments or simple models and are difficult to accurately capture the temporal changes and spatial distribution characteristics of the load, resulting in uneven power distribution and insufficient power scheduling efficiency. By grouping the loads to achieve optimal scheduling of the power system, it can schedule the power system from the perspectives of time and space, reduce the transmission loss of the power system, and improve the scheduling efficiency of the power system. Description of the Drawings

[0047] Figure 1 It is a flowchart of the power load scheduling method provided by an embodiment of the present application.

[0048] Figure 2 It is a schematic structural diagram of the power load scheduling device provided by an embodiment of the present application.

[0049] Figure 3 It is a schematic structural diagram of the electronic device provided by an embodiment of the present application.

[0050] Reference numerals: 1, acquisition module; 2, division module; 3, calculation module; 4, scheduling module; 301, processor; 302, memory; 303, communication bus. Detailed Embodiments

[0051] Next, the technical solutions in the embodiments of the present application will be clearly and completely described in conjunction with the drawings in the embodiments of the present application. Obviously, the described embodiments are only a part of the embodiments of the present application, rather than all the embodiments. Usually, the components of the embodiments of the present application described and shown here can be arranged and designed in various different configurations. Therefore, the following detailed description of the embodiments of the present application provided in the drawings is not intended to limit the scope of the present application claimed, but only represents the selected embodiments of the present application. Based on the embodiments of the present application, all other embodiments obtained by those skilled in the art without creative efforts belong to the scope of protection of the present application.

[0052] It should be noted that similar reference numerals and letters denote similar items in the following drawings. Therefore, once an item is defined in one drawing, it does not need to be further defined and explained in subsequent drawings. At the same time, in the description of the present application, terms such as "first" and "second" are only used for distinguishing descriptions and cannot be construed as indicating or implying relative importance.

[0053] Please refer to Figure 1 , Figure 1 which is a power load scheduling method in some embodiments of the present application for optimizing the scheduling of power loads, including the steps of:

[0054] Step S101, obtaining the load information of the loads in the area to be measured; the load information includes the number of all loads and the operating conditions of all loads within a preset period;

[0055] Step S102, dividing the loads by using the spectral clustering algorithm according to the (number of all loads) and the operating conditions to obtain multiple load groups;

[0056] Step S103, constructing a likelihood function based on the Markov random field based on the operating conditions of the multiple load groups, so as to calculate the planned operating conditions of the multiple load groups through Kalman filtering;

[0057] Step S104, performing operation scheduling on the corresponding load groups according to the planned operating conditions.

[0058] This power load scheduling method uses the likelihood function based on the Markov random field and applies Kalman filtering to perform operation scheduling on the load groups divided by the spectral clustering algorithm, solving the problems that the existing power scheduling methods rely on empirical judgments or simple models and are difficult to accurately capture the temporal changes and spatial distribution characteristics of loads, resulting in uneven power distribution and insufficient power scheduling efficiency. By grouping the loads to achieve the optimized scheduling of the power system, it can schedule the power system from the perspectives of time and space, reduce the transmission loss of the power system, and improve the scheduling efficiency of the power system.

[0059] Specifically, in step S101, the load information of the loads in the area to be measured is collected at multiple moments (i.e., sampling moments) within a preset period. The load information includes the number of all loads and the operating conditions of all loads within a preset period. Among them, the preset period is a time period of a preset length, the operating condition is whether the load is in an operating state or a non-operating state, and the operating conditions of the load within the preset period include the operating conditions of the load at multiple moments within the preset period; the preset period can be set according to actual needs.

[0060] For example, if the preset period contains T moments and there are n loads, then there are n×T operating conditions of the loads within the T moments.

[0061] Specifically, in step S102, according to the quantity and operating conditions, the load is partitioned using the spectral clustering algorithm to obtain multiple load groups, including:

[0062] Construct an adjacency matrix of the load regarding the operating conditions according to the quantity;

[0063] Construct an eigenvector space based on the Laplacian matrix corresponding to the adjacency matrix;

[0064] The load is partitioned using the K-means clustering algorithm based on the eigenvectors in the eigenvector space to obtain multiple load groups.

[0065] Specifically, in step S102, constructing an adjacency matrix of the load regarding the operating conditions according to the quantity includes:

[0066] Construct an operation matrix of the quantity of the load regarding the operating conditions;

[0067] Adopt the difference method to establish a differential operation matrix corresponding to the operation matrix;

[0068] Construct an adjacency matrix of the load regarding the operating conditions according to the specific values of each element in the differential operation matrix.

