CN110717619A - Multi-scale space-time load prediction method and system for bottom-up power distribution network - Google Patents

Multi-scale space-time load prediction method and system for bottom-up power distribution network Download PDF

Info

Publication number
CN110717619A
CN110717619A CN201910858169.3A CN201910858169A CN110717619A CN 110717619 A CN110717619 A CN 110717619A CN 201910858169 A CN201910858169 A CN 201910858169A CN 110717619 A CN110717619 A CN 110717619A
Authority
CN
China
Prior art keywords
load
typical
density
clustering
index
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
Application number
CN201910858169.3A
Other languages
Chinese (zh)
Inventor
张曼颖
王蕾
郑伟民
叶承晋
丁一
Current Assignee (The listed assignees may be inaccurate. Google has not performed a legal analysis and makes no representation or warranty as to the accuracy of the list.)
Zhejiang University ZJU
Economic and Technological Research Institute of State Grid Zhejiang Electric Power Co Ltd
Original Assignee
Zhejiang University ZJU
Economic and Technological Research Institute of State Grid Zhejiang Electric Power Co Ltd
Priority date (The priority date 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 date listed.)
Filing date
Publication date
Application filed by Zhejiang University ZJU, Economic and Technological Research Institute of State Grid Zhejiang Electric Power Co Ltd filed Critical Zhejiang University ZJU
Priority to CN201910858169.3A priority Critical patent/CN110717619A/en
Publication of CN110717619A publication Critical patent/CN110717619A/en
Pending legal-status Critical Current

Links

Images

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
    • G06Q10/04—Forecasting or optimisation specially adapted for administrative or management purposes, e.g. linear programming or "cutting stock problem"
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06F—ELECTRIC DIGITAL DATA PROCESSING
    • G06F18/00—Pattern recognition
    • G06F18/20—Analysing
    • G06F18/23—Clustering techniques
    • G06F18/232—Non-hierarchical techniques
    • G06F18/2321—Non-hierarchical techniques using statistics or function optimisation, e.g. modelling of probability density functions
    • G06F18/23213—Non-hierarchical techniques using statistics or function optimisation, e.g. modelling of probability density functions with fixed number of clusters, e.g. K-means clustering
    • 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
    • G06Q10/0639—Performance analysis of employees; Performance analysis of enterprise or organisation operations
    • G06Q10/06393—Score-carding, benchmarking or key performance indicator [KPI] analysis
    • 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

Landscapes

  • Engineering & Computer Science (AREA)
  • Business, Economics & Management (AREA)
  • Human Resources & Organizations (AREA)
  • Economics (AREA)
  • Theoretical Computer Science (AREA)
  • Strategic Management (AREA)
  • Physics & Mathematics (AREA)
  • General Physics & Mathematics (AREA)
  • General Business, Economics & Management (AREA)
  • Marketing (AREA)
  • Tourism & Hospitality (AREA)
  • Entrepreneurship & Innovation (AREA)
  • Development Economics (AREA)
  • Data Mining & Analysis (AREA)
  • Operations Research (AREA)
  • Quality & Reliability (AREA)
  • Game Theory and Decision Science (AREA)
  • Educational Administration (AREA)
  • Health & Medical Sciences (AREA)
  • Water Supply & Treatment (AREA)
  • General Health & Medical Sciences (AREA)
  • Primary Health Care (AREA)
  • Probability & Statistics with Applications (AREA)
  • Public Health (AREA)
  • Life Sciences & Earth Sciences (AREA)
  • Artificial Intelligence (AREA)
  • Bioinformatics & Cheminformatics (AREA)
  • Bioinformatics & Computational Biology (AREA)
  • Computer Vision & Pattern Recognition (AREA)
  • Evolutionary Biology (AREA)
  • Evolutionary Computation (AREA)
  • General Engineering & Computer Science (AREA)
  • Supply And Distribution Of Alternating Current (AREA)

Abstract

The invention discloses a multi-scale space-time load prediction method and system for a bottom-up power distribution network. The technical scheme adopted by the invention is as follows: obtaining typical load density upper and lower bounds and high, middle and low section typical values of various types of landmass by a non-parameter kernel density estimation method; extracting various user typical load curves from the electricity utilization information acquisition system by adopting a self-adaptive k-means clustering method based on DB indexes, and accurately representing the electricity utilization characteristics of the users; a bottom-to-top ground load superposition method is adopted to obtain a load increase space-time panorama of a land block-grid-power supply area step by step, and selection of simultaneous rate is avoided. The method avoids the difficulty of simultaneous rate selection, improves the accuracy of the space load prediction result, and is favorable for scientifically guiding the site selection and the volume fixing of the transformer substation, the outgoing line arrangement and the user access.

