CN113688884A - Load level classification method and device based on cloud platform - Google Patents
Load level classification method and device based on cloud platform Download PDFInfo
- Publication number
- CN113688884A CN113688884A CN202110902403.5A CN202110902403A CN113688884A CN 113688884 A CN113688884 A CN 113688884A CN 202110902403 A CN202110902403 A CN 202110902403A CN 113688884 A CN113688884 A CN 113688884A
- Authority
- CN
- China
- Prior art keywords
- monitoring data
- power load
- load monitoring
- determining
- cluster
- 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
Images
Classifications
-
- 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
-
- 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
- G06F17/16—Matrix or vector computation, e.g. matrix-matrix or matrix-vector multiplication, matrix factorization
Landscapes
- Engineering & Computer Science (AREA)
- Physics & Mathematics (AREA)
- Data Mining & Analysis (AREA)
- General Physics & Mathematics (AREA)
- Theoretical Computer Science (AREA)
- Mathematical Physics (AREA)
- General Engineering & Computer Science (AREA)
- Pure & Applied Mathematics (AREA)
- Mathematical Optimization (AREA)
- Mathematical Analysis (AREA)
- Computational Mathematics (AREA)
- Evolutionary Computation (AREA)
- Artificial Intelligence (AREA)
- Life Sciences & Earth Sciences (AREA)
- Bioinformatics & Cheminformatics (AREA)
- Computing Systems (AREA)
- Evolutionary Biology (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Probability & Statistics with Applications (AREA)
- Bioinformatics & Computational Biology (AREA)
- Algebra (AREA)
- Databases & Information Systems (AREA)
- Software Systems (AREA)
- Management, Administration, Business Operations System, And Electronic Commerce (AREA)
Abstract
A load level classification method and device based on a cloud platform belong to the technical field of load identification under a non-intrusive power detection system. Determining the average Euclidean distance, the sample density and the inter-cluster distance among the power load monitoring data of the samples, and further determining the maximum value of the Euclidean distance of the samples between two clusters; determining the density weight of the clustering centers, and determining all initial clustering centers; traversing all the power load monitoring data, and determining a cluster group where the power load monitoring data are located; traversing the power load monitoring data in each cluster, and selecting the power load monitoring data with the minimum sum of Euclidean distances as a clustering center; and taking the power load monitoring data with the minimum Euclidean distance as the characteristic value of the power load monitoring data. The invention can realize the horizontal classification and identification of the power load, and provides a technical basis for the efficient energy-saving utilization of the power load, the reasonable adjustment of the load power utilization condition, the reduction of the peak-valley difference and the reduction of the power utilization cost.
Description
Technical Field
A load level classification method and device based on a cloud platform belong to the technical field of load identification under a non-intrusive power detection system.
Background
The load level classification method plays an increasingly important role in the planning and development of the smart grid, and can play an important role in the aspects of data management, data mining and energy utilization management of the power system. Therefore, effective utilization information of the power users can be found based on the load level classification of the cloud platform, the energy and power utilization behaviors of the users are analyzed, a technical basis is provided for the energy-saving scheme of the large-scale industrial power users, meanwhile, the power utilization cost of the users can be improved, and electric energy is saved.
Currently, the K-means based method is the mainstream load level division method. The K-means method has the advantages of low time complexity, optimized iteration function and capability of overcoming the inaccuracy of clustering of a small number of samples. But at the same time, the defects are obvious, the optimal clustering number needs to be set manually, and the clustering number cannot be given by calculating a data optimization result. In addition, the center of the cluster where the data is located needs to be initialized randomly, which causes different convergence conditions of the data each time, and affects the accuracy of the clustered data. Therefore, in order to solve the problems of the K-means method, a more effective method for classifying the load level needs to be constructed.
Disclosure of Invention
The technical problem to be solved by the invention is as follows: the load level classification method and device based on the cloud platform can well determine the number of clusters and select a reasonable initial cluster center, and ensure the accuracy and effectiveness of load level classification.
The technical scheme adopted by the invention for solving the technical problems is as follows: the load level classification method based on the cloud platform is characterized by comprising the following steps: the method comprises the following steps:
acquiring power load monitoring data of M dimensions of N samples through a cloud platform, and creating a load data matrix XN×M:
Wherein x isi,jPower load monitoring data for the jth dimension of the ith sample, i 1, 2,.., N, j 1, 2.., M;
processing the power load monitoring data, determining the average Euclidean distance, the sample density and the inter-cluster distance between the power load monitoring data of the samples, and further determining the maximum value of the Euclidean distance of the samples between two clusters;
determining the density weight of the clustering centers, and determining all initial clustering centers;
traversing all the power load monitoring data, and determining a cluster group where the power load monitoring data are located;
traversing the power load monitoring data in each cluster, and selecting the power load monitoring data with the minimum sum of Euclidean distances as a clustering center;
and taking the power load monitoring data with the minimum Euclidean distance as the characteristic value of the power load monitoring data, and outputting a load level classification result.