[0069] In step S102, for the operating conditions at T moments, an operation matrix X corresponding to the quantity n of the load is constructed (i.e., an n×T matrix). Each element in the operation matrix can be represented by 0 or 1, where 0 indicates the non-operating state and 1 indicates the operating state. Among them, indicates that load i (i≤n) is in the operating state at the j-th (j≤T) moment ( is the element in the i-th row and j-th column of X), indicates that load i is in the non-operating state at the j-th moment, indicates the operating conditions of all loads at the j-th moment ( is the vector composed of the elements in the j-th column of X).

[0070] Adopt the difference method to subtract the operating conditions of all loads at any moment from the operating conditions of all loads at the corresponding previous moment to obtain a differential operation matrix. For example, for j<T, subtract the operating conditions of all loads at the (j + 1)-th moment from the operating conditions of all loads at the j-th moment to obtain the differential operation vector at the j-th moment , that is , for j=T, . The differential operation matrix is obtained from the differential operation vectors at T moments.

[0071] Construct the adjacency matrix \(W\) of the load regarding its operating conditions. The initial adjacency matrix \(W\) is a zero matrix (the size of the adjacency matrix \(W\) is also \(n\times T\)). According to the specific values of the elements in the differential operation matrix, perform iterative updates to determine each element of the adjacency matrix \(W\). When the element of the differential operation matrix at time \(j\) for load \(i\) is (where is the element in the \(i\)-th row and \(j\)-th column of the differential operation matrix), for all loads in the differential operation matrix that satisfy the element is the element in the \(k\)-th row and \(j\)-th column of the differential operation matrix), for the load ), let , that is, replace the value of the element representing the operating condition relationship between load \(i\) and load \(k\) in the adjacency matrix \(W\) with the value of + 1 ( is the element in the \(i\)-th row and \(k\)-th column of \(W\)). That is, when the elements of the differential operation matrix at the same time are the same, the values of the corresponding elements representing the operating condition relationship between the loads in the adjacency matrix are the same, and when all the elements of the differential operation matrix are 1, the values of the corresponding elements representing the operating condition relationship between the loads in the adjacency matrix will be updated to 1.

[0072] Specifically, in step S102, calculate the degree matrix \(D\) of the adjacency matrix \(W\). The degree matrix \(D\) is a diagonal matrix, and its diagonal elements are ( represents the adjacency matrix of load \(i\) regarding its operating conditions at time \(j\)), calculate the normalized adjacency matrix , and thus calculate the Laplacian matrix , is the identity matrix. Among them, the calculation process of the degree matrix is prior art and will not be elaborated here.

[0073] Perform eigenvalue decomposition on the Laplacian matrix , and use the eigenvectors corresponding to the non-zero eigenvalues to construct the eigenvector space. Among them, the non-zero eigenvalues can be set according to actual needs.

[0074] Specifically, in step S102, through the K-means clustering algorithm, based on the eigenvectors in the eigenvector space, divide the loads to obtain multiple load groups, including:

[0075] Based on the eigenvectors in the eigenvector space, perform dimensionality reduction representation on the operating conditions of the loads to obtain the load information after dimensionality reduction;

[0076] According to the load information after dimensionality reduction, select multiple loads from all loads as multiple initial clustering centers;

[0077] Based on the initial cluster centers, all loads are grouped through the K-means clustering algorithm to obtain multiple load groups.

[0078] In step S102, based on the number corresponding to the non-zero eigenvalues of the eigenvectors in the feature vector space, the number of the operating conditions of the loads is reduced to the corresponding number through existing dimensionality reduction methods such as the principal component analysis (PCA) method or the linear discriminant analysis (LDA) method, etc., to obtain the dimensionality-reduced load information. Among them, existing dimensionality reduction methods such as the principal component analysis method or the linear discriminant analysis method are prior arts and will not be elaborated herein.

[0079] Regarding the loads as points, multiple loads are randomly selected from the dimensionality-reduced load information as multiple initial cluster centers, and other loads are assigned to the initial cluster center with the minimum distance from them, forming load groups mainly centered on the initial cluster centers (i.e., the cluster centers of these load groups are the initial cluster centers). After each grouping is completed, the cluster centers of each load group need to be re-divided (the division method is to select the load closest to the group center in the load group as the new cluster center), and based on the divided cluster centers, the other loads in each load group are re-assigned (the assignment method is to assign the other loads to the cluster center with the minimum distance from them) until all the loads in the re-assigned load groups no longer need to be re-assigned, forming multiple load groups.

[0080] Specifically, in step S102, based on the initial cluster centers, all loads are grouped through the K-means clustering algorithm to obtain multiple load groups, including:

[0081] Step A1, calculate the distances between the other loads except the initial cluster centers among all loads and the multiple initial cluster centers;

[0082] Step A2, assign the other loads to the initial cluster center with the closest distance to form primary load groups;

[0083] Step A3, recalculate the cluster centers in the primary load groups, and based on the distances between the cluster centers and all the other loads except the cluster centers, re-assign all the other loads except the cluster centers;

[0084] Step A4, repeat step A3 until the cluster centers no longer change, obtaining multiple load groups.