Description

Multi-scale space-time load prediction method and system for bottom-up power distribution network
Technical Field
The invention belongs to the field of power distribution network space load prediction under a big data background, and particularly relates to a multi-dimensional data-driven multi-scale space-time load prediction method and system for a bottom-up power distribution network.
Background
The basis of medium and low voltage distribution network planning is space load prediction, and the accuracy of a space load prediction result has a critical influence on the applicability of a distribution network planning scheme. The load density index space load prediction method which is most widely applied at present has low applicability, and the method faces two difficulties in practical application.
Difficulty 1 is the determination of the load density indicator. In actual power distribution network planning, the basis is usually urban power planning specifications (GB/T50293-2014), but due to differences in climate, economic development level, industry types, and the like in various regions, the applicability of a unified load density index system is limited, a spatial load density index needs to be selected in a targeted manner according to local characteristics, and an instructive method for determining a load density differential index is currently lacking.
Difficulty 2 is the selection of coincidence. Different loads have different peak-valley time distributions, the maximum load capacity cannot be simply considered when a load density index method is used for superposing the landmasses with different properties, and a certain proportion is required to be considered, namely the coincidence rate. Obviously, different simultaneous rates exist among different types of land, but the existing planning standard does not provide a simultaneous rate selection standard, and a planner can only set the simultaneous rate selection standard through experience in most cases, so that the subjectivity is high.
The drawbacks of the prior art are summarized as follows: the load density index does not consider the difference of each place in the aspects of development level, climate and the like, and the coincidence rate cannot be scientifically selected when the plots are overlapped, so that the error of a prediction result is large.
Disclosure of Invention
Aiming at the defects in the prior art, the invention provides a multi-dimensional data-driven multi-scale space-time load prediction method and system for a bottom-up power distribution network.
Therefore, the technical scheme adopted by the invention is as follows: a multi-scale space-time load prediction method for a bottom-up power distribution network comprises the following steps:
1) typical values of the upper and lower bounds and the high, middle and low sections of the typical load density of each type of land parcel are obtained by a non-parameter kernel density estimation method, so that the scientificity of the load density index is improved;
2) extracting various user typical load curves from the electricity utilization information acquisition system by adopting a self-adaptive k-means clustering method based on DB indexes, and accurately representing the electricity utilization characteristics of the users;
3) a bottom-to-top ground load superposition method is adopted to obtain a load increase space-time panorama of a land block-grid-power supply area step by step, and selection of simultaneous rate is avoided.
The method is based on load density data and a typical load curve of a large number of sample plots, has the advantage of large data robustness, and the multi-scale refers to space scales with different sizes such as plot-grid-power supply area.
Further, the predicted area is an area where a detailed planning of control has been carried out, and the land use properties are known.
Further, a typical index of the space load density is counted by a non-parameter kernel density estimation method, and the specific content is as follows:
establishing a space load density probability model based on statistical data of a large number of mature plots, and setting x1,x2,…,xnIs the load density value of n samples of a certain geographic type, the probability density function of the load density is f (x), then the kernel estimate of the probability density function is:
Figure BDA0002198866580000021
in the formula, n is the sample capacity; h is the bandwidth, serving as the smoothing factor; k () is a kernel function; selecting a standard Gaussian function as a kernel function, and setting the bandwidth to be 0.5, then the kernel density estimation of the load density is as follows:
Figure BDA0002198866580000022
after obtaining the kernel density estimation curve of the load density value, in order to remove the influence of the extreme data in the sample, respectively removing 5% from the minimum and maximum load density data, then equally dividing the residual range, and selecting the extreme from each section as the typical value of the load density of the section.
The pole indicates that the section of the load density is most concentrated in the distribution around the value, and thus the load density is most representative. The space load density distribution rule of the land type can be relatively comprehensively described by matching a plurality of typical values.
Further, the DB index is adopted to measure the load curve clustering effect, and the specific content is as follows:
acquiring a large number of user daily load curves of the land types through the electricity utilization information acquisition system, clustering the daily load curves, taking the center line of each type as a typical load curve of the land type, and determining the appropriate clustering type number by adopting a DB index:
Figure BDA0002198866580000031
Figure BDA0002198866580000032
in the formula, CiDenotes the ith type, zjIs CiVector of (1), AiCluster center ofiP represents the number of zjAnd AiThe DB index is finally calculated according to the formula.
Is apparent SiMeasure CiDegree of correspondence within a type; | | Ai-Aj||pMeasure type CiAnd CjThe degree of difference in (c). Therefore, the DB index simultaneously considers the consistency of the same type and the difference of the different types and is the comprehensive measurement of the clustering effectiveness. The purpose of clustering the load curves is to search for the smallest DB value, so as to achieve the best clustering effect, and this search may be performed by an iterative algorithm.
Further, various typical load curves are extracted by adopting a self-adaptive clustering algorithm based on DB indexes, and the specific content is as follows:
the optimal clustering number of the k-means algorithm is searched by iteration based on DB indexes, and the process comprises 9 steps:
step 1: setting k to be 2, and inputting a load curve set to be clustered;
step 2: running a k-means algorithm;
and 3, step 3: calculating the DB index size of the current clustering scheme, and recording the DB index size as DB (k);
and 4, step 4: updating k to k + 1;
and 5, step 5: judging whether the maximum allowable clustering number is reached, namely whether k is more than kmaxIf yes, jumping to the step 9, and if not, entering the step 6;
and 6, step 6: running a k-means algorithm;
and 7, step 7: calculating the DB index size of the current clustering scheme, and recording the DB index size as DB (k);
and 8, step 8: judging whether the clustering effect is better or not based on the DB index, namely whether DB (k) is less than DB (k-1), if so, returning to the step 4, and if not, entering the step 9;
