CN116937563A - Space load prediction method and system considering space-time correlation - Google Patents
Space load prediction method and system considering space-time correlation Download PDFInfo
- Publication number
- CN116937563A CN116937563A CN202310912496.9A CN202310912496A CN116937563A CN 116937563 A CN116937563 A CN 116937563A CN 202310912496 A CN202310912496 A CN 202310912496A CN 116937563 A CN116937563 A CN 116937563A
- Authority
- CN
- China
- Prior art keywords
- load
- class
- cell
- cells
- time
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Pending
Links
- 238000000034 method Methods 0.000 title claims abstract description 52
- 239000011159 matrix material Substances 0.000 claims abstract description 104
- 238000010586 diagram Methods 0.000 claims abstract description 78
- 238000004364 calculation method Methods 0.000 claims description 22
- 238000001514 detection method Methods 0.000 claims description 18
- 230000002159 abnormal effect Effects 0.000 claims description 11
- 230000008859 change Effects 0.000 description 5
- 230000000694 effects Effects 0.000 description 5
- 238000005259 measurement Methods 0.000 description 5
- 238000005065 mining Methods 0.000 description 5
- 238000010276 construction Methods 0.000 description 4
- 238000005516 engineering process Methods 0.000 description 4
- 238000013450 outlier detection Methods 0.000 description 4
- 238000007781 pre-processing Methods 0.000 description 4
- 238000007405 data analysis Methods 0.000 description 3
- 230000003247 decreasing effect Effects 0.000 description 3
- 230000003993 interaction Effects 0.000 description 3
- 230000002123 temporal effect Effects 0.000 description 3
- 229940060587 alpha e Drugs 0.000 description 2
- 238000004458 analytical method Methods 0.000 description 2
- 238000013473 artificial intelligence Methods 0.000 description 2
- 230000008901 benefit Effects 0.000 description 2
- 230000004807 localization Effects 0.000 description 2
- 238000005192 partition Methods 0.000 description 2
- NAWXUBYGYWOOIX-SFHVURJKSA-N (2s)-2-[[4-[2-(2,4-diaminoquinazolin-6-yl)ethyl]benzoyl]amino]-4-methylidenepentanedioic acid Chemical compound C1=CC2=NC(N)=NC(N)=C2C=C1CCC1=CC=C(C(=O)N[C@@H](CC(=C)C(O)=O)C(O)=O)C=C1 NAWXUBYGYWOOIX-SFHVURJKSA-N 0.000 description 1
- 238000013528 artificial neural network Methods 0.000 description 1
- 230000000295 complement effect Effects 0.000 description 1
- 238000000354 decomposition reaction Methods 0.000 description 1
- 230000007547 defect Effects 0.000 description 1
- 238000009792 diffusion process Methods 0.000 description 1
- 230000006872 improvement Effects 0.000 description 1
- 238000000691 measurement method Methods 0.000 description 1
- 230000008569 process Effects 0.000 description 1
- 230000000750 progressive effect Effects 0.000 description 1
- 230000009467 reduction Effects 0.000 description 1
- 238000005070 sampling Methods 0.000 description 1
- 238000010183 spectrum analysis Methods 0.000 description 1
- 230000007480 spreading Effects 0.000 description 1
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06Q—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
- G06Q10/00—Administration; Management
- G06Q10/06—Resources, workflows, human or project management; Enterprise or organisation planning; Enterprise or organisation modelling
- G06Q10/063—Operations research, analysis or management
- G06Q10/0631—Resource planning, allocation, distributing or scheduling for enterprises or organisations
- G06Q10/06312—Adjustment or analysis of established resource schedule, e.g. resource or task levelling, or dynamic rescheduling
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F17/00—Digital computing or data processing equipment or methods, specially adapted for specific functions
- G06F17/10—Complex mathematical operations
- G06F17/11—Complex mathematical operations for solving equations, e.g. nonlinear equations, general mathematical optimization problems
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06Q—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
- G06Q10/00—Administration; Management
- G06Q10/06—Resources, workflows, human or project management; Enterprise or organisation planning; Enterprise or organisation modelling
- G06Q10/063—Operations research, analysis or management
- G06Q10/0631—Resource planning, allocation, distributing or scheduling for enterprises or organisations
- G06Q10/06313—Resource planning in a project environment
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06Q—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
- G06Q50/00—Information and communication technology [ICT] specially adapted for implementation of business processes of specific business sectors, e.g. utilities or tourism
- G06Q50/06—Energy or water supply
-
- H—ELECTRICITY
- H02—GENERATION; CONVERSION OR DISTRIBUTION OF ELECTRIC POWER
- H02J—CIRCUIT ARRANGEMENTS OR SYSTEMS FOR SUPPLYING OR DISTRIBUTING ELECTRIC POWER; SYSTEMS FOR STORING ELECTRIC ENERGY
- H02J3/00—Circuit arrangements for ac mains or ac distribution networks
- H02J3/003—Load forecast, e.g. methods or systems for forecasting future load demand
-
- H—ELECTRICITY
- H02—GENERATION; CONVERSION OR DISTRIBUTION OF ELECTRIC POWER
- H02J—CIRCUIT ARRANGEMENTS OR SYSTEMS FOR SUPPLYING OR DISTRIBUTING ELECTRIC POWER; SYSTEMS FOR STORING ELECTRIC ENERGY
- H02J2203/00—Indexing scheme relating to details of circuit arrangements for AC mains or AC distribution networks
- H02J2203/20—Simulating, e g planning, reliability check, modelling or computer assisted design [CAD]
Landscapes
- Business, Economics & Management (AREA)
- Engineering & Computer Science (AREA)
- Human Resources & Organizations (AREA)
- Physics & Mathematics (AREA)
- Economics (AREA)
- General Physics & Mathematics (AREA)
- Theoretical Computer Science (AREA)
- Strategic Management (AREA)
- Entrepreneurship & Innovation (AREA)
- Mathematical Physics (AREA)
- General Business, Economics & Management (AREA)
- Operations Research (AREA)
- Tourism & Hospitality (AREA)
- Marketing (AREA)
- Pure & Applied Mathematics (AREA)
- Data Mining & Analysis (AREA)
- Educational Administration (AREA)
- Game Theory and Decision Science (AREA)
- Quality & Reliability (AREA)
- Health & Medical Sciences (AREA)
- Development Economics (AREA)
- Mathematical Optimization (AREA)
- Mathematical Analysis (AREA)
- Computational Mathematics (AREA)
- Primary Health Care (AREA)
- General Health & Medical Sciences (AREA)
- Water Supply & Treatment (AREA)
- Public Health (AREA)
- Biodiversity & Conservation Biology (AREA)
- Algebra (AREA)
- Life Sciences & Earth Sciences (AREA)
- Databases & Information Systems (AREA)
- Software Systems (AREA)
- General Engineering & Computer Science (AREA)
- Power Engineering (AREA)
- Management, Administration, Business Operations System, And Electronic Commerce (AREA)
Abstract
The invention discloses a space load prediction method and a space load prediction system considering space-time correlation, and relates to the technical field of power distribution network planning. The method comprises the following steps: obtaining a power geographic information system consisting of a plurality of I-type cells corresponding to an area to be planned and a power geographic information system consisting of a plurality of II-type cells; obtaining geographic parameters between every two class I cells according to a power geographic information system consisting of a plurality of class I cells; obtaining a space weight matrix according to the geographic parameters; obtaining a time weight matrix according to the load time sequence of each class I cell; obtaining local outlier factors of each point in the time-load power scatter diagram according to the load time sequence; obtaining a reference year load matrix according to local outlier factors of each point; and obtaining the space load predicted value of each type II cell in the area to be planned according to the three matrixes and the electric power geographic information system consisting of a plurality of type II cells. The method and the device can improve the accuracy of the space load prediction result.