[0085] In step S102, calculate the distances between other loads in the dimension-reduced load information and the initial cluster centers, group them with the initial cluster centers as the central nodes, extract the initial cluster center with the minimum distance from other loads from all the initial cluster centers, and assign the other loads to the group where the initial cluster center with the minimum distance from them is located. After each assignment, it is necessary to recalculate the load closest to the center point position as the cluster center of the load group according to the positions of the loads within each load group, recalculate the distances between other loads except the cluster center and multiple cluster centers, and assign the other loads to the group where the cluster center with the minimum distance from them is located. Repeat the steps of calculating the cluster center and assigning loads until the cluster center no longer changes or the loads in each load group no longer change, and determine the finally calculated load groups as multiple load groups.

[0086] Specifically, in step S103, based on the operating conditions of multiple load groups, construct a likelihood function based on the Markov random field, and calculate the planned operating conditions of multiple load groups through Kalman filtering, including:

[0087] Construct a corresponding Markov random function based on the operating conditions of multiple load groups;

[0088] Add a correction term to the Markov random function to construct a likelihood function based on the Markov random field;

[0089] Iterate the likelihood function based on the Markov random field through Kalman filtering to calculate the planned operating conditions of multiple load groups.

[0090] In step S103, regard the operating conditions of all loads over time as a Markov chain, use the above normalization adjacency matrix as the state transition matrix to obtain the corresponding Markov random function. The Markov random function is specifically:

[0091] ;

[0092] where is the operating matrix at time T + j (time T is the current time).

[0093] Through the Markov random function, starting from the operating matrix at time T (the current time), the operating conditions of the load at times T + 1, T + 2,..., to T + j can be calculated through an iterative method.

[0094] Specifically, the correction term includes an observation matrix correction term and a noise correction term; in step S103, adding a correction term to the Markov random function to construct a likelihood function based on the Markov random field includes:

[0095] Add a noise correction term to the Markov random function to obtain an optimized Markov random function;

[0096] According to the actual operation of the optimized Markov random function and the corresponding load grouping, combined with the observation matrix correction term, construct an operation condition function;

[0097] Based on the noise correction term, the optimized Markov random function, the observation matrix correction term, and the operation condition function, construct a likelihood function based on the Markov random field.

[0098] In step S103, the influence of noise cannot be ignored in the iteration of the operation condition. Consider the zero-mean noise in the iteration process , add a noise correction term to the Markov random function to obtain an optimized Markov random function. The optimized Markov random function is specifically:

[0099] ;

[0100] where, is the noise correction term at time (i.e., zero-mean noise).

[0101] Use the observation matrix correction term to construct a relational expression for the actual operation of the optimized Markov random function and the corresponding load grouping to obtain an operation condition function. The operation condition function is specifically:

[0102] ;

[0103] where, is the observable operation condition (the operation condition includes the observable operation condition and the unobservable operation condition); is the observation matrix correction term, , is the observation matrix, is the observation matrix degree matrix. The degree matrix is a diagonal matrix, and its diagonal elements are , is the element corresponding to load i at time j in the observation matrix . The observation matrix can be set according to actual needs.

[0104] For the observation matrix , when its diagonal element it means that load i is an anchor point, it means that load i is not an anchor point. When load k and load i are assigned to the same load grouping, let the observation matrix the elements in . Among them, the anchor points are the loads closest to the cluster center in each load group, as well as the load with the smallest abscissa, the load with the largest abscissa, the load with the smallest ordinate, and the load with the largest ordinate among all loads. That is, there are always K + 4 anchor points, where K is the number of load groups.

[0105] In some embodiments, after obtaining multiple load groups, the load closest to the cluster center can be selected as an anchor point in each load group, and the load with the smallest abscissa, the load with the largest abscissa, the load with the smallest ordinate, and the load with the largest ordinate can be selected as anchor points among all loads. And after determining the anchor points, algorithms such as the XGBoost algorithm can be used to predict the spatial positions of the loads other than the anchor points and the cluster center.

[0106] In summary, based on the noise correction term, the optimized Markov random function, the observation matrix correction term, and the operating condition function, a likelihood function based on the Markov random field is constructed. The likelihood function based on the Markov random field is specifically:

[0107] ;

[0108] where p is the probability of the occurrence of under the given condition ( ), S is the covariance matrix of the zero-mean noise . Among them, the covariance matrix S is equivalent to the variance of the current moment's operation matrix , that is , and the superscript T is the transpose symbol.

[0109] In step S103, through Kalman filtering, the likelihood function based on the Markov random field is iterated to maximize the likelihood function of the Markov random field (i.e., calculate the maximum value of p), so as to determine the corresponding to the maximum value, and obtain the planned operating conditions of multiple load groups. Among them, Kalman filtering is a prior art and will not be elaborated here.