step 9: and terminating the iteration and outputting the optimal clustering number k.
According to the invention, reference information is provided for the superposition of the plot peak load through the clustering load curve, so that the load curve in the sample library is selected in the month of the regional maximum load. Consider that big data by itself has good fault tolerance when there are enough sample curves. Therefore, it is not necessary to specify a specific date of load curve selection.
Further, the multi-scale space-time load prediction method for the bottom-up power distribution network comprises the following specific contents:
step 1: based on the controlled detailed planning of the area to be predicted and the river and the traffic main road as boundaries, dividing plots, grids and power supply areas of the area to be predicted;
step 2: based on a power consumption information acquisition system and a geographic information system, searching through the Internet, collecting various mature user space load density databases in the region where the block to be predicted is located, and obtaining typical load density indexes of various types of blocks through non-parameter kernel density estimation;
and 3, step 3: extracting a large number of typical daily load curves of representative users of various types of plots based on a power utilization information acquisition system, and obtaining the typical load curves of the typical plots through a K-means clustering algorithm based on DB indexes;
and 4, step 4: classifying and matching the types of the plots to be predicted to obtain load density and a typical load curve corresponding to each plot, and calculating the maximum load of each plot by combining a load density index method and an area;
and 5, step 5: based on typical load curves of each plot, a load increase space-time panorama of a plot-grid-power supply area is obtained by a bottom-up superposition method.
Further, the maximum load value of the area to be predicted is calculated by the following formula:
in the formula, max is the symbol of maximum value, n is the number of functional blocks of the block to be predicted, and the typical daily load curve of each block comprises m-dimensional load points, wherein P is the number of functional blocksijIs the value of the jth load point, P, of the plot iiIs the maximum load of plot i, calculated by the following load density index method:
Pi＝Siρi，
in the formula, SiAnd ρiTypical values for the area and load density of the plot i, respectively.
Further, the adopted k-means clustering algorithm is based on load curve distance measure, and the following procedures are adopted:
step 1: optionally selecting k initial clustering centers Z in the load curve data set to be clustered1(l),Z2(l)…Zk(l)；
Step 2: calculating the distance from each load curve sample to k clustering center curves, and classifying according to a nearest rule; if | | | X-Zj(l)||＜||X-Zi(l) If X belongs to Gj(l) 1,2, … k, i ≠ j, wherein: gj(l) As a cluster center Zj(l) Clustering the samples; distributing all samples from X to k clustering centers in the first iteration;
and 3, step 3: calculating a new clustering center from the calculation result of the step 2;
Figure BDA0002198866580000052
wherein j is 1,2, …, k;
new clustering center should be J of criterion functionjThe value reaches the minimum:
Figure BDA0002198866580000053
wherein: gjIs the jth load curve class; n is a radical ofjThe number of load curve samples of the jth class; zjClustering the center of the load curve of the jth sample;
and 4, step 4: if the new cluster center is equal to the previous cluster center, then:
Zj(l+1)＝Zj(l),j＝1,2,…,k；
the algorithm is converged and clustering is finished; otherwise, go to step 2.
Further, the daily load curves are all subjected to normalization processing in advance, and the specific contents are as follows:
assuming that the typical daily load curve is an m-dimensional array, the ith load curve is represented as Xi＝(xi,1,xi,2,…,xi,m) The normalization result of the j dimension load value is as follows:
Figure BDA0002198866580000061
in the formula, min and max are the operation signs of taking the minimum value and the maximum value respectively.
The other technical scheme adopted by the invention is as follows: a multi-scale space-time load prediction system for a bottom-up power distribution network, comprising:
a region to be predicted dividing unit: based on the controlled detailed planning of the area to be predicted and the river and the traffic main road as boundaries, dividing plots, grids and power supply areas of the area to be predicted;
a typical load density index acquisition unit: based on a power consumption information acquisition system and a geographic information system, searching through the Internet, collecting various mature user space load density databases in the region where the block to be predicted is located, and obtaining typical load density indexes of various types of blocks through non-parameter kernel density estimation;
typical load curve calculation unit: extracting a large number of typical daily load curves of representative users of various types of plots based on a power utilization information acquisition system, and obtaining the typical load curves of the typical plots through a K-means clustering algorithm based on DB indexes;
a maximum load calculation unit: classifying and matching the types of the plots to be predicted to obtain load density and a typical load curve corresponding to each plot, and calculating the maximum load of each plot by combining a load density index method and an area;
load growth spatiotemporal panorama acquisition unit: based on typical load curves of each plot, a load increase space-time panorama of a plot-grid-power supply area is obtained by a bottom-up superposition method.
The load density index given by the sum density estimation method comprises an upper boundary and a lower boundary as well as a high typical value, a middle typical value and a low typical value, and has clear physical significance, wherein the upper boundary and the lower boundary of the density give a possible range of the load density of the area, and can be used for judging and verifying a load prediction result; and the three distribution characteristics of the low section density typical value, the middle section density typical value and the high section density typical value are the most possible load density values obtained by a large number of statistics and can be directly used for the planning load standard of the power distribution network.
The invention is based on a power distribution network gridding system, and the space scale of a land block-grid-power supply area is a common unit for planning and managing the power distribution network in China at present. The plot is a basic unit of planning intensity assignment in a gridding planning system, the range corresponds to a functional plot in land utilization planning and controllability detailed planning, generally, the plot corresponds to a distribution transformer in the load magnitude, plot load prediction is carried out, and the method is beneficial to determining distribution transformer capacity and the number of distribution transformers, determining switch station distribution points and formulating a user access scheme. The power grid corresponds to the medium-voltage line level in the load magnitude and corresponds to the medium-voltage grid structure, power grid load prediction is carried out, and the power grid power. The power supply area generally reaches a main power supply point of a high-voltage distribution network on the load capacity, power balance of a transformer substation is corresponded, load prediction of the power supply area is conducted, total requirements of the transformer substation in a planning subarea, including capacity and quantity of the transformer substation, the range of the power supply area of the 110kV transformer substation is determined, and the site of the transformer substation is optimized.