Preferably, M power load monitoring data of each of the N time intervals are acquired by the cradle head.
Preferably, the method further comprises determining the euclidean distance d between the electrical load monitoring data of the samplesi,j;
Wherein x isi′N, i' is a sample other than sample i, 1, 2.
Preferably, the method further comprises calculating an average euclidean distance avg _ d of all the power load monitoring data in the sample;
where n is the number of all power load monitoring data in the sample.
Preferably, the method further comprises determining a sample density ρ of a cluster corresponding to a cluster centeri:
ρi=ρi+1,(xi-xi′)≤avg_d。
Preferably, the method further comprises: determining inter-cluster distance si:
si=max di,j。
Preferably, the method further comprises calculating an average distance α between samples of the clusteri:
Determining inter-cluster distance si:
si=max di,j。
Preferably, the method further comprises calculating a density weight w of the cluster centeri:
The utility model provides a load level classification device based on cloud platform which characterized in that: comprises that
The data acquisition module is used for acquiring power load monitoring data of M dimensions of each of N samples through the cloud platform and creating a load data matrix;
the data calculation module is used for processing the power load monitoring data, determining the average Euclidean distance, the sample density and the inter-cluster distance between the power load monitoring data of the samples, and further determining the maximum value of the Euclidean distance of the samples between the two clusters;
the data classification module is used for determining the density weight of the clustering centers and determining all initial clustering centers;
the data traversing module is used for traversing all the power load monitoring data and determining the cluster in which the power load monitoring data is positioned;
the data selection module is used for traversing the power load monitoring data in each cluster and selecting the power load monitoring data with the minimum sum of Euclidean distances as a clustering center;
and the data output module is used for taking the power load monitoring data with the minimum Euclidean distance as the characteristic value of the power load monitoring data and outputting the load level classification result.
Compared with the prior art, the invention has the beneficial effects that:
the load level classification method based on the cloud platform can realize level classification and identification of the power load on the basis of the cloud platform for monitoring the power load data under the condition that the privacy of users is not invaded.
According to the load level classification method based on the cloud platform, the initial calculation mode of the cluster center is improved, the initial center can be determined without manual setting, the center is determined by calculating the minimum sum of the distances between data and other data curves, the influence of the random initialization cluster center value on the accuracy and accuracy of the whole data classification is avoided, the initial cluster center and the cluster number can be determined well, and the accuracy and precision of the load classification are improved.
Drawings
Fig. 1 is a flowchart of a load level classification method based on a cloud platform.
Detailed Description
The present invention is further described with reference to the following detailed description, however, it should be understood by those skilled in the art that the detailed description given herein with respect to the accompanying drawings is for better explanation and that the present invention is not necessarily limited to the specific embodiments, but rather, for equivalent alternatives or common approaches, may be omitted from the detailed description, while still remaining within the scope of the present application.
Specifically, the method comprises the following steps: as shown in fig. 1: the load level classification method based on the cloud platform comprises the following steps:
and acquiring power load monitoring data of M dimensions of each of the N samples through the cloud platform.
In this embodiment, a month time is obtained as a sample through the cloud platform, and one piece of power load monitoring data is recorded every 15 minutes, so that a load data matrix X of 96 dimensions of 30 samples can be obtained30×96:
And processing the power load monitoring data, determining the average Euclidean distance, the sample density and the inter-cluster distance among the power load monitoring data of the samples, and further determining the maximum value of the Euclidean distance of the samples between the two clusters.
Determining Euclidean distance d between power load monitoring data of samplesi,j:
Calculating the average Euclidean distance avg _ d of all data in the sample;
determining sample density rho of cluster corresponding to cluster centeri:
ρi=ρi+1,(xi-xi′)≤avg_d。
Selecting the data object with the maximum sample density as a first clustering center J1:
J1=max(ρ)。
Calculating the average distance alpha between samples of a clusteri:
Determining inter-cluster distance si:
si=max di,j。
Determining the density weight of the cluster centers and determining all initial cluster centers.