Description
Technical Field
The invention relates to the technical field of power distribution network planning, in particular to a space load prediction method and a space load prediction system considering space-time correlation.
Background
With the continuous development and construction of cities, population flow is accelerated, and the spatial characteristics and the regional characteristics of the electric power load of the urban power grid are increasingly prominent. In the load prediction, only the magnitude of the predicted load does not give the spatial and regional distribution situation of the load, and the fine requirement of urban network planning cannot be met, so that the spatial power load prediction, also called spatial load prediction, is needed.
The space load prediction is an essential link before urban network planning, and accurate space load prediction has important significance for improving the reliability and economy of urban network planning, but the existing space load prediction method focuses on the time sequence rule of the mining load, namely only the time correlation among the loads is considered, and the problem of insufficient consideration of the space change condition of the load is solved, so that the final load prediction result is inaccurate.
Disclosure of Invention
The invention aims to provide a space load prediction method and a space load prediction system considering space-time correlation, which can improve the accuracy of a space load prediction result.
In order to achieve the above object, the present invention provides the following solutions:
a spatial load prediction method considering spatio-temporal correlation, comprising:
Processing the power geographic information system of the area to be planned to obtain a power geographic information system composed of a plurality of I-type cells and a power geographic information system composed of a plurality of II-type cells; the electric power geographic information system formed by a plurality of I-type cells is obtained by drawing a cell layer in the electric power geographic information system of the area to be planned; the power geographic information system formed by a plurality of class II cells is obtained by carrying out grid division on the power geographic information system of the area to be planned;
obtaining geographic parameters between every two I-type cells in all I-type cells according to the electric power geographic information system consisting of a plurality of I-type cells; the geographic parameters comprise distance and shared boundary coefficients;
obtaining a space weight matrix of the load of the area to be planned according to the geographic parameters between every two I-type cells in all I-type cells;
obtaining a time weight matrix of the load of the area to be planned according to the Pearson correlation coefficient of the load time sequence of every two I-type cells in each I-type cell; the load time sequence comprises load power at a plurality of historical moments;
for any I-type cell, obtaining a time-load power scatter diagram of the I-type cell according to the load time sequence of the I-type cell, and processing the time-load power scatter diagram of the I-type cell by adopting a grid local outlier factor detection algorithm to obtain local outlier factors of each point in the time-load power scatter diagram of the I-type cell;
Obtaining the reference load of the class I cell according to the local outlier factors of each point in the time-load power scatter diagram of the class I cell and the time-load power scatter diagram of the class I cell;
determining the reference load of all the class I cells as a reference year load matrix of the area to be planned;
and obtaining the space load predicted value of each type II cell in the area to be planned according to the reference year load matrix of the area to be planned, the space weight matrix of the load of the area to be planned, the time weight matrix of the load of the area to be planned and a power geographic information system consisting of a plurality of type II cells.
Optionally, obtaining the shared boundary coefficient between every two class I cells in all class I cells according to the power geographic information system composed of a plurality of class I cells specifically includes:
for the ith and jth class I cells, according to the formula
Calculating a shared boundary coefficient between the ith class I cell and the jth class I cell; wherein b ij For the shared boundary coefficient between the ith class I cell and the jth class I cell, l ij For the length of the shared boundary between the ith class I cell and the jth class I cell, l i,min Is the most shared boundary length among all class I cells adjacent to the I-th class I cell Small value, l i,max Is the maximum value of the shared boundary length in all class I cells adjacent to the I-th class I cell.
Optionally, the spatial weight matrix of the load of the area to be planned is obtained according to the geographic parameters between every two class I cells in all class I cells, which specifically includes:
obtaining the space weight between the loads of every two cells in all the class I cells according to the geographic parameters between every two class I cells in all the class I cells;
and determining a matrix formed by the space weights between every two cell loads in all the class I cells as a space weight matrix of the area load to be planned.
Optionally, the time weight matrix of the load of the area to be planned is obtained according to pearson correlation coefficients of the load time sequence of every two class I cells in each class I cell, which specifically includes:
calculating the pearson correlation coefficient of the load time sequences of every two class I cells in all class I cells according to the load time sequences of all class I cells to obtain the time correlation coefficient between the loads of every two cells;
and determining a matrix formed by time correlation coefficients between loads of every two cells in all the class I cells as a load time weight matrix of the area to be planned.
Optionally, obtaining the reference load of the class I cell according to the local outlier factor of each point in the time-load power scatter diagram of the class I cell and the time-load power scatter diagram of the class I cell, which specifically includes:
Deleting points with the local outlier factor larger than a set threshold value from the time-load power scatter diagram of the class I cell to obtain an abnormal scatter diagram of the class I cell;
and determining the maximum value of the load power corresponding to all points in the abnormal scatter diagram of the class I cell removal as the reference load of the class I cell.
Optionally, the obtaining the space load prediction value of each II-type cell in the area to be planned according to the reference year load matrix of the area to be planned, the space weight matrix of the load of the area to be planned, the time weight matrix of the load of the area to be planned and the power geographic information system composed of a plurality of II-type cells specifically includes:
obtaining a space load predicted value of each class I cell in the area to be planned according to the reference year load matrix of the area to be planned, the space weight matrix of the load of the area to be planned and the time weight matrix of the load of the area to be planned;
and obtaining the space load predicted value of each class II cell in the area to be planned according to the space load predicted value of each class I cell in the area to be planned and a power geographic information system consisting of a plurality of class II cells.
Optionally, obtaining a time-load power scatter diagram of the class I cell according to the load time sequence of the class I cell, and processing the time-load power scatter diagram of the class I cell by adopting a grid local outlier factor detection algorithm to obtain local outliers of each point in the time-load power scatter diagram of the class I cell, which specifically includes:
obtaining a time-load power scatter diagram of the class I cell according to the load time sequence of the class I cell;
performing grid division on the time-load power scatter diagram of the cell to obtain a grid scatter diagram; each grid in the grid scatter plot comprises one or more points;
and calculating local outlier factors of points in all boundary grids in the grid scatter diagram by adopting an LOF algorithm.