The method is used for predicting the space load from bottom to top under the 'plot-grid-power supply area' system, and is helpful for improving the connectivity of power distribution network planning and municipal planning.
The invention has the beneficial effects that:
the invention provides a load density index system reflecting regional characteristics by applying a kernel density estimation method, and adopts a bottom-up load superposition method based on a load curve, thereby avoiding the difficulty of simultaneous rate selection, improving the precision of a space load prediction result and improving the applicability of a power distribution network planning scheme.
Drawings
FIG. 1 is a flow chart of example 2 of the present invention;
FIG. 2 is a schematic diagram of a typical load density value extraction process based on non-parametric kernel density estimation in an application example of the present invention;
FIG. 3 is a flow chart of a load curve adaptive clustering algorithm based on DB indexes in an application example of the present invention;
FIG. 4 is a diagram illustrating the effect of extracting a typical load curve in a region according to an embodiment of the present invention;
FIG. 5 is a graph of predicted load curves obtained by applying the method of the present invention to regions from bottom to top.
Detailed Description
The invention is further described with reference to the drawings and examples.
Example 1
The embodiment provides a multi-dimensional data-driven multi-scale space-time load prediction method for a bottom-up power distribution network, which comprises the following steps:
1) firstly, typical load density upper and lower bounds and typical values of high, middle and low sections of various types of landmass are obtained by a nonparametric kernel density estimation method, and the scientificity of a load density index is improved;
2) then, providing a self-adaptive k-means clustering method based on DB indexes, extracting various user typical load curves from the electricity utilization information acquisition system, and accurately representing the electricity utilization characteristics of the users;
3) finally, a bottom-up ground load superposition method is provided, and a space-time panorama of 'plot-grid-power supply area' for load increase is obtained step by step, so that selection of a simultaneous rate is avoided.
The predicted area is an area where a controlled detailed planning has been carried out, and the land use properties are known.
The typical indexes of the space load density are counted by a nonparametric kernel density estimation method, and the specific contents are as follows:
establishing a space load density probability model based on statistical data of a large number of mature plots, and setting x1,x2,…,xnIs the load density value of n samples of a certain geographic type, the probability density function of the load density is f (x), then the kernel estimate of the probability density function is:
in the formula, n is the sample capacity; h is the bandwidth, serving as the smoothing factor; k () is a kernel function; selecting a standard Gaussian function as a kernel function, and setting the bandwidth to be 0.5, then the kernel density estimation of the load density is as follows:
after obtaining the kernel density estimation curve of the load density value, in order to remove the influence of the extreme data in the sample, respectively removing 5% from the minimum and maximum load density data, then equally dividing the residual range, and selecting the extreme from each section as the typical value of the load density of the section.
Measuring the clustering effect of the load curve by using DB indexes, wherein the specific contents are as follows:
acquiring a large number of user daily load curves of the land types through the electricity utilization information acquisition system, clustering the daily load curves, taking the center line of each type as a typical load curve of the land type, and determining the appropriate clustering type number by adopting a DB index:
Figure BDA0002198866580000091
in the formula, CiDenotes the ith type, zjIs CiVector of (1), AiCluster center ofiP represents the number of zjAnd AiThe DB index is finally calculated according to the formula.
Extracting various typical load curves by adopting a self-adaptive clustering algorithm based on DB indexes, wherein the specific contents are as follows:
the optimal clustering number of the k-means algorithm is searched by iteration based on DB indexes, and the process comprises 9 steps:
step 1: setting k to be 2, and inputting a load curve set to be clustered;
step 2: running a k-means algorithm;
and 3, step 3: calculating the DB index size of the current clustering scheme, and recording the DB index size as DB (k);
and 4, step 4: updating k to k + 1;
and 5, step 5: judging whether the maximum allowable clustering number is reached, namely whether k is more than kmaxIf yes, jumping to the step 9, and if not, entering the step 6;
and 6, step 6: running a k-means algorithm;
and 7, step 7: calculating the DB index size of the current clustering scheme, and recording the DB index size as DB (k);
and 8, step 8: judging whether the clustering effect is better or not based on the DB index, namely whether DB (k) is less than DB (k-1), if so, returning to the step 4, and if not, entering the step 9;
step 9: and terminating the iteration and outputting the optimal clustering number k.
The adopted k-means clustering algorithm is based on load curve distance measurement and adopts the following procedures:
step 1: optionally selecting k initial clustering centers Z in the load curve data set to be clustered1(l),Z2(l)…Zk(l)；
Step 2: calculating the distance from each load curve sample to k clustering center curves, and classifying according to a nearest rule; if | | | X-Zj(l)||＜||X-Zi(l) If X belongs to Gj(l) 1,2, … k, i ≠ j, wherein: gj(l) As a cluster center Zj(l) Clustering the samples; distributing all samples from X to k clustering centers in the first iteration;
and 3, step 3: calculating a new clustering center from the calculation result of the step 2;
Figure BDA0002198866580000101
wherein j is 1,2, …, k;
new clustering center should be J of criterion functionjThe value reaches the minimum:
Figure BDA0002198866580000102
wherein: gjIs the jth load curve class; n is a radical ofjLoad profile for the jth classThe number of the books; zjClustering the center of the load curve of the jth sample;
and 4, step 4: if the new cluster center is equal to the previous cluster center, then:
Zj(l+1)＝Zj(l),j＝1,2,…,k；
the algorithm is converged and clustering is finished; otherwise, go to step 2.
Example 2
The embodiment provides a multi-dimensional data-driven multi-scale space-time load prediction method for a bottom-up power distribution network, as shown in fig. 1, the method includes the following steps:
step 1: based on the controlled detailed planning of the area to be predicted and the river and the traffic main road as boundaries, dividing plots, grids and power supply areas of the area to be predicted;