Calculating a density weight w of a cluster centeri:
Selecting the maximum density weight of the cluster center as the secondTwo cluster centers J2:
J2=max(ρ)。
And eliminating the first cluster center, the second cluster center and the data contained in the cluster group in which the first cluster center, the second cluster center and the second cluster center are positioned, repeatedly calculating the density weight of the rest power load monitoring data, selecting one cluster center, and determining all initial cluster centers.
And traversing all the power load monitoring data, and determining the cluster group where the power load monitoring data is located.
And traversing and calculating each power load monitoring data in the load data matrix, classifying the power load monitoring data, determining the cluster where the power load monitoring data is located, and obtaining all initial cluster centers and cluster numbers.
And traversing the power load monitoring data in each cluster, and selecting the power load monitoring data with the minimum sum of Euclidean distances as a clustering center.
Traversing the data in each cluster according to a K-medoids clustering algorithm, and selecting the power load monitoring data with the minimum sum of Euclidean distances as a clustering center.
And taking the power load monitoring data with the minimum Euclidean distance as the characteristic value of the power load monitoring data, and outputting a load level classification result.
A load level classification device based on a cloud platform comprises
The data acquisition module is used for acquiring power load detection data of M dimensions of each of N samples through the cloud platform and creating a load data matrix;
the data calculation module is used for processing the power load monitoring data, determining the average Euclidean distance, the sample density and the inter-cluster distance between the power load monitoring data of the samples, and further determining the maximum value of the Euclidean distance of the samples between the two clusters;
the data classification module is used for determining the density weight of the clustering centers and determining all initial clustering centers;
the data traversing module is used for traversing all the power load monitoring data and determining the cluster in which the power load monitoring data is positioned;
the data evaluation module is used for traversing the power load monitoring data in each cluster and selecting the power load monitoring data with the minimum sum of Euclidean distances as a clustering center;
and the data output module is used for taking the power load monitoring data with the minimum Euclidean distance as the characteristic value of the power load monitoring data and outputting the load level classification result.
The foregoing is directed to preferred embodiments of the present invention, other and further embodiments of the invention may be devised without departing from the basic scope thereof, and the scope thereof is determined by the claims that follow. However, any simple modification, equivalent change and modification of the above embodiments according to the technical essence of the present invention are within the protection scope of the technical solution of the present invention.
Claims (9)
1. A load level classification method based on a cloud platform is characterized by comprising the following steps: the method comprises the following steps:
acquiring power load monitoring data of M dimensions of N samples through a cloud platform, and creating a load data matrix XN×M:
Wherein x isi,jPower load monitoring data for the jth dimension of the ith sample, i 1, 2,.., N, j 1, 2.., M;
processing the power load monitoring data, determining the average Euclidean distance, the sample density and the inter-cluster distance between the power load monitoring data of the samples, and further determining the maximum value of the Euclidean distance of the samples between two clusters;
determining the density weight of the clustering centers, and determining all initial clustering centers;
traversing all the power load monitoring data, and determining a cluster group where the power load monitoring data are located;
traversing the power load monitoring data in each cluster, and selecting the power load monitoring data with the minimum sum of Euclidean distances as a clustering center;
and taking the power load monitoring data with the minimum Euclidean distance as the characteristic value of the power load monitoring data, and outputting a load level classification result.
2. The cloud platform-based load level classification method according to claim 1, wherein: and acquiring M power load monitoring data of N time intervals through the holder.
5. The cloud platform-based load level classification method of claim 3, wherein: the method also comprises the step of determining the sample density rho of the cluster corresponding to one cluster centeri:
ρi=ρi+1,(xi-xi)≤avg_d。
6. The cloud platform-based load level classification method of claim 5, wherein: the method further comprises the following steps: determining inter-cluster distance si:
si=max di,j。
9. The utility model provides a load level classification device based on cloud platform which characterized in that: comprises that
The data acquisition module is used for acquiring power load monitoring data of M dimensions of each of N samples through the cloud platform and creating a load data matrix;
the data calculation module is used for processing the power load monitoring data, determining the average Euclidean distance, the sample density and the inter-cluster distance between the power load monitoring data of the samples, and further determining the maximum value of the Euclidean distance of the samples between the two clusters;
the data classification module is used for determining the density weight of the clustering centers and determining all initial clustering centers;
the data traversing module is used for traversing all the power load monitoring data and determining the cluster in which the power load monitoring data is positioned;
the data selection module is used for traversing the power load monitoring data in each cluster and selecting the power load monitoring data with the minimum sum of Euclidean distances as a clustering center;
and the data output module is used for taking the power load monitoring data with the minimum Euclidean distance as the characteristic value of the power load monitoring data and outputting the load level classification result.