[0110] Specifically, in step S104, according to the planned operating conditions of each load group, the corresponding load group is scheduled for operation.

[0111] As can be seen from the above, in the power load scheduling method, by obtaining the load information of the loads in the area to be measured, where the load information includes the number of all loads and the operating conditions of all loads within a preset period, and based on the number and operating conditions, using the spectral clustering algorithm to divide the loads, multiple load groups are obtained. Based on the operating conditions of the multiple load groups, a likelihood function based on the Markov random field is constructed to calculate the planned operating conditions of the multiple load groups through Kalman filtering. According to the planned operating conditions, the corresponding load groups are scheduled for operation. Thus, through the likelihood function based on the Markov random field and using Kalman filtering, the load groups divided by the spectral clustering algorithm are scheduled for operation, solving the problem that the existing power scheduling methods rely on empirical judgments or simple models and are difficult to accurately capture the temporal changes and spatial distribution characteristics of the loads, resulting in uneven power distribution and insufficient power scheduling efficiency. By grouping the loads to achieve optimal scheduling of the power system, the power system can be scheduled from the perspectives of time and space, reducing the transmission loss of the power system and improving the scheduling efficiency of the power system.

[0112] Reference Figure 2 , this application provides a power load scheduling device for optimizing the scheduling of power loads, including:

[0113] An acquisition module 1 for acquiring the load information of the loads in the area to be measured; the load information includes the number of all loads and the operating conditions of all loads within a preset period;

[0114] A division module 2 for dividing the loads using the spectral clustering algorithm according to the (number of all loads) and the operating conditions to obtain multiple load groups;

[0115] A calculation module 3 for constructing a likelihood function based on the Markov random field based on the operating conditions of the multiple load groups to calculate the planned operating conditions of the multiple load groups through Kalman filtering;

[0116] A scheduling module 4 for scheduling the operation of the corresponding load groups according to the planned operating conditions.

[0117] This power load scheduling device schedules the operation of the load groups divided by the spectral clustering algorithm through the likelihood function based on the Markov random field and using Kalman filtering, solving the problem that the existing power scheduling methods rely on empirical judgments or simple models and are difficult to accurately capture the temporal changes and spatial distribution characteristics of the loads, resulting in uneven power distribution and insufficient power scheduling efficiency. By grouping the loads to achieve optimal scheduling of the power system, the power system can be scheduled from the perspectives of time and space, reducing the transmission loss of the power system and improving the scheduling efficiency of the power system.

[0118] Specifically, when the acquisition module 1 is executing, it collects the load information of the loads in the area to be measured at multiple moments (i.e., sampling moments) within a preset period. The load information includes the quantity of all loads and the operating conditions of all loads within the preset period. Here, the preset period is a time period of a preset length, and the operating condition is whether the load is in an operating state or a non-operating state. The operating conditions of the load within the preset period include the operating conditions at multiple moments within the preset period; the preset period can be set according to actual needs.

[0119] For example, if the preset period contains T moments and there are n loads, then there are n×T operating conditions of the loads within the T moments.

[0120] Specifically, when the partitioning module 2 partitions the loads using the spectral clustering algorithm based on the quantity and operating conditions to obtain multiple load groups, it executes:

[0121] Construct an adjacency matrix of the loads regarding the operating conditions according to the quantity;

[0122] Construct an eigenvector space based on the Laplacian matrix corresponding to the adjacency matrix;

[0123] Partition the loads through the K-means clustering algorithm based on the eigenvectors in the eigenvector space to obtain multiple load groups.

[0124] Specifically, when the partitioning module 2 constructs an adjacency matrix of the loads regarding the operating conditions according to the quantity, it executes:

[0125] Construct an operation matrix of the quantity of the loads regarding the operating conditions;

[0126] Adopt the difference method to establish a differential operation matrix corresponding to the operation matrix;

[0127] Construct an adjacency matrix of the loads regarding the operating conditions according to the specific values of the elements in the differential operation matrix.

[0128] When the partitioning module 2 is executing, for the operating conditions at T moments, it constructs an operation matrix X corresponding to the quantity n of the loads (i.e., an n×T matrix). Each element in the operation matrix can be represented by 0 or 1, where 0 represents the non-operating state and 1 represents the operating state. Here, represents that load i (i≤n) is in the operating state at the j-th (j≤T) moment ( is the element in the i-th row and j-th column of X), represents that load i is in the non-operating state at the j-th moment, represents the operating conditions of all loads at the j-th moment ( is the vector composed of the elements in the j-th column of X).