step 2: based on a power consumption information acquisition system and a geographic information system, searching through the Internet, collecting various mature user space load density databases in the region where the block to be predicted is located, and obtaining typical load density indexes of various types of blocks through non-parameter kernel density estimation;
and 3, step 3: extracting a large number of typical daily load curves of representative users of various types of plots based on a power utilization information acquisition system, and obtaining the typical load curves of the typical plots through a K-means clustering algorithm based on DB indexes;
and 4, step 4: classifying and matching the types of the plots to be predicted to obtain load density and a typical load curve corresponding to each plot, and calculating the maximum load of each plot by combining a load density index method and an area;
and 5, step 5: based on typical load curves of each plot, a load increase space-time panorama of a plot-grid-power supply area is obtained by a bottom-up superposition method.
The predicted area is an area where a controlled detailed planning has been carried out, and the land use properties are known.
The typical indexes of the space load density are counted by a nonparametric kernel density estimation method, and the specific contents are as follows:
building a space based on statistical data of a large number of mature plotsProbability model of load density, let x1,x2,…,xnIs the load density value of n samples of a certain geographic type, the probability density function of the load density is f (x), then the kernel estimate of the probability density function is:
in the formula, n is the sample capacity; h is the bandwidth, serving as the smoothing factor; k () is a kernel function; selecting a standard Gaussian function as a kernel function, and setting the bandwidth to be 0.5, then the kernel density estimation of the load density is as follows:
Figure BDA0002198866580000112
after obtaining the kernel density estimation curve of the load density value, in order to remove the influence of the extreme data in the sample, respectively removing 5% from the minimum and maximum load density data, then equally dividing the residual range, and selecting the extreme from each section as the typical value of the load density of the section.
Measuring the clustering effect of the load curve by using DB indexes, wherein the specific contents are as follows:
acquiring a large number of user daily load curves of the land types through the electricity utilization information acquisition system, clustering the daily load curves, taking the center line of each type as a typical load curve of the land type, and determining the appropriate clustering type number by adopting a DB index:
Figure BDA0002198866580000113
Figure BDA0002198866580000114
in the formula, CiDenotes the ith type, zjIs CiVector of (1), AiCluster center ofiP represents the number of zjAnd AiThe DB index is finally calculated according to the formula.
Extracting various typical load curves by adopting a self-adaptive clustering algorithm based on DB indexes, wherein the specific contents are as follows:
the optimal clustering number of the k-means algorithm is searched by iteration based on DB indexes, and the process comprises 9 steps:
step 1: setting k to be 2, and inputting a load curve set to be clustered;
step 2: running a k-means algorithm;
and 3, step 3: calculating the DB index size of the current clustering scheme, and recording the DB index size as DB (k);
and 4, step 4: updating k to k + 1;
and 5, step 5: judging whether the maximum allowable clustering number is reached, namely whether k is more than kmaxIf yes, jumping to the step 9, and if not, entering the step 6;
and 6, step 6: running a k-means algorithm;
and 7, step 7: calculating the DB index size of the current clustering scheme, and recording the DB index size as DB (k);
and 8, step 8: judging whether the clustering effect is better or not based on the DB index, namely whether DB (k) is less than DB (k-1), if so, returning to the step 4, and if not, entering the step 9;
step 9: and terminating the iteration and outputting the optimal clustering number k.
The maximum load value of the area to be predicted is calculated by the following formula:
in the formula, max is the symbol of maximum value, n is the number of functional blocks of the block to be predicted, and the typical daily load curve of each block comprises m-dimensional load points, wherein P is the number of functional blocksijIs the value of the jth load point, P, of the plot iiIs the maximum load of plot i, calculated by the following load density index method:
Pi＝Siρi，
in the formula, SiAnd ρiTypical values for the area and load density of the plot i, respectively.
The adopted k-means clustering algorithm is based on load curve distance measurement and adopts the following procedures:
step 1: optionally selecting k initial clustering centers Z in the load curve data set to be clustered1(l),Z2(l)…Zk(l)；
Step 2: calculating the distance from each load curve sample to k clustering center curves, and classifying according to a nearest rule; if | | | X-Zj(l)||＜||X-Zi(l) If X belongs to Gj(l) 1,2, … k, i ≠ j, wherein: gj(l) As a cluster center Zj(l) Clustering the samples; distributing all samples from X to k clustering centers in the first iteration;
and 3, step 3: calculating a new clustering center from the calculation result of the step 2;
Figure BDA0002198866580000131
wherein j is 1,2, …, k;
new clustering center should be J of criterion functionjThe value reaches the minimum:
Figure BDA0002198866580000132
wherein: gjIs the jth load curve class; n is a radical ofjThe number of load curve samples of the jth class; zjClustering the center of the load curve of the jth sample;
and 4, step 4: if the new cluster center is equal to the previous cluster center, then:
Zj(l+1)＝Zj(l),j＝1,2,…,k；
the algorithm is converged and clustering is finished; otherwise, go to step 2.
The daily load curves are all subjected to normalization processing in advance, and the specific contents are as follows:
assuming that the typical daily load curve is an m-dimensional array, the ith load curve is represented as Xi＝(xi,1,xi,2,…,xi,m) The normalization result of the j dimension load value is as follows:
in the formula, min and max are the operation signs of taking the minimum value and the maximum value respectively.
Example 3
The embodiment provides a multiscale space-time load prediction system of a bottom-up power distribution network, which comprises:
a region to be predicted dividing unit: based on the controlled detailed planning of the area to be predicted and the river and the traffic main road as boundaries, dividing plots, grids and power supply areas of the area to be predicted;
a typical load density index acquisition unit: based on a power consumption information acquisition system and a geographic information system, searching through the Internet, collecting various mature user space load density databases in the region where the block to be predicted is located, and obtaining typical load density indexes of various types of blocks through non-parameter kernel density estimation;
typical load curve calculation unit: extracting a large number of typical daily load curves of representative users of various types of plots based on a power utilization information acquisition system, and obtaining the typical load curves of the typical plots through a K-means clustering algorithm based on DB indexes;
a maximum load calculation unit: classifying and matching the types of the plots to be predicted to obtain load density and a typical load curve corresponding to each plot, and calculating the maximum load of each plot by combining a load density index method and an area;