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN202110902403.5A CN113688884A (en) | 2021-08-06 | 2021-08-06 | Load level classification method and device based on cloud platform |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN202110902403.5A CN113688884A (en) | 2021-08-06 | 2021-08-06 | Load level classification method and device based on cloud platform |
Publications (1)
Publication Number | Publication Date |
---|---|
CN113688884A true CN113688884A (en) | 2021-11-23 |
Family
ID=78579052
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN202110902403.5A Pending CN113688884A (en) | 2021-08-06 | 2021-08-06 | Load level classification method and device based on cloud platform |
Country Status (1)
Country | Link |
---|---|
CN (1) | CN113688884A (en) |
Citations (3)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
WO2021073462A1 (en) * | 2019-10-15 | 2021-04-22 | 国网浙江省电力有限公司台州供电公司 | 10 kv static load model parameter identification method based on similar daily load curves |
CN112785457A (en) * | 2021-02-02 | 2021-05-11 | 东南大学 | Typical scene construction method of distributed energy supply system based on panel data |
CN112819299A (en) * | 2021-01-21 | 2021-05-18 | 上海电力大学 | Differential K-means load clustering method based on center optimization |
-
2021
- 2021-08-06 CN CN202110902403.5A patent/CN113688884A/en active Pending
Patent Citations (3)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
WO2021073462A1 (en) * | 2019-10-15 | 2021-04-22 | 国网浙江省电力有限公司台州供电公司 | 10 kv static load model parameter identification method based on similar daily load curves |
CN112819299A (en) * | 2021-01-21 | 2021-05-18 | 上海电力大学 | Differential K-means load clustering method based on center optimization |
CN112785457A (en) * | 2021-02-02 | 2021-05-11 | 东南大学 | Typical scene construction method of distributed energy supply system based on panel data |
Non-Patent Citations (2)
Title |
---|
张承畅等: "基于云计算和改进K-means算法的海量用电数据分析方法", 计算机应用, vol. 38, no. 1, pages 159 - 164 * |
彭勃;李作红;李猛;杨燕;徐蔚;麻敏华;: "基于近邻传播算法的电力用户负荷曲线聚类分析", 机电工程技术, no. 04, pages 197 - 200 * |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
WO2022110557A1 (en) | Method and device for diagnosing user-transformer relationship anomaly in transformer area | |
CN106022509B (en) | Consider the Spatial Load Forecasting For Distribution method of region and load character double differences | |
CN109492667A (en) | A kind of feature selecting discrimination method for non-intrusive electrical load monitoring | |
EP3968040A1 (en) | Abnormal electricity use recognition method and device, and computer readable storage medium | |
CN106228554A (en) | Fuzzy coarse central coal dust image partition methods based on many attribute reductions | |
CN109947815B (en) | Power theft identification method based on outlier algorithm | |
CN110795690A (en) | Wind power plant operation abnormal data detection method | |
CN111340065B (en) | User load electricity stealing model mining system and method based on complex user behavior analysis | |
CN105825232A (en) | Classification method and device for electromobile users | |
CN110738232A (en) | grid voltage out-of-limit cause diagnosis method based on data mining technology | |
CN117113235B (en) | Cloud computing data center energy consumption optimization method and system | |
CN112002949A (en) | Monitoring management method and system for active equalization of power battery | |
CN110991527A (en) | Similarity threshold determination method considering voltage curve average fluctuation rate | |
CN113125903A (en) | Line loss anomaly detection method, device, equipment and computer-readable storage medium | |
CN115422788A (en) | Power distribution network line loss analysis management method, device, storage medium and system | |
CN109948738B (en) | Energy consumption abnormity detection method and device for coating drying chamber | |
CN113094448B (en) | Analysis method and analysis device for residence empty state and electronic equipment | |
CN109389172B (en) | Radio signal data clustering method based on non-parameter grid | |
CN112528762B (en) | Harmonic source identification method based on data correlation analysis | |
CN106651630B (en) | Crucial electricity consumption industry recognition methods and system | |
CN108875801A (en) | A kind of Classification of Load Curves system based on smart grid | |
CN107274025B (en) | System and method for realizing intelligent identification and management of power consumption mode | |
CN113688884A (en) | Load level classification method and device based on cloud platform | |
CN116846332A (en) | Photovoltaic power station equipment power generation performance diagnosis method, device and medium | |
CN109783586B (en) | Water army comment detection method based on clustering resampling |
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 |