[go: up one dir, main page]
More Web Proxy on the site http://driver.im/

CN111324790A - Load type identification method based on support vector machine classification - Google Patents

Load type identification method based on support vector machine classification Download PDF

Info

Publication number
CN111324790A
CN111324790A CN202010110937.XA CN202010110937A CN111324790A CN 111324790 A CN111324790 A CN 111324790A CN 202010110937 A CN202010110937 A CN 202010110937A CN 111324790 A CN111324790 A CN 111324790A
Authority
CN
China
Prior art keywords
load
year
data
support vector
vector machine
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
CN202010110937.XA
Other languages
Chinese (zh)
Inventor
李宁
袁铁江
郭泽林
杨金成
王永超
齐尚敏
白银平
李娜
潘成龙
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.)
Dalian University of Technology
Electric Power Research Institute of State Grid Xinjiang Electric Power Co Ltd
Original Assignee
Dalian University of Technology
Electric Power Research Institute of State Grid Xinjiang 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 Dalian University of Technology, Electric Power Research Institute of State Grid Xinjiang Electric Power Co Ltd filed Critical Dalian University of Technology
Priority to CN202010110937.XA priority Critical patent/CN111324790A/en
Publication of CN111324790A publication Critical patent/CN111324790A/en
Pending legal-status Critical Current

Links

Images

Classifications

    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06F—ELECTRIC DIGITAL DATA PROCESSING
    • G06F16/00—Information retrieval; Database structures therefor; File system structures therefor
    • G06F16/90—Details of database functions independent of the retrieved data types
    • G06F16/906—Clustering; Classification
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06F—ELECTRIC DIGITAL DATA PROCESSING
    • G06F18/00—Pattern recognition
    • G06F18/20—Analysing
    • G06F18/24—Classification techniques
    • G06F18/241—Classification techniques relating to the classification model, e.g. parametric or non-parametric approaches
    • G06F18/2411—Classification techniques relating to the classification model, e.g. parametric or non-parametric approaches based on the proximity to a decision surface, e.g. support vector machines
    • 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
    • Y—GENERAL TAGGING OF NEW TECHNOLOGICAL DEVELOPMENTS; GENERAL TAGGING OF CROSS-SECTIONAL TECHNOLOGIES SPANNING OVER SEVERAL SECTIONS OF THE IPC; TECHNICAL SUBJECTS COVERED BY FORMER USPC CROSS-REFERENCE ART COLLECTIONS [XRACs] AND DIGESTS
    • Y02—TECHNOLOGIES OR APPLICATIONS FOR MITIGATION OR ADAPTATION AGAINST CLIMATE CHANGE
    • Y02D—CLIMATE CHANGE MITIGATION TECHNOLOGIES IN INFORMATION AND COMMUNICATION TECHNOLOGIES [ICT], I.E. INFORMATION AND COMMUNICATION TECHNOLOGIES AIMING AT THE REDUCTION OF THEIR OWN ENERGY USE
    • Y02D10/00—Energy efficient computing, e.g. low power processors, power management or thermal management

Landscapes

  • Engineering & Computer Science (AREA)
  • Theoretical Computer Science (AREA)
  • Data Mining & Analysis (AREA)
  • Business, Economics & Management (AREA)
  • General Physics & Mathematics (AREA)
  • Physics & Mathematics (AREA)
  • General Engineering & Computer Science (AREA)
  • Databases & Information Systems (AREA)
  • Economics (AREA)
  • Health & Medical Sciences (AREA)
  • Computer Vision & Pattern Recognition (AREA)
  • Human Resources & Organizations (AREA)
  • Evolutionary Biology (AREA)
  • Bioinformatics & Computational Biology (AREA)
  • Bioinformatics & Cheminformatics (AREA)
  • Artificial Intelligence (AREA)
  • Public Health (AREA)
  • Water Supply & Treatment (AREA)
  • General Health & Medical Sciences (AREA)
  • Evolutionary Computation (AREA)
  • Marketing (AREA)
  • Primary Health Care (AREA)
  • Strategic Management (AREA)
  • Tourism & Hospitality (AREA)
  • General Business, Economics & Management (AREA)
  • Life Sciences & Earth Sciences (AREA)
  • Information Retrieval, Db Structures And Fs Structures Therefor (AREA)