load growth spatiotemporal panorama acquisition unit: based on typical load curves of each plot, a load increase space-time panorama of a plot-grid-power supply area is obtained by a bottom-up superposition method.
The predicted area is an area where a controlled detailed planning has been carried out, and the land use properties are known.
The typical indexes of the space load density are counted by a nonparametric kernel density estimation method, and the specific contents are as follows:
establishing a space load density probability model based on statistical data of a large number of mature plots, and setting x1,x2,…,xnIs n samples of a certain plot typeThe load density value and the probability density function of the load density are f (x), and the kernel estimation of the probability density function is as follows:
Figure BDA0002198866580000141
in the formula, n is the sample capacity; h is the bandwidth, serving as the smoothing factor; k () is a kernel function; selecting a standard Gaussian function as a kernel function, and setting the bandwidth to be 0.5, then the kernel density estimation of the load density is as follows:
Figure BDA0002198866580000142
after obtaining the kernel density estimation curve of the load density value, in order to remove the influence of the extreme data in the sample, respectively removing 5% from the minimum and maximum load density data, then equally dividing the residual range, and selecting the extreme from each section as the typical value of the load density of the section.
Measuring the clustering effect of the load curve by using DB indexes, wherein the specific contents are as follows:
acquiring a large number of user daily load curves of the land types through the electricity utilization information acquisition system, clustering the daily load curves, taking the center line of each type as a typical load curve of the land type, and determining the appropriate clustering type number by adopting a DB index:
Figure BDA0002198866580000151
in the formula, CiDenotes the ith type, zjIs CiVector of (1), AiCluster center ofiP represents the number of zjAnd AiThe DB index is finally calculated according to the formula.
Extracting various typical load curves by adopting a self-adaptive clustering algorithm based on DB indexes, wherein the specific contents are as follows:
the optimal clustering number of the k-means algorithm is searched by iteration based on DB indexes, and the process comprises 9 steps:
step 1: setting k to be 2, and inputting a load curve set to be clustered;
step 2: running a k-means algorithm;
and 3, step 3: calculating the DB index size of the current clustering scheme, and recording the DB index size as DB (k);
and 4, step 4: updating k to k + 1;
and 5, step 5: judging whether the maximum allowable clustering number is reached, namely whether k is more than kmaxIf yes, jumping to the step 9, and if not, entering the step 6;
and 6, step 6: running a k-means algorithm;
and 7, step 7: calculating the DB index size of the current clustering scheme, and recording the DB index size as DB (k);
and 8, step 8: judging whether the clustering effect is better or not based on the DB index, namely whether DB (k) is less than DB (k-1), if so, returning to the step 4, and if not, entering the step 9;
step 9: and terminating the iteration and outputting the optimal clustering number k.
The maximum load value of the area to be predicted is calculated by the following formula:
Figure BDA0002198866580000153
in the formula, max is the symbol of maximum value, n is the number of functional blocks of the block to be predicted, and the typical daily load curve of each block comprises m-dimensional load points, wherein P is the number of functional blocksijIs the value of the jth load point, P, of the plot iiIs the maximum load of plot i, calculated by the following load density index method:
Pi＝Siρi，
in the formula, SiAnd ρiTypical values for the area and load density of the plot i, respectively.
The adopted k-means clustering algorithm is based on load curve distance measurement and adopts the following procedures:
step 1: in the load curve data set to be clusteredOptionally k initial clustering centers Z1(l),Z2(l)…Zk(l)；
Step 2: calculating the distance from each load curve sample to k clustering center curves, and classifying according to a nearest rule; if | | | X-Zj(l)||＜||X-Zi(l) If X belongs to Gj(l) 1,2, … k, i ≠ j, wherein: gj(l) As a cluster center Zj(l) Clustering the samples; distributing all samples from X to k clustering centers in the first iteration;
and 3, step 3: calculating a new clustering center from the calculation result of the step 2;
Figure BDA0002198866580000161
wherein j is 1,2, …, k;
new clustering center should be J of criterion functionjThe value reaches the minimum:
Figure BDA0002198866580000162
wherein: gjIs the jth load curve class; n is a radical ofjThe number of load curve samples of the jth class; zjClustering the center of the load curve of the jth sample;
and 4, step 4: if the new cluster center is equal to the previous cluster center, then:
Zj(l+1)＝Zj(l),j＝1,2,…,k；
the algorithm is converged and clustering is finished; otherwise, go to step 2.
The daily load curves are all subjected to normalization processing in advance, and the specific contents are as follows:
assuming that the typical daily load curve is an m-dimensional array, the ith load curve is represented as Xi＝(xi,1,xi,2,…,xi,m) The normalization result of the j dimension load value is as follows:
Figure BDA0002198866580000171
in the formula, min and max are the operation signs of taking the minimum value and the maximum value respectively.
Application example
According to the specific implementation scheme of the embodiment 2 of the invention, the block with the detailed control plan is taken as an object, and the method verification is carried out by taking the warp knitting industrial park of the Haining Zhejiang as an example, and the implementation process is as follows:
1) first, typical load density upper and lower bounds and typical high, middle and low section values of each type of parcel are obtained by the non-parametric kernel density estimation method shown in fig. 2, and the results are shown in table 1.
TABLE 1 extraction results of typical distribution characteristics of earth load density of various types of tannin
Figure BDA0002198866580000172
2) Then, extracting various user typical load curves from the electricity utilization information acquisition system by adopting a self-adaptive k-means clustering method based on DB indexes shown in FIG. 3, and accurately representing the electricity utilization characteristics of the users, wherein the result is shown in FIG. 4;
after the load density typical index and the typical load curve of the plot of the area to be predicted are obtained, a bottom-up superposition method based on the load curve is adopted, the load value of the whole prediction park is obtained to be 161.89MW, and the load curve of the park obtained by superposition is shown in figure 5.
Finally, it is also noted that the above description of the disclosed embodiments is provided to enable any person skilled in the art to make or use the present invention. Various modifications to these embodiments will be readily apparent to those skilled in the art, and the generic principles defined herein may be applied to other embodiments without departing from the spirit or scope of the invention. Thus, the present invention is not intended to be limited to the embodiments shown herein but is to be accorded the widest scope consistent with the principles and novel features disclosed herein.