A spatial load prediction system that accounts for spatio-temporal correlation, comprising:
the power geographic information system processing module is used for processing a power geographic information system of an area to be planned to obtain a power geographic information system composed of a plurality of I-type cells and a power geographic information system composed of a plurality of II-type cells; the electric power geographic information system formed by a plurality of I-type cells is obtained by drawing a cell layer in the electric power geographic information system of the area to be planned; the power geographic information system formed by a plurality of class II cells is obtained by carrying out grid division on the power geographic information system of the area to be planned;
The geographic parameter determining module is used for obtaining geographic parameters between every two I-type cells in all I-type cells according to the electric power geographic information system consisting of a plurality of I-type cells; the geographic parameters comprise distance and shared boundary coefficients;
the space weight matrix calculation module is used for obtaining a space weight matrix of the load of the area to be planned according to the geographic parameters between every two I-type cells in all I-type cells;
the time weight matrix calculation module obtains a time weight matrix of the load of the area to be planned according to the Pearson correlation coefficient of the load time sequence of each two I-type cells in each I-type cell; the load time sequence comprises load power at a plurality of historical moments;
the local outlier factor calculation module is used for obtaining a time-load power scatter diagram of any I-type cell according to the load time sequence of the I-type cell, and processing the time-load power scatter diagram of the I-type cell by adopting a grid local outlier factor detection algorithm to obtain local outlier factors of each point in the time-load power scatter diagram of the I-type cell;
the reference load determining module is used for obtaining the reference load of the class I cell according to the local outlier factors of each point in the time-load power scatter diagram of the class I cell and the time-load power scatter diagram of the class I cell;
The reference year load matrix determining module is used for determining that the reference load of all the class I cells is the reference year load matrix of the area to be planned;
the space load prediction value calculation module is used for obtaining the space load prediction value of each type II cell in the area to be planned according to the reference year load matrix of the area to be planned, the space weight matrix of the load of the area to be planned, the time weight matrix of the load of the area to be planned and the power geographic information system consisting of a plurality of type II cells.
According to the specific embodiment provided by the invention, the invention discloses the following technical effects:
according to the method, geographic parameters between every two I-type cells in all I-type cells are obtained according to a power geographic information system formed by a plurality of I-type cells, then a space weight matrix of the load of a region to be planned is obtained according to the geographic parameters between every two I-type cells in all I-type cells, the space change of the load is considered, a time weight matrix of the load of the region to be planned is obtained according to the load time sequence of each I-type cell, the time sequence change of the load is considered, the historical load power which does not meet the condition is removed by adopting a grid local outlier factor detection algorithm, a reference year load matrix is obtained, the space load matrix, the space weight matrix and a power geographic information system formed by a plurality of II-type cells are obtained, the space load prediction value of each II-type cell in the region to be planned is mined from the space-time aspect, the defects that the existing space load prediction method is focused on mining the time sequence rule and the space correlation of the load is insufficient are overcome, the accuracy of the space load prediction result is improved, and the urban space load prediction value of each II-type cell in the region to be planned in the recent region can be guided.
Drawings
In order to more clearly illustrate the embodiments of the present invention or the technical solutions of the prior art, the drawings that are needed in the embodiments will be briefly described below, it being obvious that the drawings in the following description are only some embodiments of the present invention, and that other drawings may be obtained according to these drawings without inventive effort for a person skilled in the art.
FIG. 1 is a flow chart of a spatial load prediction method considering space-time correlation provided in embodiment 1 of the present invention;
FIG. 2 is a flowchart of a spatial load prediction method considering space-time correlation according to embodiment 2 of the present invention;
FIG. 3 is a schematic diagram of the generation of class I cells according to embodiments 1 and 2 of the present invention;
FIG. 4 is a schematic diagram of the generation of class II cells according to embodiments 1 and 2 of the present invention;
FIG. 5 is a schematic diagram of the relationship between nodes and edges corresponding to class I cells according to embodiment 2 of the present invention;
FIG. 6 is a schematic diagram of a specific prediction strategy according to embodiment 3 of the present invention;
fig. 7 is a graph showing the spatial load prediction result and quasi-actual measurement result of each prediction method based on class II cells according to embodiment 3 of the present invention.
Detailed Description
The following description of the embodiments of the present invention will be made clearly and completely with reference to the accompanying drawings, in which it is apparent that the embodiments described are only some embodiments of the present invention, but not all embodiments. All other embodiments, which can be made by those skilled in the art based on the embodiments of the invention without making any inventive effort, are intended to be within the scope of the invention.
In order that the above-recited objects, features and advantages of the present invention will become more readily apparent, a more particular description of the invention will be rendered by reference to the appended drawings and appended detailed description.
Example 1
The embodiment of the invention provides a space load prediction method considering space-time correlation, which is shown in fig. 1 and comprises the following steps:
s1: processing the power geographic information system of the area to be planned to obtain a power geographic information system composed of a plurality of I-type cells as shown in figure 3 and a power geographic information system composed of a plurality of II-type cells as shown in figure 4; the electric power geographic information system formed by a plurality of I-type cells is obtained by drawing a cell layer in the electric power geographic information system of the area to be planned; the power geographic information system formed by a plurality of class II cells is obtained by carrying out grid division on the power geographic information system of the area to be planned.
S2: obtaining geographic parameters between every two I-type cells in all I-type cells according to the electric power geographic information system consisting of a plurality of I-type cells; the geographic parameters include distance and shared boundary coefficients.
S3: and obtaining the space weight matrix of the load of the area to be planned according to the geographic parameters between every two I-type cells in all I-type cells.
S4: obtaining a time weight matrix of the load of the area to be planned according to the Pearson correlation coefficient of the load time sequence of every two I-type cells in each I-type cell; the load time series includes load power at a plurality of historical moments.
S5: for any I-type cell, obtaining a time-load power scatter diagram of the I-type cell according to the load time sequence of the I-type cell, and processing the time-load power scatter diagram of the I-type cell by adopting a grid local outlier factor detection algorithm to obtain local outlier factors of points in the time-load power scatter diagram of the I-type cell.
S6: and obtaining the reference load of the class I cell according to the local outlier factors of each point in the time-load power scatter diagram of the class I cell and the time-load power scatter diagram of the class I cell.
S7: and determining the reference load of all the class I cells as a reference year load matrix of the area to be planned.
S8: and obtaining the space load predicted value of each type II cell in the area to be planned according to the reference year load matrix of the area to be planned, the space weight matrix of the load of the area to be planned, the time weight matrix of the load of the area to be planned and a power geographic information system consisting of a plurality of type II cells.