[0129] Using the difference method, the operating conditions of all loads at any moment are subtracted from the operating conditions of all loads at the corresponding previous moment to obtain a differential operation matrix. For example, for j < T, the operating conditions of all loads at the (j + 1)-th moment are subtracted from the operating conditions of all loads at the j-th moment to obtain the differential operation vector at the j-th moment , that is . For j = T . The differential operation matrix is obtained from the differential operation vectors at T moments

[0130] Construct an adjacency matrix W of the loads regarding the operating conditions. The initial adjacency matrix W is a zero matrix (the size of the adjacency matrix W is also n × T). According to the specific values of the elements in the differential operation matrix, iterative updates are performed to determine each element of the adjacency matrix W. When the element in the differential operation matrix of load i at the j-th moment is the element in the i-th row and j-th column of the differential operation matrix , for all loads in the differential operation matrix that satisfy the element is the element in the k-th row and j-th column of the differential operation matrix ), let ), that is, replace the value of the element representing the operating condition relationship between load i and load k in the adjacency matrix W with the value of +1 ( is the element in the i-th row and k-th column of W), that is, when the elements of the differential operation matrix at the same moment are the same, the values of the elements representing the operating condition relationships of the corresponding loads in the adjacency matrix are the same, and when the elements of the differential operation matrix are all 1, the values of the elements representing the operating condition relationships of the corresponding loads in the adjacency matrix will be updated to 1

[0131] Specifically, when the partitioning module 2 is executed, the degree matrix D of the adjacency matrix W is calculated. The degree matrix D is a diagonal matrix, and its diagonal elements are ]]), where represents the adjacency matrix of load i at the j-th moment regarding the operating conditions. The normalized adjacency matrix is calculated, and then the Laplacian matrix is calculated , where

[0132] is the identity matrix. The calculation process of the degree matrix is prior art and will not be elaborated here The Laplacian matrix is eigen-decomposed, and the eigenvectors corresponding to the number of non-zero eigenvalues are used to construct the eigenvector space. The non-zero eigenvalues can be set according to actual needs

[0133] Specifically, when the partitioning module 2 partitions the load through the K-means clustering algorithm based on the feature vectors in the feature vector space to obtain multiple load groups, it performs the following:

[0134] Perform dimensionality reduction representation on the operating conditions of the load based on the feature vectors in the feature vector space to obtain the load information after dimensionality reduction;

[0135] Select multiple loads from all loads as multiple initial clustering centers according to the load information after dimensionality reduction;

[0136] Group all loads through the K-means clustering algorithm based on the initial clustering centers to obtain multiple load groups.

[0137] When the partitioning module 2 executes, based on the number corresponding to the non-zero eigenvalues corresponding to the feature vectors in the feature vector space, through existing dimensionality reduction methods such as the principal component analysis (PCA) method or the linear discriminant analysis (LDA) method, reduce the number of the operating conditions of the load to the corresponding number to obtain the load information after dimensionality reduction. Among them, existing dimensionality reduction methods such as the principal component analysis method or the linear discriminant analysis method are prior arts, and details thereof are not described herein.

[0138] Regard the load as a point, randomly select multiple loads from the load information after dimensionality reduction as multiple initial clustering centers, assign other loads to the initial clustering center with the smallest distance from them, and form a load group mainly based on the initial clustering center (that is, the clustering center of this load group is the initial clustering center). After each grouping is completed, re-partition the clustering center of each load group (the partitioning method is to select the load closest to the group center in the load group as the new clustering center), and based on the partitioned clustering center, re-assign other loads in each load group (the assignment method is to assign other loads to the clustering center with the smallest distance from them) until all loads in the load group after re-assignment are no longer re-assigned, thus forming multiple load groups.

[0139] Specifically, when the partitioning module 2 groups all loads through the K-means clustering algorithm based on the initial clustering centers to obtain multiple load groups, it performs the following:

[0140] Step A1, calculate the distances between other loads except the initial clustering centers among all loads and the multiple initial clustering centers;

[0141] Step A2, assign other loads to the initial clustering center with the closest distance to form a primary load group;

[0142] Step A3: Recalculate the cluster centers in the primary load groups, and reassign all other loads outside the cluster centers according to the distances between the cluster centers and all other loads outside the cluster centers.

[0143] Step A4: Repeat Step A3 until the cluster centers no longer change, and multiple load groups are obtained.

[0144] When the partitioning module 2 is executed, calculate the distances between other loads in the load information after dimensionality reduction and the initial cluster centers, group them with the initial cluster centers as the central nodes, extract the initial cluster centers with the smallest distances from all initial cluster centers to other loads, and assign other loads to the groups where the initial cluster centers with the smallest distances to them are located. After each assignment is completed, it is necessary to recalculate the load closest to the center point position in each load group as the cluster center of the load group, recalculate the distances between other loads except the cluster centers and multiple cluster centers, and assign other loads to the groups where the cluster centers with the smallest distances to them are located. Repeat the steps of calculating the cluster centers and assigning loads until the cluster centers no longer change or the loads in each load group no longer change, and determine the finally calculated load groups as multiple load groups.