Claims (10)

1. A multi-scale space-time load prediction method for a bottom-up power distribution network is characterized by comprising the following steps:
1) obtaining typical load density upper and lower bounds and high, middle and low section typical values of various types of landmass by a non-parameter kernel density estimation method;
2) extracting various user typical load curves from the electricity utilization information acquisition system by adopting a self-adaptive k-means clustering method based on DB indexes, and accurately representing the electricity utilization characteristics of the users;
3) a bottom-to-top ground load superposition method is adopted to obtain a load increase space-time panorama of a land block-grid-power supply area step by step, and selection of simultaneous rate is avoided.
2. The method for multi-scale space-time load prediction of a bottom-up distribution network according to claim 1, characterized in that the predicted areas are areas where detailed planning of control has been carried out and land use properties are known.
3. The multi-scale space-time load prediction method for the bottom-up power distribution network according to claim 2, characterized in that the typical indexes of the space load density are counted by a non-parameter kernel density estimation method, and the specific contents are as follows:
establishing a space load density probability model based on statistical data of a large number of mature plots, and setting x1,x2,…,xnIs the load density value of n samples of a certain geographic type, the probability density function of the load density is f (x), then the kernel estimate of the probability density function is:
Figure FDA0002198866570000011
in the formula, n is the sample capacity; h is the bandwidth, serving as the smoothing factor; k () is a kernel function; selecting a standard Gaussian function as a kernel function, and setting the bandwidth to be 0.5, then the kernel density estimation of the load density is as follows:
Figure FDA0002198866570000012
after obtaining the kernel density estimation curve of the load density value, in order to remove the influence of the extreme data in the sample, respectively removing 5% from the minimum and maximum load density data, then equally dividing the residual range, and selecting the extreme from each section as the typical value of the load density of the section.
4. The multi-scale space-time load prediction method for the bottom-up power distribution network according to claim 2, characterized in that a DB index is adopted to measure the load curve clustering effect, and the specific contents are as follows:
acquiring a large number of user daily load curves of the land types through the electricity utilization information acquisition system, clustering the daily load curves, taking the center line of each type as a typical load curve of the land type, and determining the appropriate clustering type number by adopting a DB index:
Figure FDA0002198866570000021
Figure FDA0002198866570000022
in the formula, CiDenotes the ith type, zjIs CiVector of (1), AiCluster center ofiP represents the number of zjAnd AiThe DB index is finally calculated according to the formula.
5. The multi-scale space-time load prediction method for the bottom-up power distribution network according to claim 4, characterized in that a DB index-based adaptive clustering algorithm is adopted to extract various typical load curves, and the specific contents are as follows:
the optimal clustering number of the k-means algorithm is searched by iteration based on DB indexes, and the process comprises 9 steps:
step 1: setting k to be 2, and inputting a load curve set to be clustered;
step 2: running a k-means algorithm;
and 3, step 3: calculating the DB index size of the current clustering scheme, and recording the DB index size as DB (k);
and 4, step 4: updating k to k + 1;
and 5, step 5: judging whether the maximum allowable clustering number is reached, namely whether k is more than kmaxIf yes, jumping to the step 9, and if not, entering the step 6;
and 6, step 6: running a k-means algorithm;
and 7, step 7: calculating the DB index size of the current clustering scheme, and recording the DB index size as DB (k);
and 8, step 8: judging whether the clustering effect is better or not based on the DB index, namely whether DB (k) is less than DB (k-1), if so, returning to the step 4, and if not, entering the step 9;
step 9: and terminating the iteration and outputting the optimal clustering number k.
6. The multi-scale space-time load prediction method for the bottom-up power distribution network according to any one of claims 1 to 5, characterized by comprising the following specific contents:
step 1: based on the controlled detailed planning of the area to be predicted and the river and the traffic main road as boundaries, dividing plots, grids and power supply areas of the area to be predicted;
step 2: based on a power consumption information acquisition system and a geographic information system, searching through the Internet to collect space load density databases of various mature users in the region of a block to be predicted, and obtaining typical load density indexes of various types of blocks through nonparametric kernel density estimation;
and 3, step 3: extracting a large number of typical daily load curves of representative users of various types of plots based on a power utilization information acquisition system, and obtaining the typical load curves of the typical plots through a K-means clustering algorithm based on DB indexes;
and 4, step 4: classifying and matching the types of the plots to be predicted to obtain load density and a typical load curve corresponding to each plot, and calculating the maximum load of each plot by combining a load density index method and an area;
and 5, step 5: based on typical load curves of each plot, a load increase space-time panorama of a plot-grid-power supply area is obtained by a bottom-up superposition method.
7. The multi-scale space-time load prediction method for a bottom-up power distribution network according to claim 6, wherein the maximum load value of the area to be predicted is calculated by the following formula:
Figure FDA0002198866570000031
in the formula, max is the symbol of maximum value, n is the number of functional blocks of the block to be predicted, and the typical daily load curve of each block comprises m-dimensional load points, wherein P is the number of functional blocksijIs the value of the jth load point, P, of the plot iiIs the maximum load of plot i, calculated by the following load density index method:
Pi＝Siρi，
in the formula, SiAnd ρiTypical values for the area and load density of the plot i, respectively.
8. The method for predicting the multi-scale space-time load of the bottom-up power distribution network according to claim 6, wherein the adopted k-means clustering algorithm is based on the distance measure of the load curve, and the following procedures are adopted:
step 1: optionally selecting k initial clustering centers Z in the load curve data set to be clustered1(l),Z2(l)…Zk(l)；
Step 2: calculating the distance from each load curve sample to k clustering center curves, and classifying according to a nearest rule; if | | | X-Zj(l)||＜||X-Zi(l) If X belongs to Gj(l) 1,2, … k, i ≠ j, wherein: gj(l) As a cluster center Zj(l) Clustering the samples; distributing all samples from X to k clustering centers in the first iteration;
and 3, step 3: calculating a new clustering center from the calculation result of the step 2;
Figure FDA0002198866570000041
wherein j is 1,2, …, k;
new clustering center should be J of criterion functionjThe value reaches the minimum:
wherein: gjIs the jth load curve class; n is a radical ofjThe number of load curve samples of the jth class; zjClustering the center of the load curve of the jth sample;
and 4, step 4: if the new cluster center is equal to the previous cluster center, then:
Zj(l+1)＝Zj(l),j＝1,2,…,k；
the algorithm is converged and clustering is finished; otherwise, go to step 2.
9. The method for predicting the multi-scale space-time load of the bottom-up power distribution network according to claim 6, wherein the daily load curves are all normalized in advance, and the method comprises the following specific contents:
assuming that the typical daily load curve is an m-dimensional array, the ith load curve is represented as Xi＝(xi,1,xi,2,…,xi,m) The normalization result of the j dimension load value is as follows:
Figure FDA0002198866570000043
in the formula, min and max are the operation signs of taking the minimum value and the maximum value respectively.
10. The utility model provides a distribution network multiscale space-time load prediction system from bottom to top which characterized in that includes:
a region to be predicted dividing unit: based on the controlled detailed planning of the area to be predicted and the river and the traffic main road as boundaries, dividing plots, grids and power supply areas of the area to be predicted;
a typical load density index acquisition unit: based on a power consumption information acquisition system and a geographic information system, searching through the Internet, collecting various mature user space load density databases in the region where the block to be predicted is located, and obtaining typical load density indexes of various types of blocks through non-parameter kernel density estimation;
typical load curve calculation unit: extracting a large number of typical daily load curves of representative users of various types of plots based on a power utilization information acquisition system, and obtaining the typical load curves of the typical plots through a K-means clustering algorithm based on DB indexes;
a maximum load calculation unit: classifying and matching the types of the plots to be predicted to obtain load density and a typical load curve corresponding to each plot, and calculating the maximum load of each plot by combining a load density index method and an area;
load growth spatiotemporal panorama acquisition unit: based on typical load curves of each plot, a load increase space-time panorama of a plot-grid-power supply area is obtained by a bottom-up superposition method.
CN201910858169.3A 2019-09-11 2019-09-11 Multi-scale space-time load prediction method and system for bottom-up power distribution network Pending CN110717619A (en)