In practical application, the spatial weight matrix of the load of the area to be planned is obtained according to the geographic parameters between every two I-type cells in all I-type cells, which specifically comprises:
and obtaining the space weight between the loads of every two cells in all the class I cells according to the geographic parameters between every two class I cells in all the class I cells.
And determining a matrix formed by the space weights between every two cell loads in all the class I cells as a space weight matrix of the area load to be planned.
In practical application, the time weight matrix of the load of the area to be planned is obtained according to pearson correlation coefficient of load time sequences of every two I-type cells in each I-type cell, which specifically comprises the following steps:
and calculating the pearson correlation coefficient of the load time sequences of every two class I cells in all class I cells according to the load time sequences of all class I cells to obtain the time correlation coefficient between the loads of every two cells.
And determining a matrix formed by time correlation coefficients between loads of every two cells in all the class I cells as a load time weight matrix of the area to be planned.
In practical application, obtaining the reference load of the class I cell according to the local outlier factor of each point in the time-load power scatter diagram of the class I cell and the time-load power scatter diagram of the class I cell, specifically including:
And deleting the points with the local outlier factors larger than a set threshold value from the time-load power scatter diagram of the class I cell to obtain an abnormal scatter diagram of the class I cell.
And determining the maximum value of the load power corresponding to all points in the abnormal scatter diagram of the class I cell removal as the reference load of the class I cell.
In practical application, according to the reference year load matrix of the area to be planned, the space weight matrix of the load of the area to be planned, the time weight matrix of the load of the area to be planned and the power geographic information system composed of a plurality of class II cells, the method for obtaining the space load prediction value of each class II cell in the area to be planned specifically includes:
and obtaining the space load predicted value of each class I cell in the area to be planned according to the reference year load matrix of the area to be planned, the space weight matrix of the load of the area to be planned and the time weight matrix of the load of the area to be planned.
And obtaining the space load predicted value of each class II cell in the area to be planned according to the space load predicted value of each class I cell in the area to be planned and a power geographic information system consisting of a plurality of class II cells.
In practical application, a time-load power scatter diagram of the class I cell is obtained according to the load time sequence of the class I cell, and the grid local outlier factor detection algorithm is adopted to process the time-load power scatter diagram of the class I cell to obtain local outlier factors of points in the time-load power scatter diagram of the class I cell, which specifically comprises:
and obtaining a time-load power scatter diagram of the class I cell according to the load time sequence of the class I cell.
Performing grid division on the time-load power scatter diagram of the cell to obtain a grid scatter diagram; each grid within the grid scatter plot includes one or more points therein.
And calculating local outlier factors of points in all boundary grids in the grid scatter diagram by adopting an LOF algorithm.
Example 2
The present invention provides a more specific embodiment for describing the above method in detail, referring to fig. 2, the specific steps are as follows:
1) Spatial localization of metropolitan network load (localization of metropolitan network spatial load).
1. Establishing a power geographic information system, drawing a cell layer in the power geographic information system, wherein the positioning of loads in the urban power grid takes cells as minimum units, the cells refer to power supply small areas divided according to different modes, the power supply areas of the urban power grid are divided according to the power distribution range of unit facilities, and each irregularly-powered cell after division is simply called as a class I cell; and the urban power grid coverage area is partitioned according to regular grids with equal size, each partitioned grid is called a class II cell, and the class I cell and the class II cell are shown in fig. 3 and 4.
2. And positioning the load space load of the urban network based on the class I cells.
3. The load of the class I cells is converted into the load of the class II cells through a gridding technology, so that the positioning of the load space load of the urban network based on the class II cells is realized.
2) Measurement of load spatiotemporal correlation (measurement of spatiotemporal correlation between individual class I cell loads).
1. Measuring the spatial correlation between the class I cell loads:
establishing a space weight matrix W of an area to be planned s ，W s The spatial interdependence degree and interaction between the cell loads of each class I in the area to be planned are represented as a formula (1), wherein the expression is a symmetric matrix with 0 diagonal elements:
wherein: m is the number of class I cells in the area to be planned;the spatial weight between the I-th class I cell load and the j-th class I cell load; i=1, 2, …, m; j=1, 2, …, m; and when i=j,/is> The calculation formula of (2) is shown as the formula:
wherein: b ij A shared boundary coefficient between the ith class I cell and the jth cell; d, d ij The Euclidean distance between the I-th class I cell load and the j-th class I cell load; lambda is an exogenous coefficient of a magnitude equal to d in the area to be planned ij Average value of (2).
The closer the spatial distance between class I cell loads, the stronger the interaction force between them, the more likely propagation and diffusion will occur; thus, the spatial weight between the I-th and j-th class I cell loads Distance d from the two class I cells ij Exhibit the inverse distance effect, i.e. d ij The greater the spatial weight ∈ ->The smaller; spatial distance d between cells ij Is calculated by Euclidean distance formula of the geometric center of each class I cell, as shown in formula (3),
wherein: d, d ij Is the Euclidean distance between the I-th class I cell load and the j-th cell load; x is x i And x j Respectively the abscissa of the geometric centers of the ith class I cell and the jth class I cell; y is i And y j The ordinate of the geometric centers of the ith class I cell and the jth class I cell respectively.
In addition, the difference of the shared boundary lengths between the adjacent I-type cells also causes the difference of the influence intensity of the load between the cells, and the longer the shared boundary, the more frequent the flow and interaction between the cells, the stronger the influence, i.e. the spatial correlation coefficient between the two I-type cell loads and the shared boundary length of the two I-type cells show positive correlation, and the shared boundary coefficient b between the ith I-type cell and the jth I-type cell ij The calculation formula of (2) is shown as formula (4):
wherein: l (L) ij The length of the shared boundary between the ith class I cell and the jth class I cell; l (L) i,min And l i,max Respectively the minimum value and the maximum value of the shared boundary length in all the I type cells adjacent to the I type cell; i=1, 2, …, m; j=1, 2, …, m.
2. Measuring the time correlation between the cell loads of the class I cells:
the invention utilizes the pearson correlation coefficient to measure the correlation between the load time sequences of all I cells, and the calculation formula of the pearson correlation coefficient is shown as the formula (5):
wherein:a time correlation coefficient between the I-th class I cell load and the j-th class I cell load; p is p i And p j The load time sequence of the ith class I cell and the load time sequence of the jth class I cell are respectively; cov (p) i ,p j ) Is p i And p j Covariance between; sigma (p) i ) Sum sigma (p) j ) P is respectively i And p j Standard deviation of (2).
Unlike the spatial weight matrix that measures the spatial correlation of class I cell load, the temporal weight matrix W that measures the temporal correlation of class I cell load t Is a symmetrical matrix with diagonal elements of 1, and the expression is shown in formula (6);
3) Construction of a spatial load prediction model (construction of a spatial load prediction model).