[0145] Specifically, when the calculation module 3 constructs a likelihood function based on the Markov random field based on the operation conditions of multiple load groups to calculate the planned operation conditions of multiple load groups through Kalman filtering, it executes:

[0146] Construct a corresponding Markov random function based on the operation conditions of multiple load groups;

[0147] Add a correction term to the Markov random function to construct a likelihood function based on the Markov random field;

[0148] Iterate the likelihood function based on the Markov random field through Kalman filtering to calculate the planned operation conditions of multiple load groups.

[0149] When the calculation module 3 is executed, regard the operation conditions of all loads in time series as a Markov chain, use the above normalized adjacency matrix as the state transition matrix to obtain the corresponding Markov random function, and the Markov random function is specifically:

[0150] ;

[0151] Where is the operation matrix at time T + j (time T is the current time).

[0152] Through the Markov random function, from the operation matrix at time T (the current time) Starting from the beginning, through an iterative approach, the operating conditions of the load at times T + 1, T + 2, … to T + j can be calculated.

[0153] Specifically, the correction terms include an observation matrix correction term and a noise correction term; when the calculation module 3 adds the correction terms in the Markov random function to construct the likelihood function based on the Markov random field, it performs:

[0154] Adding the noise correction term in the Markov random function to obtain an optimized Markov random function;

[0155] According to the optimized Markov random function and the actual operating conditions of the corresponding load groups, combined with the observation matrix correction term, construct an operating condition function;

[0156] Based on the noise correction term, the optimized Markov random function, the observation matrix correction term, and the operating condition function, construct a likelihood function based on the Markov random field.

[0157] When the calculation module 3 executes, the iteration of the operating conditions cannot ignore the influence of noise, and zero-mean noise in the iterative process is considered , adding the noise correction term in the Markov random function to obtain an optimized Markov random function, and the optimized Markov random function is specifically:

[0158] ;

[0159] Where, is the noise correction term at time

[0160] Using the observation matrix correction term, construct a relational expression between the optimized Markov random function and the actual operating conditions of the corresponding load groups to obtain an operating condition function, and the operating condition function is specifically:

[0161] ;

[0162] Where, is the observable operating condition (the operating condition includes the observable operating condition and the unobservable operating condition); is the observation matrix correction term, , is the observation matrix, is the observation matrix 's degree matrix, and the degree matrix is a diagonal matrix, whose diagonal elements are , is the observation matrix The element corresponding to load i at time j. The observation matrix can be set according to actual needs.

[0163] For the observation matrix its diagonal element indicates that load i is an anchor point when indicates that load i is not an anchor point when. When load k and load i are assigned to the same load group, let the element in the observation matrix . Among them, the anchor points are the loads closest to the cluster center in each load group, as well as the load with the smallest abscissa, the load with the largest abscissa, the load with the smallest ordinate, and the load with the largest ordinate among all loads. That is, there are always K + 4 anchor points, where K is the number of load groups.

[0164] In some embodiments, after obtaining multiple load groups, the load closest to the cluster center can be selected as an anchor point in each load group, and the load with the smallest abscissa, the load with the largest abscissa, the load with the smallest ordinate, and the load with the largest ordinate can be selected as anchor points among all loads. And after determining the anchor points, algorithms such as the XGBoost algorithm can be used to predict the spatial positions of the loads other than the anchor points and the cluster center.

[0165] In summary, based on the noise correction term, the optimized Markov random function, the observation matrix correction term, and the operating condition function, a likelihood function based on the Markov random field is constructed. The likelihood function based on the Markov random field is specifically:

[0166] ;

[0167] where p is the probability of appearing under the given condition ( ), S is the covariance matrix of the zero-mean noise . Among them, the covariance matrix S is equivalent to the variance of the current-time operating matrix , that is , and the superscript T is the transpose symbol.

[0168] When the calculation module 3 executes, through Kalman filtering, the likelihood function based on the Markov random field is iterated to maximize the likelihood function of the Markov random field (that is, calculate the maximum value of p), so as to determine the corresponding to the maximum value, and obtain the planned operating conditions of multiple load groups Among them, Kalman filtering is a prior art and will not be elaborated here.

[0169] Specifically, when the scheduling module 4 executes, it performs operation scheduling on the corresponding load group according to the planned operation conditions of each load group.