Priority Applications (1)

Application Number Priority Date Filing Date Title
CN201910858169.3A CN110717619A (en) 2019-09-11 2019-09-11 Multi-scale space-time load prediction method and system for bottom-up power distribution network

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
CN201910858169.3A CN110717619A (en) 2019-09-11 2019-09-11 Multi-scale space-time load prediction method and system for bottom-up power distribution network

Publications (1)

Publication Number Publication Date
CN110717619A true CN110717619A (en) 2020-01-21

Family

ID=69210371

Family Applications (1)

Application Number Title Priority Date Filing Date
CN201910858169.3A Pending CN110717619A (en) 2019-09-11 2019-09-11 Multi-scale space-time load prediction method and system for bottom-up power distribution network

Country Status (1)

Country Link
CN (1) CN110717619A (en)

Cited By (8)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN111461197A (en) * 2020-03-27 2020-07-28 国网上海市电力公司 Spatial load distribution rule research method based on feature extraction
CN111461921A (en) * 2020-03-31 2020-07-28 国网湖南省电力有限公司 Load modeling typical user database updating method based on machine learning
CN111612031A (en) * 2020-04-03 2020-09-01 华电电力科学研究院有限公司 Regional building dynamic load prediction method based on high-dimensional spatial clustering neighbor search
CN111723975A (en) * 2020-05-18 2020-09-29 国网新疆电力有限公司经济技术研究院 Power distribution network electric power tight balance method based on distributed power supply output time sequence
CN112232381A (en) * 2020-09-25 2021-01-15 国网上海市电力公司 Model parameter post-processing method and device for leading load parameter noise identification
CN113887809A (en) * 2021-10-11 2022-01-04 国网新疆电力有限公司巴州供电公司 Power distribution network supply and demand balance method, system, medium and computing equipment under double-carbon target
CN114418328A (en) * 2021-12-27 2022-04-29 宁波诺丁汉大学 City functional area planning method
CN114418328B (en) * 2021-12-27 2024-06-25 宁波诺丁汉大学 Urban functional area planning method

Citations (4)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US20040102937A1 (en) * 2002-11-21 2004-05-27 Honeywell International Inc. Energy forecasting using model parameter estimation
CN106022509A (en) * 2016-05-07 2016-10-12 国网浙江省电力公司经济技术研究院 Power distribution network space load prediction method taking region and load property dual differences into consideration
CN108491969A (en) * 2018-03-16 2018-09-04 国家电网公司 Spatial Load Forecasting model building method based on big data
CN109492950A (en) * 2018-12-26 2019-03-19 广东电网有限责任公司 One kind can meet big regional scope space saturation load forecasting method based on GIS technology

Patent Citations (4)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US20040102937A1 (en) * 2002-11-21 2004-05-27 Honeywell International Inc. Energy forecasting using model parameter estimation
CN106022509A (en) * 2016-05-07 2016-10-12 国网浙江省电力公司经济技术研究院 Power distribution network space load prediction method taking region and load property dual differences into consideration
CN108491969A (en) * 2018-03-16 2018-09-04 国家电网公司 Spatial Load Forecasting model building method based on big data
CN109492950A (en) * 2018-12-26 2019-03-19 广东电网有限责任公司 One kind can meet big regional scope space saturation load forecasting method based on GIS technology

Non-Patent Citations (2)

* Cited by examiner, † Cited by third party
Title
刘思 等: "应用聚类分析与非参数核密度估计的空间负荷分布规律", 《电网技术》 *
郑伟民等: "基于Softmax概率分类器的数据驱动空间负荷预测", 《电力系统自动化》 *

Cited By (10)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN111461197A (en) * 2020-03-27 2020-07-28 国网上海市电力公司 Spatial load distribution rule research method based on feature extraction
CN111461921A (en) * 2020-03-31 2020-07-28 国网湖南省电力有限公司 Load modeling typical user database updating method based on machine learning
CN111461921B (en) * 2020-03-31 2023-11-24 国网湖南省电力有限公司 Load modeling typical user database updating method based on machine learning
CN111612031A (en) * 2020-04-03 2020-09-01 华电电力科学研究院有限公司 Regional building dynamic load prediction method based on high-dimensional spatial clustering neighbor search
CN111723975A (en) * 2020-05-18 2020-09-29 国网新疆电力有限公司经济技术研究院 Power distribution network electric power tight balance method based on distributed power supply output time sequence
CN112232381A (en) * 2020-09-25 2021-01-15 国网上海市电力公司 Model parameter post-processing method and device for leading load parameter noise identification
CN112232381B (en) * 2020-09-25 2024-03-01 国网上海市电力公司 Model parameter post-processing method and device for dominant load parameter noise identification
CN113887809A (en) * 2021-10-11 2022-01-04 国网新疆电力有限公司巴州供电公司 Power distribution network supply and demand balance method, system, medium and computing equipment under double-carbon target
CN114418328A (en) * 2021-12-27 2022-04-29 宁波诺丁汉大学 City functional area planning method
CN114418328B (en) * 2021-12-27 2024-06-25 宁波诺丁汉大学 Urban functional area planning method

Similar Documents

Publication Publication Date Title
CN110717619A (en) Multi-scale space-time load prediction method and system for bottom-up power distribution network
US11043808B2 (en) Method for identifying pattern of load cycle
CN106022509B (en) Consider the Spatial Load Forecasting For Distribution method of region and load character double differences
CN106447206A (en) Power utilization analysis method based on acquisition data of power utilization information
CN112149873B (en) Low-voltage station line loss reasonable interval prediction method based on deep learning
CN110929939B (en) Landslide hazard susceptibility spatial prediction method based on clustering-information coupling model
CN108345908A (en) Sorting technique, sorting device and the storage medium of electric network data
CN108364187A (en) A kind of power failure sensitive users based on power failure sensitivity characteristic determine method and system
CN106952027A (en) A kind of 10kV distribution network lines plan access capacity computational methods
CN110119556B (en) Space-time evolution analysis method for regional water source conservation function
CN114519514B (en) Low-voltage transformer area reasonable line loss value measuring and calculating method, system and computer equipment
CN105868906A (en) Optimized method for analyzing maturity of regional development
CN112288172A (en) Prediction method and device for line loss rate of transformer area
CN115329899A (en) Clustering equivalent model construction method, system, equipment and storage medium
CN110543660A (en) low-impact development simulation method, system and related device
CN109858667A (en) It is a kind of based on thunder and lightning weather to the short term clustering method of loading effects
CN110264010B (en) Novel rural power saturation load prediction method
CN112328851A (en) Distributed power supply monitoring method and device and electronic equipment
CN116720750A (en) Comprehensive energy station equipment capacity planning method, device, terminal and storage medium
CN109889981B (en) Positioning method and system based on binary classification technology
CN110852547B (en) Public service facility grading method based on position data and clustering algorithm
CN111417132B (en) Cell division method, device and equipment
CN115905319B (en) Automatic identification method and system for abnormal electricity fees of massive users
CN111144628A (en) Distributed energy supply type cooling, heating and power load prediction model system and method
CN110188964A (en) A kind of photovoltaic power generation output forecasting method based on correlation

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
WD01 Invention patent application deemed withdrawn after publication
WD01 Invention patent application deemed withdrawn after publication

Application publication date: 20200121