1. Data outlier detection module for constructing class I cell load time sequence
In the abnormal value detection module, carrying out abnormal value detection of the load by adopting a grid local outlier factor detection algorithm; the grid data analysis is a proprietary term of a grid local outlier factor detection algorithm, and the grid division corresponding to the grid data analysis refers to grid division of a load time sequence scatter diagram plane of each class I cell, and has no relation with grid division of a geographic area to be planned corresponding to generation of class II cells.
A. Grid data analysis of class I cell load time series
First, t-p scatter diagrams of the load time series of each class I cell are generated respectively, tFor sampling time, p is the load power of the class I cells; then dividing the t-p plane where each I-type cell load time sequence is located into a limited grid containing a certain number of load points (namely points in a scatter diagram), and distributing each load point in the grid; by comparing the number of load points contained in adjacent grids of each grid with a threshold value N th To determine whether the grid is an internal grid or a boundary grid; finally, only the load points in the boundary grid are reserved to participate in LOF algorithm detection, while the load points in the internal grid are not participated, and the detailed steps are as follows:
a. respectively inputting the load time sequences of the I-type cells, generating a t-p scatter diagram of the load time sequences of each I-type cell, and determining the grid number h divided by the t-p plane where the load time sequences of each I-type cell are located t ×h p 。
b. Calculating the lateral length l of each grid t Longitudinal length l p As shown in formulas (7) and (8):
wherein: l (L) t For the lateral length of each grid; l (L) p For the longitudinal length of each grid; h is a t The number of the grid partitions is transversely divided; h is a p The number of the grid partitions is divided longitudinally; t is t max Loading the last moment of the time sequence for each class I cell; t is t min Loading an initial time of the time sequence for each class I cell; p is p max Loading a maximum load power in the time series for each class I cell; p is p min The minimum load power in the time series is loaded for each class I cell.
c. And determining the number of the grid where each load point is located in each class I cell load time sequence, and counting the number of the load points in each grid.
d. Comparing eachThe adjacent grids of the grids contain load points and a threshold value N th The magnitude relation between adjacent grids contains load points less than a threshold value N th The grids of the class I cells are judged to be boundary grids, the boundary grids of each class I cell and the load points in the boundary grids are output, and only the load points in the boundary grids are reserved for data outlier detection based on a local outlier factor algorithm.
B. Local outlier factor detection for load points in boundary grids of each class I cell
a. Determining the kth distance of load points in the boundary grid of each class I cell time sequence t-p plane:
in the boundary grid data set D of each class I cell, the distance between the point p and the point o is denoted as D (p, o), and for any positive integer k, the kth distance of the point p is denoted as k_dist (p), while D (p, o) satisfying the following 2 conditions is equal to k_dist (p):
Condition 1: in the boundary grid dataset D of each class I cell time series t-p plane, there are at least k points o 'such that D (p, o'). Ltoreq.d (p, o).
Condition 2: in the boundary grid dataset D of each class I cell time series t-p plane there are at most k-1 points o 'such that D (p, o') < D (p, o).
b. Determining the k neighborhood of each load point in the boundary grid of each class I cell time sequence t-p plane:
k-th distance neighborhood N of point p k (p) is a set of points where all distances to the point p do not exceed the kth-distance, as shown in formula (9):
N k (p)＝{q∈D\{p}|d(q,p)≤k_dist(p)} (9)
c. determining the reachable distance (p, o) of each load point in the boundary grid of each class I cell time sequence t-p plane:
the reachable distance of the point p with respect to the point o is denoted as reach_dist (p, o), as shown in formula (10):
reach_dist(p,o)＝max{d(p,o),k_dist(p,o)} (10)
d. calculating the local reachable density lrd of each load point in the boundary grid of each class I cell time sequence t-p plane k (p)：
The local reachable density of a point p is defined as the inverse of the average reachable distance from the point to the point p in the kth neighborhood of the point p, denoted as lrd k (p) the calculation formula is shown as formula (11);
wherein: n k (p) is the number of points in the k-distance neighborhood of point p.
e. Calculating local outlier factors LOF of each load point in each class I cell boundary grid k (p)：
The local outlier factor of point p is point N in the kth neighborhood of point p k (p) average value of ratio of local reachable density to local reachable density of point p, expressed as LOF k (p) as shown in formula (12):
wherein: lrd k (o) local reachable densities for points in the k neighborhood of p; lrd k (p) is the locally attainable density of p-dots.
For points p with higher outliers, LOF k (p) is relatively large; for points p with a low degree of outlier, LOF k (P) is small, so that load points with local outlier factors larger than a set threshold value are deleted from the scatter diagram, the maximum value of load power corresponding to all points in the scatter diagram of the deleted load points is used as the reference load of the class I cells, and a matrix formed by the reference loads of all the class I cells is determined to be the reference year load matrix P of the area to be planned 0 。
2. Space-time rule mining module for constructing I-type cell load time sequence
The invention abstracts the time-space relationship between the load time sequence of the class I cells and the load of the class I cells in the area to be planned into node and edge representation: each class I cell is abstracted into a node, the number of the nodes is equal to the number m of the class I cells in the area to be planned,i.e.The spatial relationship between two class I cells is abstracted into edges, and a schematic diagram of the relationship between the nodes corresponding to the class I cells and the edges is shown in fig. 5.
The spatial influence of a node i by neighboring nodes can be abstracted into a functional expression, as shown in equation (13):
wherein:is node x i Is influenced by the space of other nodes; p is p j Is the load of node j; />Is the spatial correlation weight of the ith node and the jth node.
The timing effect of a node i on neighboring nodes can be abstracted into a functional expression, as shown in equation (14):
wherein:is node x i Affected by the timing of other nodes; p is p j Is the load of node j; />Is the time correlation weight of the ith node and the jth node.
Spreading to all nodes of the whole area to be planned, wherein the load predicted value of each node is the load predicted value P of each I-type cell fore The calculated spatiotemporal convolution expression is shown as (15):
P fore ＝P 0 +f ts P 0 ＝P 0 +αW s *(1-α)W t P 0 (15)
wherein:space-time influence operators for each node in the whole area to be planned, which are influenced by other nodes; p (P) 0 The reference year load matrix of all nodes in the area to be planned is obtained; alpha E [0,1 ]]Is a space-time weight coefficient; * Is a convolution symbol; when only the spatial influence is considered, ftsP0 in the formula (15) is reduced to +.>Performing calculation, i.e.)>Wherein->Obtained by the formula (13); when only the spatial influence is considered, ftsP0 in the formula (15) is reduced to +.>Performing calculations, i.e. Wherein->Obtained by the formula (14); when both temporal and spatial effects are considered, according to P fore ＝P 0 +αW s *(1-α)W t P 0 Calculation of P fore Irrespective of the contents of equation (13) and equation (14), P 0 The reference year load matrix of all nodes in the area to be planned is obtained; alpha E [0,1 ]]Is a space-time weight coefficient; * Is a convolution symbol.