[0170] As can be seen from the above, the power load scheduling device obtains the load information of the loads in the area to be measured. The load information includes the number of all loads and the operation conditions of all loads within a preset period. According to the number and operation conditions, the spectral clustering algorithm is used to divide the loads to obtain multiple load groups. Based on the operation conditions of the multiple load groups, a likelihood function based on the Markov random field is constructed to calculate the planned operation conditions of the multiple load groups through Kalman filtering. According to the planned operation conditions, operation scheduling is performed on the corresponding load groups. Therefore, through the likelihood function based on the Markov random field and by using Kalman filtering, operation scheduling is performed on the load groups divided by the spectral clustering algorithm, solving the problem that the existing power scheduling method depends on empirical judgment or simple models and is difficult to accurately capture the temporal changes and spatial distribution characteristics of loads, resulting in uneven power distribution and insufficient power scheduling efficiency. By grouping the loads to achieve optimal scheduling of the power system, the power system can be scheduled from the perspectives of time and space, reducing the transmission loss of the power system and improving the scheduling efficiency of the power system.

[0171] Please refer to Figure 3 , Figure 3 which is a schematic structural diagram of an electronic device provided in an embodiment of the present application. The present application provides an electronic device, including: a processor 301 and a memory 302. The processor 301 and the memory 302 are interconnected and communicate with each other through a communication bus 303 and / or other forms of connection mechanisms (not marked). The memory 302 stores a computer program executable by the processor 301. When the electronic device runs, the processor 301 executes the computer program to perform the power load scheduling method in any optional implementation manner of the above embodiment to achieve the following functions: obtaining the load information of the loads in the area to be measured, where the load information includes the number of all loads and the operation conditions of all loads within a preset period, dividing the loads by using the spectral clustering algorithm according to the number and operation conditions to obtain multiple load groups, constructing a likelihood function based on the Markov random field based on the operation conditions of the multiple load groups to calculate the planned operation conditions of the multiple load groups through Kalman filtering, and performing operation scheduling on the corresponding load groups according to the planned operation conditions.

[0172] An embodiment of the present application provides a computer-readable storage medium, on which a computer program is stored. When the computer program is executed by a processor, it executes the power load scheduling method in any optional implementation manner of the above embodiment to achieve the following functions: obtaining load information of loads in a to-be-tested area, where the load information includes the number of all loads and the operating conditions of all loads in a preset period, dividing the loads by using a spectral clustering algorithm according to the number and the operating conditions to obtain multiple load groups, constructing a likelihood function based on a Markov random field based on the operating conditions of the multiple load groups, calculating the planned operating conditions of the multiple load groups through Kalman filtering, and performing operating scheduling on the corresponding load groups according to the planned operating conditions. Among them, the storage medium can be implemented by any type of volatile or non-volatile storage device or a combination thereof, such as static random access memory (SRAM for short), electrically erasable programmable read-only memory (EEPROM for short), erasable programmable read-only memory (EPROM for short), programmable read-only memory (PROM for short), read-only memory (ROM for short), magnetic memory, flash memory, a magnetic disk or an optical disc.

[0173] In the embodiments provided in the present application, it should be understood that the disclosed device and method can be implemented in other ways. The device embodiments described above are only illustrative. For example, the division of the units is only a logical function division, and there may be other division methods in actual implementation. For another example, multiple units or components can be combined or integrated into another system, or some features can be ignored or not executed. Another point is that the displayed or discussed mutual coupling or direct coupling or communication connection may be through some communication interfaces. The indirect coupling or communication connection of the device or unit may be in an electrical, mechanical or other form.

[0174] In addition, the units described as separate components may or may not be physically separated, and the components displayed as units may or may not be physical units. They may be located in one place or distributed to multiple network units. Some or all of the units can be selected according to actual needs to achieve the purpose of the solution of this embodiment.

[0175] Furthermore, in each embodiment of the present application, each functional module may be integrated together to form an independent part, or each module may exist alone, or two or more modules may be integrated to form an independent part.

[0176] In this text, relational terms such as first and second are only used to distinguish one entity or operation from another entity or operation, and do not necessarily require or imply any actual relationship or order between these entities or operations.

[0177] The above description is only for the embodiments of the present application and is not intended to limit the protection scope of the present application. For those skilled in the art, the present application may have various changes and modifications. Any modification, equivalent replacement, improvement, etc. made within the spirit and principle of the present application shall be included in the protection scope of the present application.