Abstract

A load type identification method based on support vector machine classification analyzes the electricity consumption behavior habit characteristics of industrial, agricultural, commercial and municipal residents, obtains daily load and annual load electricity consumption data of four typical load types of the industrial, agricultural, commercial and municipal residents based on the user electricity consumption data acquired by an electricity consumption information acquisition system, and constructs a multi-dimensional load characteristic classification index; establishing a plurality of vector machine classifiers of the data according to a one-to-many strategy; searching a maximum value element and a minimum value element in the load data, and performing dispersion standardization processing on data of different types of loads in various orders of magnitude; dividing the acquired electricity consumption data into a training group and an experimental group, wherein the training group is used for training a support vector machine classifier, and the experimental group is used for classifying and identifying the load type of the experimental group; inputting the training set into a support vector machine classifier for training; and inputting the experimental group data into a support vector machine classifier to obtain a load type identification result of each experimental group.

Description

Load type identification method based on support vector machine classification
Technical Field
The invention relates to a power load type identification method.
Background
With the development of the current economy and scientific technology in China, the intelligentization and informatization levels of the power grid are continuously improved, and the electric power plays a significant role in the current energy pattern. In order to better improve the service quality of power supply companies to users, the electricity utilization information acquisition system has been greatly improved in recent years. The initial electricity consumption information acquisition system is only used for acquiring the electricity consumption of the user to charge, and later along with the continuous development of related technologies, the data acquisition types of the electricity consumption information acquisition system are continuously increased, so that an electric power company can obtain abundant original data, and the development of electric power big data related research is promoted.
In recent years, a great deal of relevant literature research has been conducted on the load curve clustering problem. At present, most scholars focus on unsupervised clustering algorithm research, such as a K-means clustering algorithm and the like, and the algorithm can achieve the purpose of load curve clustering, but has some problems and disadvantages. For example, the K-means clustering algorithm needs to manually specify the number of clusters K and the cluster center before use, which makes the selection of the K value very difficult, and when the data size is quite large, the K-means clustering algorithm will consume a lot of time and cannot meet the relevant requirements under the background of large data. In addition, in the context of big data, user load data is quite huge, and although an unsupervised clustering algorithm can cluster load curves, it is difficult to identify the types to which the load curves belong, so that additional artificial load curve analysis is required.
In summary, currently, there are few studies considering the use of a support vector classifier model in load type classification studies.
Disclosure of Invention
The invention provides a load type identification method based on a support vector machine classifier, aiming at the defects of the existing research on load type classification research. The method comprises the steps of collecting power consumption data of power consumers with different load types in a power consumption information collection system, constructing load characteristic indexes including load rate, peak-valley difference rate, average load, load change rate and the like, introducing load type corresponding labels, establishing load characteristic models corresponding to daily load curves and annual load curves, and realizing load type identification and classification of load data to be predicted by dividing a training group and an experimental group and utilizing a support vector machine classifier.
1. The characteristics of the electricity utilization behavior habits of industrial, agricultural, commercial and municipal residential users are researched and analyzed. Acquiring user power consumption data based on a power consumption information acquisition system, acquiring daily load and annual load power consumption data of four typical load types of industrial, agricultural, commercial and municipal residential users, and constructing a multi-dimensional load characteristic classification index;
2. establishing a support vector machine classifier model, and realizing multi-classification of data according to a one-to-many strategy;