Predicting the load predictive value P of each class I cell in the area to be planned fore Passing through a network in a geographic information systemThe meshing technology can be converted into a space load prediction result of the II type cell.
The embodiment is based on urban network space load positioning of cells, class I cell load space-time correlation measurement based on space-time matrix, grid local outlier factor detection algorithm and space-time convolution space load prediction model construction, and is characterized in that an electric geographic information system is established, class I cells and class II cells are generated in the electric geographic information system, on the basis of positioning urban network space load by taking cells as minimum units, the space position of each class I cell and the propagation rule and time sequence characteristic of the load are combined to form a space weight matrix between the class I cell loads and a time weight matrix between the corresponding class I cell loads in a region to be planned respectively, the space-time correlation between the class I cell loads is measured through the space weight matrix and the time weight matrix, a space load prediction model comprising an abnormal data detection module adopting the grid local outlier factor algorithm and a space-time rule mining module adopting the space-time convolution algorithm is constructed, and finally the class I cell-based prediction result output by the model is converted into a class II cell-based prediction result by adopting a grid technology.
Example 3
The embodiment of the invention predicts the space power load of a administrative district 2014 by using the prior method and the method provided by the invention by utilizing the historical load data of a development district of a city in 2009-2013 and the power supply range thereof and the land information in the administrative district, and a specific prediction strategy is shown in figure 6.
1) Positioning urban network load based on class I cells
And generating I-type cells according to the power supply range of the 10kV feeder line in the area to be planned, and realizing the positioning of the urban network load space distribution in the area to be planned.
2) Measuring spatiotemporal correlation
And calculating time sequence correlation among the class I cells in the area to be planned, wherein the calculated result is shown in table 1.
TABLE 1 timing correlation metric results
As can be seen from table 1, there is a certain degree of timing correlation between each class I cell. Among them, the time correlation between the cells 21 and 22 is strongest because: the historic load curve change rules of the cells 21 and 22 are similar: both have a strong annual periodicity, with higher and lower annual load values at the beginning and end of the year.
The spatial correlation between the class I cells in the area to be planned is calculated, and the calculation result is shown in table 2.
As can be seen from table 2, only a part of the cells of each class I have a spatial correlation, i.e. the spatial weight matrix has a high sparsity. Among them, the spatial correlation between the cells 16 and 19, and between the cells 21 and 22 is strong because: cells 16 and 19, 21 and 22 are spatially adjacent geographic locations and have a longer spatial sharing boundary.
TABLE 2 spatial correlation measurement results
3) Spatial load prediction
And detecting abnormal values of the type I cell historical load data by adopting a grid LOF algorithm: taking h t ＝h p =50. Outlier detection is performed on points in the boundary grid using the LOF algorithm, k=100 is set, and points with local outlier factors greater than 1 are determined to be outlier points. Reasonable maximum values after the abnormal values are removed from each class I cell of the area to be planned are shown in table 3.
TABLE 3 reasonable maximum load for each class I cell
The comparison invention provides prediction accuracy of the prediction method and 8 existing spatial load prediction methods, and the 8 prediction methods are shown in table 4.
Table 48 names and numbers of spatial load prediction methods
The 8 spatial load prediction methods in table 4 include 3 "traditional prediction methods" (F5, F7, F8), 3 "data preprocessing+traditional prediction methods" combined models (F2, F3, F4), 2 "data preprocessing+artificial intelligence algorithm prediction methods" (F1, F6). The method relates to algorithms and theories such as a noise reduction self-encoder (DAE), singular Spectrum Analysis (SSA), a long-short-term neural network (LSTM), complementary set empirical mode decomposition (CEEMD), gray theory (GM) and the like.
In order to facilitate urban network planning and use and analysis of designers, the spatial load prediction result of the class I cells is finally required to be converted into the spatial load prediction result of the class II cells through a meshing technology, the prediction result of the class II cells of each method is shown as part (a) in fig. 7 to part (h) in fig. 7, part (a) in fig. 7 is the spatial load prediction result of the class II cells based on the class II cells, part (b) in fig. 7 is the spatial load prediction result of the class II cells based on the class F2 method, part (c) in fig. 7 is the spatial load prediction result of the class II cells based on the F3 method, part (d) in fig. 7 is the spatial load prediction result of the class II cells based on the class F4 method, part (e) in fig. 7 is the spatial load prediction result of the class II cells based on the part (F6) in fig. 7, part (g) in fig. 7 is the spatial load prediction result of the class II cells based on the spatial load prediction result of the class j in fig. 7.
In order to further quantify the prediction effect of the different prediction methods, the relative error RE, root mean square error RMSE and mean absolute error MAE of each prediction method based on the class II cell prediction result were calculated separately, and the calculation results are shown in table 5.
Table 5 prediction error for each method based on class II cells
As can be seen from table 5, the RMSE index was decreased by 0.0524MW, 0.0456MW, 0.0421MW, respectively, compared to the three conventional prediction methods F8, F7, F5; MAE index decreased by 0.0536MW, 0.0482MW, 0.0433MW, respectively. The relative errors are respectively reduced by 29.18 percent, 26.21 percent and 23.56 percent, the average improvement rate of the prediction accuracy exceeds 25 percent, and the prediction accuracy is greatly improved.
Compared with the combined models F4, F3 and F2 of three data pretreatment and traditional prediction methods, the RMSE indexes are respectively reduced by 0.0290MW, 0.0197MW and 0.0232MW; MAE index decreased by 0.0332MW, 0.0226MW, 0.0261MW, respectively. The relative errors are reduced by 17.49%, 12.29% and 14.22%, and the prediction accuracy is improved by 14.67% on average, so that compared with the traditional combined model, the combined prediction model combining the grid LOF algorithm and the space-time convolution algorithm adopted by the invention has the advantage that the prediction accuracy is obviously improved.
Compared with advanced artificial intelligence algorithm combined data preprocessing models F6 and F1, the RMSE indexes are respectively reduced by 0.0381MW and 0.0092MW; the MAE index is respectively reduced by 0.0405MW and 0.0119MW; the RE indexes are respectively reduced by 22.02 percent and 6.46 percent, and the change rule of the load is not comprehensive enough when the load is mined from the time sequence angle, and compared with the prediction method of the invention, the prediction precision of the combined data preprocessing model of the advanced artificial intelligent algorithm can be improved.
In this embodiment, first, positioning of urban network load spatial distribution is realized based on cells. And then, combining the spatial positions and the shared boundaries of the class I cells, providing a measurement method of the spatial load correlation of the class I cells according to a geographic rule, and measuring the time sequence correlation of the loads of all class I cells by adopting a pearson correlation coefficient. And finally, constructing a space load prediction model comprising a data outlier detection module adopting a grid local outlier factor algorithm and a space-time rule mining module adopting a space-time convolution algorithm. Finally, the effectiveness of the method provided by the invention is verified through calculation example analysis.