Claims

1. A power load scheduling method for optimizing the scheduling of power loads, characterized in that, Including the steps: Obtain the load information of the loads within the area to be measured; the load information includes the number of all loads and the operating conditions of all loads within a preset period; According to the number and the operating conditions, use the spectral clustering algorithm to divide the loads to obtain multiple load groups; Based on the operating conditions of the multiple load groups, construct a likelihood function based on the Markov random field to calculate the planned operating conditions of the multiple load groups through Kalman filtering; According to the planned operating conditions, perform operating scheduling on the corresponding load groups; Based on the operating conditions of the multiple load groups, construct a likelihood function based on the Markov random field to calculate the planned operating conditions of the multiple load groups through Kalman filtering, including: Based on the operating conditions of the multiple load groups, construct a corresponding Markov random function; Add a correction term to the Markov random function to construct a likelihood function based on the Markov random field; Through Kalman filtering, iterate the likelihood function based on the Markov random field to calculate the planned operating conditions of the multiple load groups; The correction term includes an observation matrix correction term and a noise correction term; adding a correction term to the Markov random function to construct a likelihood function based on the Markov random field includes: Add the noise correction term to the Markov random function to obtain an optimized Markov random function; According to the optimized Markov random function and the actual operating conditions of the corresponding load groups, combined with the observation matrix correction term, construct an operating condition function; Based on the noise correction term, the optimized Markov random function, the observation matrix correction term, and the operating condition function, construct a likelihood function based on the Markov random field.

2. The power load scheduling method according to claim 1, wherein According to the number and the operating conditions, use the spectral clustering algorithm to divide the loads to obtain multiple load groups, including: According to the number, construct an adjacency matrix of the loads regarding the operating conditions; According to the Laplacian matrix corresponding to the adjacency matrix, construct an eigenvector space; Through the K-means clustering algorithm, based on the eigenvectors in the eigenvector space, divide the loads to obtain multiple load groups.

3. The power load scheduling method according to claim 2, wherein According to the number, construct an adjacency matrix of the loads regarding the operating conditions, including: Construct an operating matrix of the number of the loads regarding the operating conditions; Adopt the difference method to establish a differential operating matrix corresponding to the operating matrix; According to the specific values of the elements in the differential operating matrix, construct an adjacency matrix of the loads regarding the operating conditions.

4. The power load scheduling method according to claim 2, wherein Through the K-means clustering algorithm, based on the eigenvectors in the eigenvector space, divide the loads to obtain multiple load groups, including: Based on the eigenvectors in the eigenvector space, perform dimensionality reduction representation on the operating conditions of the loads to obtain load information after dimensionality reduction; According to the load information after dimensionality reduction, select multiple loads from all the loads as multiple initial clustering centers; Based on the initial cluster centers, group all the loads through the K-means clustering algorithm to obtain multiple load groups.

5. The power load scheduling method according to claim 4, wherein Based on the initial cluster centers, group all the loads through the K-means clustering algorithm to obtain multiple load groups, including: Step A1, calculate the distances between the other loads except the initial cluster centers among all the loads and the multiple initial cluster centers; Step A2, allocate the other loads to the nearest initial cluster center to form primary load groups; Step A3, recalculate the cluster centers in the primary load groups, and re-allocate all the other loads except the cluster centers according to the distances between the cluster centers and all the other loads; Step A4, repeatedly execute Step A3 until the cluster centers no longer change, and obtain multiple load groups.

6. A power load scheduling device for optimizing the scheduling of power loads, characterized in that, Including: An acquisition module, configured to acquire the load information of the loads in the area to be measured; the load information includes the quantity of all the loads and the operating conditions of all the loads within a preset period; A partitioning module, configured to partition the loads by using the spectral clustering algorithm according to the quantity and the operating conditions to obtain multiple load groups; A calculation module, configured to construct a likelihood function based on the Markov random field based on the operating conditions of the multiple load groups, so as to calculate the planned operating conditions of the multiple load groups through Kalman filtering; A scheduling module, configured to perform operating scheduling on the corresponding load groups according to the planned operating conditions; The calculation module, configured to construct a likelihood function based on the Markov random field based on the operating conditions of the multiple load groups, so as to calculate the planned operating conditions of the multiple load groups through Kalman filtering, including: Construct a corresponding Markov random function based on the operating conditions of the multiple load groups; Add a correction term to the Markov random function to construct a likelihood function based on the Markov random field; Iterate the likelihood function based on the Markov random field through Kalman filtering to calculate the planned operating conditions of the multiple load groups; The correction term includes an observation matrix correction term and a noise correction term; adding a correction term to the Markov random function to construct a likelihood function based on the Markov random field includes: Add the noise correction term to the Markov random function to obtain an optimized Markov random function; Construct an operating condition function according to the optimized Markov random function and the actual operating conditions of the corresponding load groups in combination with the observation matrix correction term; Construct a likelihood function based on the Markov random field based on the noise correction term, the optimized Markov random function, the observation matrix correction term, and the operating condition function.

7. An electronic device, characterized in that, Including a processor and a memory, the memory stores a computer program executable by the processor, and when the processor executes the computer program, it runs the steps in the power load scheduling method according to any one of claims 1-5.

8. A computer-readable storage medium having a computer program stored thereon, characterized in that, When the computer program is executed by a processor, it runs the steps in the power load scheduling method according to any one of claims 1-5.

Images