3. searching a maximum value element and a minimum value element in the load data, and performing dispersion standardization processing on data of different types of loads in various orders of magnitude;
4. dividing power consumption data acquired from a power consumption information acquisition system into a training group and an experimental group according to a ratio of 8:2, wherein the training group is used for training a support vector machine classifier, and the experimental group is used for classifying and identifying the load type of the experimental group;
5. inputting the training set into a support vector machine classifier for training;
6. and inputting the experimental group data into a support vector machine classifier to obtain a load type identification result of each experimental group.
In the step 1: currently, in the process of classifying and researching the types of electric loads, the loads are generally classified into four types, namely industrial loads, agricultural loads, commercial loads, municipal loads and living loads. The four load types have respective characteristics, and the change rules of daily load and annual load power consumption are different, so that related load characteristic classification indexes can be constructed according to the characteristics, and therefore, the power consumption behavior characteristics of the users under the four load types need to be analyzed.
The daily load value of the industrial load is large, the industrial load is basically maintained at a high level all day long, and the requirement on the reliability of power supply is high; the most obvious characteristics of agricultural load are that the season is strong, the agricultural load is extremely unbalanced in the year, and the agricultural load is characterized by uneven distribution and low density; the load curves of a large mall, an office building and the like in a day under the load generally show that the difference between the peak and the valley of electricity utilization is large, namely the difference between the peak and the valley is large; the municipal load refers to facilities for serving residents in a city, the daily load change rule of the facilities is similar to the load of residents, and the municipal load is usually expressed in that the peak hours of electricity consumption occur at about 12 noon and 8 pm.
Because the daily load electricity consumption can describe the change rule of the load electricity consumption in one day, for industrial, commercial and residential users, the daily load electricity consumption can present electricity consumption conditions of different load types, and a daily load curve can be conveniently obtained by utilizing the electricity consumption information acquisition system. However, the agricultural load presents obvious seasonality, the judgment is difficult to be carried out according to the daily load curve, a load curve capable of describing the change rule of the agricultural load in one year needs to be found, and the method adopts the annual load curve as the basis to construct the related load characteristic classification indexes to carry out the judgment on the agricultural load type. The classification result needs to satisfy the condition that the user electricity load curves have higher similarity under the same load type, and the load curves under different load types have larger dissimilarity.
The load electricity utilization data of 24 hours a day and the average load electricity utilization data of 12 months a year can be acquired from the electricity utilization information acquisition system.
The daily average load refers to the average of the load electricity consumption data 24 hours a day, and the monthly average load refers to the average of the load electricity consumption of 12 months a year, as shown in the following formula:
Figure BDA0002388424370000021
Figure BDA0002388424370000022
wherein, ηDay(s)Representing average daily load, ηYear of yearDenotes the average load per month, x1，x2，...，x24Load power consumption per hour for 24 hours a day, n1，n2，...，n12Representing the monthly load electricity usage of 12 months a year.
The load average value/load maximum value is the load rate. The maximum value of daily load electricity consumption of 24 hours a day is used as a denominator, and daily average load is used as a numerator, so that the daily load rate and the annual load rate of various load types are constructed.
Figure BDA0002388424370000031
Figure BDA0002388424370000032
Wherein, deltaDay(s)、δYear of yearThe daily load rate and the annual load rate are indicated, respectively.
The quotient of the difference between the maximum value of the load electricity consumption and the minimum value of the load electricity consumption divided by the maximum value of the load electricity consumption is the peak-valley difference rate:
Figure BDA0002388424370000033
Figure BDA0002388424370000034
in the above formula, λDay(s)、λYear of yearRespectively representing the daily peak-valley difference rate and the annual peak-valley difference rate.
The difference between the electricity usage amount in one hour after and one hour before the 24 hours is the daily load change rate, and the difference between the average electricity usage amount in one month after and one month before the 12 months is the annual load change rate:
γday(s)＝(x2-x1,x3-x2,...,x24-x23)
γYear of year＝(n2-n1,n3-n2,...,n12-n11)
In the above formula, γDay(s)、γYear of yearRespectively representing the daily load change rate and the annual load change.
The load classification index obtained finally is as follows:
Yday(s)＝(ηDay(s)，δDay(s)，λDay(s),x2-x1,x3-x2,...,x24-x23,di)
YYear of year＝(ηYear of year，δYear of year，λYear of year,n2-n1,n3-n2,...,n12-n11,di)
In the above formula, YDay(s)Indicates a daily load classification index, YYear of yearRepresenting annual load classification index, ηDay(s)Representing average daily load, ηYear of yearRepresenting the annual average load, delta representing the load rate, lambda representing the peak-to-valley difference rate, xiIndicating the amount of electricity used, diAnd indicating a label corresponding to the load type. Wherein, the dimension of the daily load classification index is 27, and the dimension of the annual load classification index is 15.
In the step 2, the basic principle and the model establishing process of the support vector machine classifier are as follows:
the model of the support vector machine is actually a linear classifier that maximizes the separation of the samples in feature space. Assume that there is a linear function:
g(x)＝wx+b
where x represents sample data, w represents the slope of the linear classification function, and b represents the intercept of the linear function. Setting the threshold value to 0, when there is a sample xiWhen the judgment is needed, calculating a linear function g (x) corresponding to the samplei) If greater than zero, those are classified as one class, otherwise they are classified as another class.
Taking binary classification as an example, in the linear classification problem, the classification flags are set to 1 and-1. The Euclidean distance of a sample to the hyperplane can be represented by the normalized following equation:
Figure BDA0002388424370000041
where | w | | represents the norm of the vector w.
Defining the distance from the nearest point of the hyperplane in a group of data sample points as the distance from the group of sample points to a certain hyperplane, namely a geometric interval, wherein the geometric interval has the following relation with the fraction error of the samples:
Figure BDA0002388424370000042
wherein δ represents the geometric interval from the data sample set to the classification hyperplane, and R ═ max | | xiAnd R represents the value of the longest vector length among all data samples. The above-mentioned number of wrong divisions represents the error of the support vector classifier to some extent, and the larger the geometric interval is, the smaller the upper limit of the number of wrong divisions is, that is, the smaller the upper limit of the error is. Therefore, maximizing the geometric spacing is the goal of the data training phase.
From the above formula, the maximum geometric interval is the minimum | | | w | |. The value of w is determined by the position of the sample point and the type of the sample point, so the original g (x) can be expressed as:
Figure BDA0002388424370000043
wherein, in the above formula, aiCalled Lagrange multiplier, yiIndicates the class to which the sample point belongs, xiIs a sample point vector, i ═ 1, 2.. n, n denotes the number of sample points.<w,x>Representing the inner product of vectors w and x. In the above formula, only xiAnd x represents a vector, so a can be expressediyiTaking out, the formula becomes:
Figure BDA0002388424370000044
the problem is then transformed from solving w to solving a.
The basic function of the kernel function is to accept vectors in two low-dimensional spaces<w,x>To thereby calculate a high dimensional space<w',x'>. Taking fault tolerance into account, introducing a slack variable ζi(ii) a Considering the loss caused by fault tolerance, a penalty factor is introduced, so that the objective function of the optimization problem is as follows:
Figure BDA0002388424370000045
the constraint conditions are as follows:
yi[(wxi)+b]≥1-ζi
ζi≥0
in the above formula, w and b are both elements in a linear function g (x) ═ wx + b, and represent the slope and intercept of the linear function, respectively; | w | non-woven phosphor2A two-norm representing the slope of a linear function; c represents a penalty factor; zetaiIs a relaxation variable; y isiRepresenting classification marks, and in a binary problem, 1 and-1 can be taken; n, n denotes the number of sample points.
The support vector machine model is the minimum value of the target function searched under the constraint condition, so that the classification of the samples is realized.
The support vector machine is designed for the two-classification problem, and when the multi-classification problem is processed, a proper multi-class classifier needs to be constructed.
The one-to-many strategy sequentially classifies samples of a certain category into one category according to K category sequences during sample training, and other category samples into one category, so that K support vector machine classifiers are constructed by the samples of the K categories. The load types studied by the invention can be roughly divided into four major classes, so when a one-to-many strategy is adopted, 4 support vector machine classifiers are constructed. When the sample class is predicted, the geometric distance delta between the sample and the established 4 support vector machine classifiers is respectively judged, and the sample is classified into the class with the maximum geometric distance delta.
And in the step 3, acquiring user load information of the electric energy meters in each distribution area based on the electricity utilization information acquisition system. Acquiring data every 1 hour to obtain daily load data of 24 hours in one day of each load node; and (4) taking the average value of the electricity consumption of the users in each month to obtain the annual load data of 12 months in one year of each load node. Because the power consumption difference of each user under different load types is huge, the collected data is considered to be subjected to standardization processing, and a maximum value element and a minimum value element in the load data are searched, so that the processed data are in a uniform threshold range.
Figure BDA0002388424370000051
In the above formula, XiRepresents load data of node i, min { X }iDenotes the minimum value of the load data, max { X }iDenotes the maximum value in the load data, YiRepresenting the load data after the normalization process.
In the step 4, after the power consumption data is acquired from the power consumption information acquisition system, a training group and an experimental group are divided according to a ratio of 8:2, wherein the training group is used for training a support vector machine classifier, and the experimental group is used for classifying and identifying the load type of the experimental group.
In the step 5, a computer software program is used for sample training, training data is imported, the format of the specified sample is that the first n-1 dimension represents sample training data, and the nth dimension represents a sample training label. The training of the samples is performed using a "one-to-many" strategy,
in the step 6, a sample test is performed through a computer software program, similar to the format of the training sample data in the step 5, test data is imported, the format of the test sample is that the first n-1 dimension represents sample training data, and the nth dimension represents a sample training label. And outputting a load type label of the test sample according to the output result, and identifying the load type of the corresponding sample.
Drawings
FIG. 1 is a flow chart of a load type identification method based on support vector machine classification according to the present invention;
FIG. 2 is a graph of daily load for different load types;
fig. 3 is an annual load graph of different load types.
Detailed Description
The invention is further described with reference to the following drawings and detailed description.
As shown in fig. 1, the load type identification method based on support vector machine classification according to the present invention has the following processes:
1. the characteristics of the electricity utilization behavior habits of industrial, agricultural, commercial and municipal residential users are researched and analyzed. Acquiring user electricity data based on an electricity information acquisition system, and standardizing a daily load curve and an annual load curve of a typical load type as shown in fig. 2 and 3;
2. establishing a support vector machine classifier model, and realizing multi-classification of data according to a one-to-many strategy;
3. searching a maximum value element and a minimum value element in the load data, and carrying out dispersion standardization treatment on data of different types of loads in various orders of magnitude;
4. dividing the power utilization data acquired from the power utilization information acquisition system into a training group and an experimental group according to a certain proportion;
5. inputting the training set into a support vector machine classifier for training;
6. inputting the experimental group data into a support vector machine classifier to obtain a load type identification result of each experimental group;
in summary, the present invention applies a support vector machine classifier to load type classification. The method is a labeled classification method, avoids the condition that most clustering methods need to manually appoint a clustering center, and clustering results do not need manual identification any more.

Claims (4)

1. A load type identification method based on support vector machine classification is characterized in that: the load characteristic curves of users with different load types are used as the classification basis of the support vector machine to realize the identification of the load curves with different load types, and the steps are as follows:
step 1: the method comprises the steps of researching and analyzing the power utilization behavior habit characteristics of industrial, agricultural, commercial and municipal residential users, acquiring user power utilization data based on a power utilization information acquisition system, obtaining daily load and annual load power utilization data of four typical load types of the industrial, agricultural, commercial and municipal residential users, and constructing a multi-dimensional load characteristic classification index;
step 2: establishing a support vector machine classifier model, and realizing multi-classification of data according to a one-to-many strategy;
during training samples, classifying samples of a certain category into one category according to 4 category sequences, classifying samples of other categories into one category, and constructing 4 classifiers of support vector machines; when the sample category is predicted, respectively judging the geometric distance delta between the sample and the established 4 support vector machine classifiers, and classifying the sample into a class with the maximum geometric distance delta;
and step 3: searching a maximum value element and a minimum value element in the load data, and performing dispersion standardization processing on data of different types of loads in various orders of magnitude;
and 4, step 4: dividing power consumption data acquired from a power consumption information acquisition system into a training group and an experimental group according to a ratio of 8:2, wherein the training group is used for training a support vector machine classifier, and the experimental group is used for classifying and identifying the load type of the experimental group;
and 5: inputting the training set into a support vector machine classifier for training;
step 6: and inputting the experimental group data into a support vector machine classifier to obtain a load type identification result of each experimental group.
2. The method of claim 1, wherein the load type identification based on the classification of the support vector machine comprises: the step 1 of constructing the multi-dimensional load characteristic classification indexes comprises the following steps:
the daily average load refers to the average of the load electricity consumption data 24 hours a day, and the monthly average load refers to the average of the load electricity consumption of 12 months a year, as shown in the following formula:
Figure FDA0002388424360000011
Figure FDA0002388424360000012
wherein, ηDay(s)Representing average daily load, ηYear of yearDenotes the average load per month, x1，x2，...，x24Load power consumption per hour for 24 hours a day, n1，n2，...，n12Represents the monthly load electricity usage of 12 months a year;
the average load value/the maximum load value is the load rate; taking the maximum value of daily load electricity consumption of 24 hours a day as a denominator and daily average load as a numerator, and constructing daily load rates and annual load rates of various load types:
Figure FDA0002388424360000021
Figure FDA0002388424360000022
wherein, deltaDay(s)、δYear of yearRespectively representing daily load rate and annual load rate;
the quotient of the difference between the maximum value of the load electricity consumption and the minimum value of the load electricity consumption divided by the maximum value of the load electricity consumption is the peak-valley difference rate:
Figure FDA0002388424360000023
Figure FDA0002388424360000024
in the above formula, λDay(s)、λYear of yearRespectively representing the daily peak-valley difference rate and the annual peak-valley difference rate;
the difference between the electricity usage in the last hour and the previous hour in 24 hours is the daily load change rate, and the difference between the average electricity usage in the last month and the previous month in 12 months is the annual load change rate:
γday(s)＝(x2-x1,x3-x2,...,x24-x23)
γYear of year＝(n2-n1,n3-n2,...,n12-n11)
In the above formula, γDay(s)、γYear of yearRespectively representing daily load change rate and annual load change;
the load classification index obtained finally is as follows:
Yday(s)＝(ηDay(s)，δDay(s)，λDay(s),x2-x1,x3-x2,...,x24-x23,di)
YYear of year＝(ηYear of year，δYear of year，λYear of year,n2-n1,n3-n2,...,n12-n11,di)
In the above formula, YDay(s)Indicates a daily load classification index, YYear of yearRepresenting annual load classification index, ηDay(s)Representing average daily load, ηYear of yearRepresenting the annual average load, delta representing the load rate, lambda representing the peak-to-valley difference rate, xiIndicating the amount of electricity used, diA label corresponding to the load type; wherein the dimension of the daily load classification indexThe degree is 27 and the dimension of the annual load classification index is 15.
3. The method of claim 1, wherein the load type identification based on the classification of the support vector machine comprises: step 2, establishing a support vector machine classifier model as follows:
the objective function is as follows:
Figure FDA0002388424360000025
the constraint conditions are as follows:
yi[(wxi)+b]≥1-ζi
ζi≥0
in the above formula, w and b are both elements in a linear function g (x) ═ wx + b, and represent the slope and intercept of the linear function, respectively; | w | non-woven phosphor2A two-norm representing the slope of a linear function; c represents a penalty factor; zetaiIs a relaxation variable; y isiRepresenting classification marks, and in a binary problem, 1 and-1 can be taken; n, n represents the number of sample points;
the support vector machine model is the minimum value of the objective function found under the above constraint conditions.
4. The method of claim 1, wherein the load type identification based on the classification of the support vector machine comprises: in step 3, the daily load data and the annual load data obtained in step 1 are subjected to standardization processing, and a maximum value element and a minimum value element in the load data are searched, so that the processed load data are in a uniform threshold range:
Figure FDA0002388424360000031
in the above formula, XiRepresents load data of node i, min { X }iDenotes the minimum value of the load data, max { X }iDenotes the maximum value in the load data, YiIndicating the number of loads after normalizationAccordingly.
CN202010110937.XA 2020-02-20 2020-02-20 Load type identification method based on support vector machine classification Pending CN111324790A (en)

Priority Applications (1)

Application Number Priority Date Filing Date Title
CN202010110937.XA CN111324790A (en) 2020-02-20 2020-02-20 Load type identification method based on support vector machine classification

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
CN202010110937.XA CN111324790A (en) 2020-02-20 2020-02-20 Load type identification method based on support vector machine classification

Publications (1)

Publication Number Publication Date
CN111324790A true CN111324790A (en) 2020-06-23

Family

ID=71172867

Family Applications (1)

Application Number Title Priority Date Filing Date
CN202010110937.XA Pending CN111324790A (en) 2020-02-20 2020-02-20 Load type identification method based on support vector machine classification

Country Status (1)

Country Link
CN (1) CN111324790A (en)

Cited By (3)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN112884042A (en) * 2021-02-23 2021-06-01 新疆大学 Power transmission and distribution line maximum load identification method based on relevance vector machine
CN116304358A (en) * 2023-05-17 2023-06-23 济南安迅科技有限公司 User data acquisition method
CN118033208A (en) * 2024-04-12 2024-05-14 江苏尚研电力科技有限公司 Intelligent air switch

Citations (4)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN105023092A (en) * 2015-06-24 2015-11-04 国网山东省电力公司电力科学研究院 Power load monitoring method based on electric quantity feature analysis
CN105023054A (en) * 2015-06-24 2015-11-04 国网山东省电力公司电力科学研究院 Power load analysis and predication method based on one-class support vector machine
CN109034241A (en) * 2018-07-24 2018-12-18 南京千智电气科技有限公司 Load cluster control method and system based on support vector machines
CN109598642A (en) * 2018-12-10 2019-04-09 国网山东省电力公司电力科学研究院 A kind of method for building up of accurate cutting load system interruptible load evaluation mechanism

Patent Citations (4)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN105023092A (en) * 2015-06-24 2015-11-04 国网山东省电力公司电力科学研究院 Power load monitoring method based on electric quantity feature analysis
CN105023054A (en) * 2015-06-24 2015-11-04 国网山东省电力公司电力科学研究院 Power load analysis and predication method based on one-class support vector machine
CN109034241A (en) * 2018-07-24 2018-12-18 南京千智电气科技有限公司 Load cluster control method and system based on support vector machines
CN109598642A (en) * 2018-12-10 2019-04-09 国网山东省电力公司电力科学研究院 A kind of method for building up of accurate cutting load system interruptible load evaluation mechanism

Non-Patent Citations (1)

* Cited by examiner, † Cited by third party
Title
刘子玥: "基于混沌理论和支持向量机的短期电力负荷预测", 《中国优秀博硕士学位论文全文数据库(硕士) 信息科技辑(月刊)》 *

Cited By (4)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN112884042A (en) * 2021-02-23 2021-06-01 新疆大学 Power transmission and distribution line maximum load identification method based on relevance vector machine
CN116304358A (en) * 2023-05-17 2023-06-23 济南安迅科技有限公司 User data acquisition method
CN116304358B (en) * 2023-05-17 2023-08-08 济南安迅科技有限公司 User data acquisition method
CN118033208A (en) * 2024-04-12 2024-05-14 江苏尚研电力科技有限公司 Intelligent air switch

Similar Documents

Publication Publication Date Title
US11043808B2 (en) Method for identifying pattern of load cycle
CN112561156A (en) Short-term power load prediction method based on user load mode classification
CN111724278A (en) Fine classification method and system for power multi-load users
CN111324642A (en) Model algorithm type selection and evaluation method for power grid big data analysis
CN109376772B (en) Power load combination prediction method based on neural network model
CN110781332A (en) Electric power resident user daily load curve clustering method based on composite clustering algorithm
CN111160401A (en) Abnormal electricity utilization judging method based on mean shift and XGboost
CN111324790A (en) Load type identification method based on support vector machine classification
CN112819299A (en) Differential K-means load clustering method based on center optimization
CN111428766B (en) Power consumption mode classification method for high-dimensional mass measurement data
CN108345908A (en) Sorting technique, sorting device and the storage medium of electric network data
CN111612228A (en) User electricity consumption behavior analysis method based on electricity consumption information
CN110188221A (en) A kind of load curve hierarchy clustering method based on shape distance
CN111626614A (en) User classification method based on electric charge recovery
CN108664653A (en) A kind of Medical Consumption client&#39;s automatic classification method based on K-means
CN117113126A (en) Industry electricity utilization characteristic analysis method based on improved clustering algorithm
Chen et al. A power line loss analysis method based on boost clustering
CN114611738A (en) Load prediction method based on user electricity consumption behavior analysis
CN114722098A (en) Typical load curve identification method based on normal cloud model and density clustering algorithm
CN110503145A (en) A kind of typical load curve acquisition methods based on k-shape cluster
CN113988161A (en) User electricity consumption behavior pattern recognition method
Wang et al. Analysis of user’s power consumption behavior based on k-means
CN111768066B (en) Park electric heating load coupling relation analysis method and device based on fusion characteristics
CN115687948A (en) Power special transformer user unsupervised classification method based on load curve
CN111695599B (en) Elastic identification method for user electricity load time

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
RJ01 Rejection of invention patent application after publication
RJ01 Rejection of invention patent application after publication

Application publication date: 20200623