The invention also provides a space load prediction system which is corresponding to the method and takes the space-time correlation into consideration, comprising:
the power geographic information system processing module is used for processing a power geographic information system of an area to be planned to obtain a power geographic information system composed of a plurality of I-type cells and a power geographic information system composed of a plurality of II-type cells; the electric power geographic information system formed by a plurality of I-type cells is obtained by drawing a cell layer in the electric power geographic information system of the area to be planned; the power geographic information system formed by a plurality of class II cells is obtained by carrying out grid division on the power geographic information system of the area to be planned.
The geographic parameter determining module is used for obtaining geographic parameters between every two I-type cells in all I-type cells according to the electric power geographic information system consisting of a plurality of I-type cells; the geographic parameters include distance and shared boundary coefficients.
And the space weight matrix calculation module is used for obtaining the space weight matrix of the load of the area to be planned according to the geographic parameters between every two I-type cells in all I-type cells.
The time weight matrix calculation module is used for obtaining the time weight matrix of the load of the area to be planned according to the pearson correlation coefficient of the load time sequence of every two I-type cells in each I-type cell; the load time series includes load power at a plurality of historical moments.
The local outlier factor calculation module is used for obtaining a time-load power scatter diagram of any I-type cell according to the load time sequence of the I-type cell, and processing the time-load power scatter diagram of the I-type cell by adopting a grid local outlier factor detection algorithm to obtain local outlier factors of each point in the time-load power scatter diagram of the I-type cell.
And the reference load determining module is used for obtaining the reference load of the class I cell according to the local outlier factors of each point in the time-load power scatter diagram of the class I cell and the time-load power scatter diagram of the class I cell.
And the reference year load matrix determining module is used for determining the reference load of all the class I cells as the reference year load matrix of the area to be planned.
The space load prediction value calculation module is used for obtaining the space load prediction value of each type II cell in the area to be planned according to the reference year load matrix of the area to be planned, the space weight matrix of the load of the area to be planned, the time weight matrix of the load of the area to be planned and the power geographic information system consisting of a plurality of type II cells.
In the present specification, each embodiment is described in a progressive manner, and each embodiment is mainly described in a different point from other embodiments, and identical and similar parts between the embodiments are all enough to refer to each other. For the system disclosed in the embodiment, since it corresponds to the method disclosed in the embodiment, the description is relatively simple, and the relevant points refer to the description of the method section.
The principles and embodiments of the present invention have been described herein with reference to specific examples, the description of which is intended only to assist in understanding the methods of the present invention and the core ideas thereof; also, it is within the scope of the present invention to be modified by those of ordinary skill in the art in light of the present teachings. In view of the foregoing, this description should not be construed as limiting the invention.
Claims (8)
1. A spatial load prediction method considering spatio-temporal correlation, comprising:
processing the power geographic information system of the area to be planned to obtain a power geographic information system composed of a plurality of I-type cells and a power geographic information system composed of a plurality of II-type cells; the electric power geographic information system formed by a plurality of I-type cells is obtained by drawing a cell layer in the electric power geographic information system of the area to be planned; the power geographic information system formed by a plurality of class II cells is obtained by carrying out grid division on the power geographic information system of the area to be planned;
obtaining geographic parameters between every two I-type cells in all I-type cells according to the electric power geographic information system consisting of a plurality of I-type cells; the geographic parameters comprise distance and shared boundary coefficients;
obtaining a space weight matrix of the load of the area to be planned according to the geographic parameters between every two I-type cells in all I-type cells;
obtaining a time weight matrix of the load of the area to be planned according to the Pearson correlation coefficient of the load time sequence of every two I-type cells in each I-type cell; the load time sequence comprises load power at a plurality of historical moments;
For any I-type cell, obtaining a time-load power scatter diagram of the I-type cell according to the load time sequence of the I-type cell, and processing the time-load power scatter diagram of the I-type cell by adopting a grid local outlier factor detection algorithm to obtain local outlier factors of each point in the time-load power scatter diagram of the I-type cell;
obtaining the reference load of the class I cell according to the local outlier factors of each point in the time-load power scatter diagram of the class I cell and the time-load power scatter diagram of the class I cell;
determining the reference load of all the class I cells as a reference year load matrix of the area to be planned;
and obtaining the space load predicted value of each type II cell in the area to be planned according to the reference year load matrix of the area to be planned, the space weight matrix of the load of the area to be planned, the time weight matrix of the load of the area to be planned and a power geographic information system consisting of a plurality of type II cells.
2. The space load prediction method considering space-time correlation according to claim 1, wherein the obtaining the shared boundary coefficient between every two I-type cells in all I-type cells according to the power geographic information system composed of a plurality of I-type cells specifically comprises:
For the ith and jth class I cells, according to the formula
Calculating a shared boundary coefficient between the ith class I cell and the jth class I cell; wherein b ij For the shared boundary coefficient between the ith class I cell and the jth class I cell, l ij For the length of the shared boundary between the ith class I cell and the jth class I cell, l i,min For the minimum value of the shared boundary length in all class I cells adjacent to the I-th class I cell, l i,max Is the maximum value of the shared boundary length in all class I cells adjacent to the I-th class I cell.
3. The spatial load prediction method considering space-time correlation according to claim 1, wherein the spatial weight matrix of the area load to be planned is obtained according to the geographic parameters between every two class I cells in all class I cells, and specifically comprises:
obtaining the space weight between the loads of every two cells in all the class I cells according to the geographic parameters between every two class I cells in all the class I cells;
and determining a matrix formed by the space weights between every two cell loads in all the class I cells as a space weight matrix of the area load to be planned.
4. The spatial load prediction method considering space-time correlation according to claim 1, wherein the time weight matrix of the area load to be planned is obtained according to pearson correlation coefficient of load time series of every two class I cells in each class I cell, specifically comprising:
Calculating the pearson correlation coefficient of the load time sequences of every two class I cells in all class I cells according to the load time sequences of all class I cells to obtain the time correlation coefficient between the loads of every two cells;
and determining a matrix formed by time correlation coefficients between loads of every two cells in all the class I cells as a load time weight matrix of the area to be planned.
5. The spatial load prediction method considering space-time correlation according to claim 1, wherein the obtaining the reference load of the class I cell according to the local outlier factor of each point in the time-load power scatter diagram of the class I cell and the time-load power scatter diagram of the class I cell specifically comprises:
deleting points with the local outlier factor larger than a set threshold value from the time-load power scatter diagram of the class I cell to obtain an abnormal scatter diagram of the class I cell;
and determining the maximum value of the load power corresponding to all points in the abnormal scatter diagram of the class I cell removal as the reference load of the class I cell.
6. The space load prediction method considering space-time correlation according to claim 1, wherein the space load prediction value of each class II cell in the area to be planned is obtained according to the reference year load matrix of the area to be planned, the space weight matrix of the load of the area to be planned, the time weight matrix of the load of the area to be planned and a power geographic information system composed of a plurality of class II cells, and specifically comprises:
Obtaining a space load predicted value of each class I cell in the area to be planned according to the reference year load matrix of the area to be planned, the space weight matrix of the load of the area to be planned and the time weight matrix of the load of the area to be planned;
and obtaining the space load predicted value of each class II cell in the area to be planned according to the space load predicted value of each class I cell in the area to be planned and a power geographic information system consisting of a plurality of class II cells.
7. The spatial load prediction method considering space-time correlation according to claim 1, wherein the method comprises obtaining a time-load power scatter diagram of the class I cells according to the load time sequence of the class I cells, and processing the time-load power scatter diagram of the class I cells by using a grid local outlier factor detection algorithm to obtain local outlier factors of points in the time-load power scatter diagram of the class I cells, and specifically comprises:
obtaining a time-load power scatter diagram of the class I cell according to the load time sequence of the class I cell;
performing grid division on the time-load power scatter diagram of the cell to obtain a grid scatter diagram; each grid in the grid scatter plot comprises one or more points;
And calculating local outlier factors of points in all boundary grids in the grid scatter diagram by adopting an LOF algorithm.
8. A spatial load prediction system that considers spatio-temporal correlation, comprising:
the power geographic information system processing module is used for processing a power geographic information system of an area to be planned to obtain a power geographic information system composed of a plurality of I-type cells and a power geographic information system composed of a plurality of II-type cells; the electric power geographic information system formed by a plurality of I-type cells is obtained by drawing a cell layer in the electric power geographic information system of the area to be planned; the power geographic information system formed by a plurality of class II cells is obtained by carrying out grid division on the power geographic information system of the area to be planned;
the geographic parameter determining module is used for obtaining geographic parameters between every two I-type cells in all I-type cells according to the electric power geographic information system consisting of a plurality of I-type cells; the geographic parameters comprise distance and shared boundary coefficients;
the space weight matrix calculation module is used for obtaining a space weight matrix of the load of the area to be planned according to the geographic parameters between every two I-type cells in all I-type cells;
The time weight matrix calculation module is used for obtaining the time weight matrix of the load of the area to be planned according to the pearson correlation coefficient of the load time sequence of every two I-type cells in each I-type cell; the load time sequence comprises load power at a plurality of historical moments;
the local outlier factor calculation module is used for obtaining a time-load power scatter diagram of any I-type cell according to the load time sequence of the I-type cell, and processing the time-load power scatter diagram of the I-type cell by adopting a grid local outlier factor detection algorithm to obtain local outlier factors of each point in the time-load power scatter diagram of the I-type cell;
the reference load determining module is used for obtaining the reference load of the class I cell according to the local outlier factors of each point in the time-load power scatter diagram of the class I cell and the time-load power scatter diagram of the class I cell;
the reference year load matrix determining module is used for determining that the reference load of all the class I cells is the reference year load matrix of the area to be planned;
the space load prediction value calculation module is used for obtaining the space load prediction value of each type II cell in the area to be planned according to the reference year load matrix of the area to be planned, the space weight matrix of the load of the area to be planned, the time weight matrix of the load of the area to be planned and the power geographic information system consisting of a plurality of type II cells.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202310912496.9A CN116937563A (en) | 2023-07-25 | 2023-07-25 | Space load prediction method and system considering space-time correlation |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202310912496.9A CN116937563A (en) | 2023-07-25 | 2023-07-25 | Space load prediction method and system considering space-time correlation |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN116937563A true CN116937563A (en) | 2023-10-24 |
Family
ID=88391912
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202310912496.9A Pending CN116937563A (en) | 2023-07-25 | 2023-07-25 | Space load prediction method and system considering space-time correlation |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN116937563A (en) |
Cited By (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN117878927A (en) * | 2024-03-11 | 2024-04-12 | 国网黑龙江省电力有限公司绥化供电公司 | Power system load trend analysis method based on time sequence analysis |
-
2023
- 2023-07-25 CN CN202310912496.9A patent/CN116937563A/en active Pending
Cited By (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN117878927A (en) * | 2024-03-11 | 2024-04-12 | 国网黑龙江省电力有限公司绥化供电公司 | Power system load trend analysis method based on time sequence analysis |
| CN117878927B (en) * | 2024-03-11 | 2024-05-28 | 国网黑龙江省电力有限公司绥化供电公司 | Power system load trend analysis method based on time sequence analysis |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN108846517B (en) | Integration method for predicating quantile probabilistic short-term power load | |
| CN111563611B (en) | Cloud data center renewable energy space-time prediction method for graph rolling network | |
| CN114091766B (en) | CEEMDAN-LSTM-based space load prediction method | |
| CN116937563A (en) | Space load prediction method and system considering space-time correlation | |
| US20240014651A1 (en) | Probability estimation method for photovoltaic power based on optimized copula function and photovoltaic power system | |
| CN112054943A (en) | Traffic prediction method for mobile network base station | |
| CN113112090B (en) | Space load prediction method based on principal component analysis of comprehensive mutual informativity | |
| CN112329973A (en) | Space-time load prediction method based on graph neural network and regional gridding | |
| CN116404647B (en) | Distributed photovoltaic short-term power prediction method and system considering dynamic convergence characteristics | |
| CN112508279B (en) | Regional distributed photovoltaic prediction method and system based on spatial correlation | |
| CN112329997A (en) | Power demand load prediction method and system, electronic device, and storage medium | |
| CN114357670A (en) | Power distribution network power consumption data abnormity early warning method based on BLS and self-encoder | |
| CN108717581A (en) | A kind of random multiple attributive decision making method of reservoir operation based on Monte Carlo simulation | |
| CN115600729A (en) | Grid load prediction method considering multiple attributes | |
| Lai et al. | Sub-region division based short-term regional distributed PV power forecasting method considering spatio-temporal correlations | |
| CN114169251A (en) | Ultra-short-term wind power prediction method | |
| CN111882119A (en) | Battery SOH prediction optimization method based on SA-BP neural network | |
| Wang et al. | Electric vehicle charging load clustering and load forecasting based on long short term memory neural network | |
| CN112101673B (en) | Power grid development trend prediction method and system based on hidden Markov model | |
| CN113868812A (en) | Medium voltage distribution network cable path planning method based on intelligent ant colony algorithm | |
| CN116316589A (en) | Distribution network distributed photovoltaic bearing capacity assessment method considering source load uncertainty | |
| Shen et al. | An interval analysis scheme based on empirical error and MCMC to quantify uncertainty of wind speed | |
| CN111062538A (en) | CEEMD method for determining reasonable maximum value of cellular load in space load prediction | |
| CN113592565B (en) | Method for predicting importance degree of nodes of power system | |
| Yan et al. | Water consumption prediction model based on clustering and Multi-layer Perceptron |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination |