CN111667135A - Load structure analysis method based on typical feature extraction - Google Patents
Load structure analysis method based on typical feature extraction Download PDFInfo
- Publication number
- CN111667135A CN111667135A CN202010216897.7A CN202010216897A CN111667135A CN 111667135 A CN111667135 A CN 111667135A CN 202010216897 A CN202010216897 A CN 202010216897A CN 111667135 A CN111667135 A CN 111667135A
- Authority
- CN
- China
- Prior art keywords
- load
- data
- classification
- cluster
- sample
- 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.)
- Granted
Links
- 238000000605 extraction Methods 0.000 title claims abstract description 19
- 238000003696 structure analysis method Methods 0.000 title claims abstract description 13
- 238000004458 analytical method Methods 0.000 claims abstract description 11
- 238000012545 processing Methods 0.000 claims abstract description 10
- 238000013211 curve analysis Methods 0.000 claims abstract description 5
- 238000000034 method Methods 0.000 claims description 27
- 238000004422 calculation algorithm Methods 0.000 claims description 19
- 230000006870 function Effects 0.000 claims description 12
- 230000002159 abnormal effect Effects 0.000 claims description 9
- 238000009825 accumulation Methods 0.000 claims description 9
- 239000011159 matrix material Substances 0.000 claims description 9
- 238000012937 correction Methods 0.000 claims description 6
- 238000007621 cluster analysis Methods 0.000 claims description 4
- 238000007635 classification algorithm Methods 0.000 claims description 3
- 230000005611 electricity Effects 0.000 claims description 3
- 238000010801 machine learning Methods 0.000 claims description 3
- 238000005192 partition Methods 0.000 claims description 3
- 230000035945 sensitivity Effects 0.000 claims description 3
- 238000011524 similarity measure Methods 0.000 claims description 3
- 238000007619 statistical method Methods 0.000 claims description 3
- 238000012549 training Methods 0.000 claims description 3
- 238000005259 measurement Methods 0.000 claims description 2
- 230000008859 change Effects 0.000 abstract description 7
- 238000002372 labelling Methods 0.000 abstract description 2
- 238000005516 engineering process Methods 0.000 abstract 1
- 238000005065 mining Methods 0.000 abstract 1
- 238000011161 development Methods 0.000 description 3
- 238000010586 diagram Methods 0.000 description 2
- 238000012986 modification Methods 0.000 description 2
- 230000004048 modification Effects 0.000 description 2
- 230000008569 process Effects 0.000 description 2
- 206010063385 Intellectualisation Diseases 0.000 description 1
- 238000009412 basement excavation Methods 0.000 description 1
- 230000006399 behavior Effects 0.000 description 1
- 230000007547 defect Effects 0.000 description 1
- 230000000694 effects Effects 0.000 description 1
- 238000003064 k means clustering Methods 0.000 description 1
- 238000004519 manufacturing process Methods 0.000 description 1
- 238000011160 research Methods 0.000 description 1
- 230000001932 seasonal effect Effects 0.000 description 1
- 238000005728 strengthening Methods 0.000 description 1
- 230000007704 transition Effects 0.000 description 1
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
- G06Q10/06—Resources, workflows, human or project management; Enterprise or organisation planning; Enterprise or organisation modelling
- G06Q10/063—Operations research, analysis or management
- G06Q10/0637—Strategic management or analysis, e.g. setting a goal or target of an organisation; Planning actions based on goals; Analysis or evaluation of effectiveness of goals
-
- 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
- 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
-
- 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
- Y04—INFORMATION OR COMMUNICATION TECHNOLOGIES HAVING AN IMPACT ON OTHER TECHNOLOGY AREAS
- Y04S—SYSTEMS INTEGRATING TECHNOLOGIES RELATED TO POWER NETWORK OPERATION, COMMUNICATION OR INFORMATION TECHNOLOGIES FOR IMPROVING THE ELECTRICAL POWER GENERATION, TRANSMISSION, DISTRIBUTION, MANAGEMENT OR USAGE, i.e. SMART GRIDS
- Y04S10/00—Systems supporting electrical power generation, transmission or distribution
- Y04S10/50—Systems or methods supporting the power network operation or management, involving a certain degree of interaction with the load-side end user applications
Landscapes
- Engineering & Computer Science (AREA)
- Business, Economics & Management (AREA)
- Theoretical Computer Science (AREA)
- Physics & Mathematics (AREA)
- Data Mining & Analysis (AREA)
- Human Resources & Organizations (AREA)
- Economics (AREA)
- General Physics & Mathematics (AREA)
- Strategic Management (AREA)
- Evolutionary Computation (AREA)
- Entrepreneurship & Innovation (AREA)
- Evolutionary Biology (AREA)
- Computer Vision & Pattern Recognition (AREA)
- General Engineering & Computer Science (AREA)
- Bioinformatics & Computational Biology (AREA)
- Health & Medical Sciences (AREA)
- Bioinformatics & Cheminformatics (AREA)
- Educational Administration (AREA)
- General Business, Economics & Management (AREA)
- Tourism & Hospitality (AREA)
- Artificial Intelligence (AREA)
- Marketing (AREA)
- Life Sciences & Earth Sciences (AREA)
- Primary Health Care (AREA)
- General Health & Medical Sciences (AREA)
- Water Supply & Treatment (AREA)
- Public Health (AREA)
- Development Economics (AREA)
- Probability & Statistics with Applications (AREA)
- Game Theory and Decision Science (AREA)
- Operations Research (AREA)
- Quality & Reliability (AREA)
- Information Retrieval, Db Structures And Fs Structures Therefor (AREA)
- Investigation Of Foundation Soil And Reinforcement Of Foundation Soil By Compacting Or Drainage (AREA)
Abstract
A load structure analysis method based on typical feature extraction comprises the following steps: s1, processing and repairing historical load data; s2, classifying the loads; and S3, typical curve analysis. The load structure analysis method is characterized by labeling and finely classifying urban loads by using national industry classification standards, then carrying out industry load characteristic analysis according to classification conditions, searching for load change rules, finally extracting a load characteristic curve by using a clustering technology, deeply mining internal characteristics of the loads, and finally obtaining an accurate relation between industry load characteristics and urban load characteristics.
Description
Technical Field
The invention relates to the technical field of power grid dispatching, in particular to a load structure analysis method based on typical feature extraction.
Background
Tianjin is developing relevant research on electric power markets, market members and transaction behaviors are complex and various after market reformation, and particularly after a spot market is opened, the price of the spot market is closely related to the load prediction accuracy. Therefore, with the continuous deepening of the power market innovation and the continuous strengthening of the lean management requirement of the power grid, higher requirements are put on the accuracy, the real-time performance, the reliability and the intellectualization of the load prediction.
On the other hand, with the development of society and economy, the national economy increases and enters a new normal state, the structural change occurs in the industrial development, the power utilization structure and the load characteristics are greatly changed, and the changes increase the uncertainty of the load demand change. In addition, with the influence of access such as industry transition, new energy access and novel load, the load characteristics gradually evolve, the referency of the historical load change rule of the whole network is gradually reduced, and the accurate prediction of the future load development trend is influenced to a great extent.
Aiming at various problems, the invention provides a load structure analysis method for classifying loads and extracting features aiming at regional load structure analysis, so as to analyze regional load structure constitution and excavation load characteristics.
Disclosure of Invention
The invention aims to overcome the defects of the prior art and provides a load structure analysis method based on typical feature extraction.
A load structure analysis method based on typical feature extraction comprises the following steps:
s1, historical load data processing and repairing:
when the data anomaly point is judged, a fuzzy clustering algorithm is adopted, and the objective function of the algorithm is as follows:
in the formula (I); x is the sample used, V is the clustering center matrix, T is the probability partition matrix, V is the fuzzy matrix, c is the clustering center, m is the weighted fuzzy index, and m > 1, n is the number of samples, uijIs the fuzzy value of the jth sample point in U belonging to the ith class, and Uij∈[0,1],(1≤i≤c,1≤j≤n),tijThe probability value that the jth sample point in T belongs to the ith class is defined, and q is more than 1;
and (3) performing completion correction after the type of the abnormal data is judged, and assuming that the jth data of the ith day in the 96-point load is missing or abnormal, and the ith data of the two days before and after the jth data is normal, wherein a null data completion formula is as follows:
Pij=(P(i-1)j+P(i+1)j)/2 (2)
when the load value is abnormal for more than two consecutive days, the Lagrange interpolation method is used for correction and completion, and the formula is as follows:
wherein li(P) is an interpolation basis function;
s2, load classification:
carrying out different classification strategies on the load according to the integrity degree of the load label;
s3, typical curve analysis:
standardizing the processed data to enable the processed data to become dimensionless data so as to avoid that attributes with larger magnitude caused by the magnitude difference occupy dominant positions, avoid that the magnitude difference causes slow iterative convergence speed, and reduce the influence of an algorithm depending on sample distance on the magnitude sensitivity of the data; and standardizing the selected load sample data by using Z-score, wherein the standardized formula is as follows:
P*=(P-Pμ)/Pσ(4)
wherein P isμIs the mean value of the load sample data, PσStandard deviation of load sample data;
the method adopts a K-means algorithm when extracting a typical curve, and the principle of the algorithm is that for a given sample set, samples are divided into K clusters according to the distance between the samples, so that points in the clusters are connected together as tightly as possible, and the maximum value of the distance between the clusters is taken;
the similarity measure function of the Kmeans cluster is expressed as follows:
where E is the sum of the squared errors of all samples; k is the number of clusters; ciIs the ith cluster; p is a sample point in the clustering space; mu.siIs the cluster center of the ith cluster; an elbow rule is adopted when the clustering cluster number k is determined, the intelligent determination of the cluster number is optimized, and typical clusters are extracted finallyAnd analyzing the load characteristics after the curve to obtain the load rule characteristics.
Further, the degree of completeness of the load label in step S2 includes the following three categories:
(1) when the load classification label is accurate and complete: in the case, a step-by-step accumulation method is adopted to analyze the load structure; establishing an industry standard classification logic relation structure, matching acquired 10kV and above user load data belonging type labels according to industry standard classification subclass names, classifying electricity consumption of the industry to which the subclasses belong according to the middle class to which the industry subclasses belong, the large class to which the middle class belongs, the door class to which the large class belongs and the door class belong, and performing statistical analysis on the load structure by adopting a method of accumulating step by step from bottom to top according to the belonged relation;
(2) when the load classification label part is missing: completing the load labels by using a typical machine learning classification algorithm, and then analyzing the load structure by adopting a step-by-step accumulation method according to the labels; under the condition that an industry classification label is partially lost, firstly, the load data with the complete label is used as a training set, the load characteristics are used as analysis objects, the label is labeled in a classification mode on the load with the lost label, and then a step-by-step accumulation method is adopted for analyzing the load structure;
(3) when all the load classification labels are missing: and carrying out standardization processing on the load, analyzing the load characteristics, carrying out cluster analysis on the load according to the load characteristics, and analyzing the load structure according to the cluster type.
Moreover, the step of extracting the typical curve by using the K-means algorithm in the step S3 includes the following steps:
(1) determining the number k of clustering clusters;
(2) initializing k clustering centers mu 1, … … and mu k;
(3) calculating the distance between the samples and the center, and distributing each sample to the nearest center;
(4) updating the center of each cluster according to the sample of each cluster;
(5) and (4) iterating the steps (3) to (4) until the similarity measurement function starts to converge.
The invention has the advantages and technical effects that:
according to the load structure analysis method based on the typical feature extraction, the history conforming data is processed and repaired through the step S1, the integrity and the reliability of the processed and repaired data can meet the requirement of urban load structure analysis, the internal correlation characteristics of mass data are utilized, the potential rule of the load is mined, and bad data are analyzed, identified and marked through the transverse and longitudinal comparisons of the load data; for different types of bad data, methods such as interpolation, virtual prediction and characteristic value extraction are used for repairing the bad data; the load is subjected to classification processing by step S2; finally, the typical curve is analyzed through the step S3, because of the total load of the region formed by the loads of various industries in the region, the step S3 needs to analyze the characteristics of different load curves exhibited by the industries due to different production rules, and after the loads are classified, the load curve of each class can be extracted, the characteristics of each class of loads can be analyzed, and the load change rule can be mined.
According to the load structure analysis method based on the typical feature extraction, the load structure is analyzed through load data restoration, load classification and typical curve analysis in sequence, an accurate load change rule is obtained, and the load structure analysis method is high in creativity.
Drawings
FIG. 1 is a flowchart of the data processing of step S1 according to the present invention;
FIG. 2 is a hierarchical relationship diagram of industry classification when the load classification label is accurate and complete in step S2 according to the present invention;
FIG. 3 is a schematic diagram of the missing tag load classification when the load classification tag is partially missing in step S2 according to the present invention;
FIG. 4 is a flowchart illustrating the K-means clustering process in step S3 according to the present invention;
FIG. 5 is a 96 point raw load graph for the second industry of the present invention;
fig. 6 is a graph showing typical curve results of the second industry according to the present invention.
Detailed Description
A load structure analysis method based on typical feature extraction comprises the following steps:
s1, historical load data processing and repairing:
when the data anomaly point is judged, a fuzzy clustering algorithm is adopted, and the objective function of the algorithm is as follows:
in the formula (I); x is the sample used, V is the clustering center matrix, T is the probability partition matrix, V is the fuzzy matrix, c is the clustering center, m is the weighted fuzzy index, and m > 1, n is the number of samples, uijIs the fuzzy value of the jth sample point in U belonging to the ith class, and Uij∈[0,1],(1≤i≤c,1≤j≤n),tijThe probability value that the jth sample point in T belongs to the ith class is defined, and q is more than 1;
and (3) performing completion correction after the type of the abnormal data is judged, and assuming that the jth data of the ith day in the 96-point load is missing or abnormal, and the ith data of the two days before and after the jth data is normal, wherein a null data completion formula is as follows:
Pij=(P(i-1)j+P(i+1)j)/2 (2)
when the load value is abnormal for more than two consecutive days, the Lagrange interpolation method is used for correction and completion, and the formula is as follows:
wherein li(P) is an interpolation basis function;
the specific data processing flow is shown in fig. 1;
s2, load classification:
carrying out different classification strategies on the load according to the integrity degree of the load label;
s3, typical curve analysis:
standardizing the processed data to enable the processed data to become dimensionless data so as to avoid that attributes with larger magnitude caused by the magnitude difference occupy dominant positions, avoid that the magnitude difference causes slow iterative convergence speed, and reduce the influence of an algorithm depending on sample distance on the magnitude sensitivity of the data; and standardizing the selected load sample data by using Z-score, wherein the standardized formula is as follows:
P*=(P-Pμ)/Pσ(4)
wherein P isμIs the mean value of the load sample data, PσStandard deviation of load sample data;
the method adopts a K-means algorithm when extracting a typical curve, and the principle of the algorithm is that for a given sample set, samples are divided into K clusters according to the distance between the samples, so that points in the clusters are connected together as tightly as possible, and the maximum value of the distance between the clusters is taken;
the similarity measure function of the Kmeans cluster is expressed as follows:
where E is the sum of the squared errors of all samples; k is the number of clusters; ciIs the ith cluster; p is a sample point in the clustering space; mu.siIs the cluster center of the ith cluster; and an elbow rule is adopted when the clustering cluster number k is determined, the intelligent determination of the cluster number is optimized, and finally a typical curve is extracted and then the load characteristic is analyzed to obtain the load rule characteristic.
Further, the degree of completeness of the load label in step S2 includes the following three categories:
(1) when the load classification label is accurate and complete: in the case, a step-by-step accumulation method is adopted to analyze the load structure; establishing an industry standard classification logic relation structure, matching acquired 10kV and above user load data belonging type labels according to industry standard classification subclass names, classifying electricity for the industry to which the class belongs according to the middle class to which the industry subclass belongs, the large class to which the middle class belongs, the door class to which the large class belongs and the door class belongs, and performing statistical analysis on the load structure according to the belonging relation by adopting a method of accumulating step by step from bottom to top, wherein the industry classification hierarchical relation is shown in figure 2;
(2) when the load classification label part is missing: completing the load labels by using a typical machine learning classification algorithm, and then analyzing the load structure by adopting a step-by-step accumulation method according to the labels; under the condition that part of industry classification labels are lost, firstly, load data with complete labels are used as a training set, load characteristics are used as analysis objects, classification labeling is carried out on the loads with the lost labels, then, a step-by-step accumulation method is adopted to carry out load structure analysis, and the analysis flow is shown in figure 3;
(3) when all the load classification labels are missing: and carrying out standardization processing on the load, analyzing the load characteristics, carrying out cluster analysis on the load according to the load characteristics, and analyzing the load structure according to the cluster type.
Moreover, the step of extracting the typical curve by using the K-means algorithm in the step S3 includes the following steps:
(1) determining the number k of clustering clusters;
(2) initializing k clustering centers mu 1, a.... mu.k;
(3) calculating the distance between the samples and the center, and distributing each sample to the nearest center;
(4) updating the center of each cluster according to the sample of each cluster;
(5) and (4) iterating the steps (3) to (4) until the similarity metric function starts to converge, wherein the clustering flow chart of the K-means algorithm is shown in FIG. 4.
In order to more clearly describe the specific embodiments of the present invention, an example is provided below:
examples are: take a second industry in a certain area as an example
Taking 96-point load data of a second industry in a certain area as an example, the feasibility of the load analysis method in the area is verified, and the load characteristics are discovered. The method comprises the following implementation steps:
1) standardizing the load data;
2) calculating a clustering category k by utilizing an elbow rule;
3) entering a clustering process to realize typical curve extraction;
4) analyzing a typical curve according to the clustering result;
5) the original load curve is shown in fig. 5.
The typical curve extraction result is shown in fig. 6, and it can be known from the typical curve extraction result that the two types of curves are distributed in different time periods, the type i curve is mainly distributed on holidays, the type ii curve is mainly distributed on normal working days and saturdays, the type i curve has a small load fluctuation range but a high change frequency, although the peak-valley characteristic is obvious, the peak-valley difference is small compared with the peak-valley difference of the type ii curve, the peak-valley characteristic of the type ii curve is more obvious, and impact load occurs. From the corresponding load curves, it can be seen that the two types of curves have no obvious seasonal characteristics, and also no obvious early peak and late peak.
The invention utilizes load data of a certain area to carry out load analysis, utilizes the load labels to classify the loads, then carries out cluster analysis on each type of load, analyzes the load characteristics of each type of load, excavates the load rule and completes the load analysis.
It will be understood that modifications and variations can be made by persons skilled in the art in light of the above teachings and all such modifications and variations are intended to be included within the scope of the invention as defined by the appended claims.
Claims (6)
1. A load structure analysis method based on typical feature extraction is characterized by comprising the following steps:
s1, historical load data processing and repairing:
when the data anomaly point is judged, a fuzzy clustering algorithm is adopted, and the objective function of the algorithm is as follows:
in the formula (I); x is the sample used, V is the cluster center matrix, T is the probability partition matrix, V is the fuzzy matrix, c is the cluster center, m is the weighted fuzzy index, and m is>1, n is the number of samples, uijIs the fuzzy value of the jth sample point in U belonging to the ith class, and Uij∈[0,1],tijFor the probability value of the jth sample point in T belonging to the ith class, q>1;
And (3) performing completion correction after judging the type of the detected abnormal data, wherein if the jth data of the ith day in the 96-point load is missing or abnormal, and the ith data of the two days before and after are normal, the null data completion formula is as follows:
Pij=(P(i-1)j+P(i+1)j)/2 (2)
when the load value is abnormal for more than two consecutive days, the Lagrange interpolation method is used for correction and completion, and the formula is as follows:
wherein li(P) is an interpolation basis function;
s2, load classification:
carrying out different classification strategies on the load according to the integrity degree of the load label;
s3, typical curve analysis:
standardizing the processed data to enable the processed data to become dimensionless data so as to avoid that attributes with larger magnitude caused by the magnitude difference occupy dominant positions, avoid that the magnitude difference causes slow iterative convergence speed, and reduce the influence of an algorithm depending on sample distance on the magnitude sensitivity of the data; and standardizing the selected load sample data by using Z-score, wherein the standardized formula is as follows:
P*=(P-Pμ)/Pσ(4)
whereinPμIs the mean value of the load sample data, PσStandard deviation of load sample data;
the method adopts a K-means algorithm when extracting a typical curve, and the principle of the algorithm is that for a given sample set, samples are divided into K clusters according to the distance between the samples, so that points in the clusters are connected together as tightly as possible, and the maximum value of the distance between the clusters is taken;
the similarity measure function of the Kmeans cluster is expressed as follows:
where E is the sum of the squared errors of all samples; k is the number of clusters; ciIs the ith cluster; p is a sample point in the clustering space; mu.siIs the cluster center of the ith cluster; and an elbow rule is adopted when the clustering cluster number k is determined, the intelligent determination of the cluster number is optimized, and finally a typical curve is extracted and then the load characteristic is analyzed to obtain the load rule characteristic.
2. The method for analyzing the load structure based on the characteristic feature extraction as claimed in claim 1, wherein: the integrity level of the load label in step S2 includes the following three states: the load classification label is accurate and complete, the load classification label is partially lost, and the load classification label is completely lost.
3. The method for analyzing the load structure based on the characteristic feature extraction as claimed in claim 2, wherein: when the load classification label in the step S2 is accurate and complete: in the case, a step-by-step accumulation method is adopted to analyze the load structure; establishing an industry standard classification logic relation structure, matching the type labels of the acquired 10kV and above user load data according to the industry standard classification subclass name, classifying the electricity consumption of the industry according to the middle class, the door class and the door class to which the industry subclass belongs, and performing statistical analysis on the load structure by adopting a method of accumulating step by step from bottom to top according to the affiliation.
4. The method for analyzing the load structure based on the characteristic feature extraction as claimed in claim 2, wherein: when the load classification tag portion in said step S2 is missing: completing the load labels by using a typical machine learning classification algorithm, and then analyzing the load structure by adopting a step-by-step accumulation method according to the labels; under the condition that part of industry classification labels are lost, firstly, load data with complete labels are used as a training set, load characteristics are used as analysis objects, the loads with the lost labels are labeled in a classification mode, and then a step-by-step accumulation method is adopted for analyzing a load structure.
5. The method for analyzing the load structure based on the characteristic feature extraction as claimed in claim 2, wherein: when all the load classification tags in the step S2 are missing: and carrying out standardization processing on the load, analyzing the load characteristics, carrying out cluster analysis on the load according to the load characteristics, and analyzing the load structure according to the cluster type.
6. The method for analyzing the load structure based on the characteristic feature extraction as claimed in claim 1, wherein: the step of extracting the typical curve by using the K-means algorithm in the step S3 comprises the following steps:
(1) determining the number k of clustering clusters;
(2) initializing k clustering centers mu 1, … … and mu k;
(3) calculating the distance between the samples and the center, and distributing each sample to the nearest center;
(4) updating the center of each cluster according to the sample of each cluster;
(5) and (4) iterating the steps (3) to (4) until the similarity measurement function starts to converge.
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN202010216897.7A CN111667135B (en) | 2020-03-25 | 2020-03-25 | Load structure analysis method based on typical feature extraction |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN202010216897.7A CN111667135B (en) | 2020-03-25 | 2020-03-25 | Load structure analysis method based on typical feature extraction |
Publications (2)
Publication Number | Publication Date |
---|---|
CN111667135A true CN111667135A (en) | 2020-09-15 |
CN111667135B CN111667135B (en) | 2023-07-28 |
Family
ID=72382556
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN202010216897.7A Active CN111667135B (en) | 2020-03-25 | 2020-03-25 | Load structure analysis method based on typical feature extraction |
Country Status (1)
Country | Link |
---|---|
CN (1) | CN111667135B (en) |
Cited By (4)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN112113316A (en) * | 2020-09-18 | 2020-12-22 | 国网辽宁省电力有限公司电力科学研究院 | Method for extracting air conditioner load |
CN112766733A (en) * | 2021-01-21 | 2021-05-07 | 山东大学 | Method and system for accelerating convergence of optimized scheduling algorithm by using improved K-means algorithm |
CN113033596A (en) * | 2020-12-30 | 2021-06-25 | 国网河南省电力公司南阳供电公司 | Refined identification method for user electricity consumption behavior category and typical electricity consumption mode |
CN113553041A (en) * | 2021-09-22 | 2021-10-26 | 武汉江民网安科技有限公司 | Method, apparatus and medium for generating function code formalized structure in binary program |
Citations (10)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
JP2002107389A (en) * | 2000-10-03 | 2002-04-10 | Itochu Corp | Electric power charge calculation method and electric power charge provisional calculation method |
CN103745417A (en) * | 2014-01-27 | 2014-04-23 | 国家电网公司 | Power grid load characteristic curve analysis method |
CN103793788A (en) * | 2014-01-27 | 2014-05-14 | 国家电网公司 | Orderly power utilization management method |
CN106055918A (en) * | 2016-07-26 | 2016-10-26 | 天津大学 | Power system load data identification and recovery method |
CN108376262A (en) * | 2018-02-23 | 2018-08-07 | 新疆大学 | A kind of analysis model construction method of wind power output typical characteristics |
CN109829610A (en) * | 2018-12-20 | 2019-05-31 | 国网江苏省电力有限公司苏州供电分公司 | It is a kind of meter and load temporal characteristics Distribution Network Reliability promoted Optimal Investment method |
CN110032944A (en) * | 2019-03-20 | 2019-07-19 | 国网电力科学研究院(武汉)能效测评有限公司 | A kind of electric load feature extracting method and system |
CN110188972A (en) * | 2019-03-12 | 2019-08-30 | 贵州电网有限责任公司 | A kind of non-special customer cut-in method of 10kV power distribution network |
CN110334726A (en) * | 2019-04-24 | 2019-10-15 | 华北电力大学 | A kind of identification of the electric load abnormal data based on Density Clustering and LSTM and restorative procedure |
CN110610121A (en) * | 2019-06-20 | 2019-12-24 | 国网重庆市电力公司 | Small-scale source load power abnormal data identification and restoration method based on curve clustering |
-
2020
- 2020-03-25 CN CN202010216897.7A patent/CN111667135B/en active Active
Patent Citations (10)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
JP2002107389A (en) * | 2000-10-03 | 2002-04-10 | Itochu Corp | Electric power charge calculation method and electric power charge provisional calculation method |
CN103745417A (en) * | 2014-01-27 | 2014-04-23 | 国家电网公司 | Power grid load characteristic curve analysis method |
CN103793788A (en) * | 2014-01-27 | 2014-05-14 | 国家电网公司 | Orderly power utilization management method |
CN106055918A (en) * | 2016-07-26 | 2016-10-26 | 天津大学 | Power system load data identification and recovery method |
CN108376262A (en) * | 2018-02-23 | 2018-08-07 | 新疆大学 | A kind of analysis model construction method of wind power output typical characteristics |
CN109829610A (en) * | 2018-12-20 | 2019-05-31 | 国网江苏省电力有限公司苏州供电分公司 | It is a kind of meter and load temporal characteristics Distribution Network Reliability promoted Optimal Investment method |
CN110188972A (en) * | 2019-03-12 | 2019-08-30 | 贵州电网有限责任公司 | A kind of non-special customer cut-in method of 10kV power distribution network |
CN110032944A (en) * | 2019-03-20 | 2019-07-19 | 国网电力科学研究院(武汉)能效测评有限公司 | A kind of electric load feature extracting method and system |
CN110334726A (en) * | 2019-04-24 | 2019-10-15 | 华北电力大学 | A kind of identification of the electric load abnormal data based on Density Clustering and LSTM and restorative procedure |
CN110610121A (en) * | 2019-06-20 | 2019-12-24 | 国网重庆市电力公司 | Small-scale source load power abnormal data identification and restoration method based on curve clustering |
Non-Patent Citations (2)
Title |
---|
胡阳春: "基于改进k 均值聚类算法的 电力负荷模式识别方法研究", 中国优秀硕士学位论文全文数据库 工程科技Ⅱ辑, no. 9, pages 042 - 193 * |
贾志达: "基于VMD的负荷分类方法研究", 中国优秀硕士学位论文全文数据库 工程科技Ⅱ辑, no. 8, pages 042 - 807 * |
Cited By (4)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN112113316A (en) * | 2020-09-18 | 2020-12-22 | 国网辽宁省电力有限公司电力科学研究院 | Method for extracting air conditioner load |
CN113033596A (en) * | 2020-12-30 | 2021-06-25 | 国网河南省电力公司南阳供电公司 | Refined identification method for user electricity consumption behavior category and typical electricity consumption mode |
CN112766733A (en) * | 2021-01-21 | 2021-05-07 | 山东大学 | Method and system for accelerating convergence of optimized scheduling algorithm by using improved K-means algorithm |
CN113553041A (en) * | 2021-09-22 | 2021-10-26 | 武汉江民网安科技有限公司 | Method, apparatus and medium for generating function code formalized structure in binary program |
Also Published As
Publication number | Publication date |
---|---|
CN111667135B (en) | 2023-07-28 |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
CN111667135B (en) | Load structure analysis method based on typical feature extraction | |
CN106845717B (en) | Energy efficiency evaluation method based on multi-model fusion strategy | |
CN110223196B (en) | Anti-electricity-stealing analysis method based on typical industry feature library and anti-electricity-stealing sample library | |
CN112561156A (en) | Short-term power load prediction method based on user load mode classification | |
CN112686286B (en) | Building operation energy consumption abnormality identification method, system and computer readable storage medium | |
CN111367801B (en) | Data transformation method for cross-company software defect prediction | |
CN114676742A (en) | Power grid abnormal electricity utilization detection method based on attention mechanism and residual error network | |
CN115358347B (en) | Method for predicting remaining life of intelligent electric meter under different subsystems | |
CN107862406A (en) | Using deep learning and the method for the primary equipment risk profile for improving Apriori algorithm synthesis | |
CN113408659A (en) | Building energy consumption integrated analysis method based on data mining | |
CN110598747B (en) | Road classification method based on self-adaptive K-means clustering algorithm | |
CN114707422A (en) | Intelligent power check method based on load prediction | |
CN110163437B (en) | Day-ahead photovoltaic power generation power prediction method based on DPK-means | |
CN111861785A (en) | Special transformer industry fault identification method based on power utilization characteristics and outlier detection | |
CN114372835B (en) | Comprehensive energy service potential customer identification method, system and computer equipment | |
CN111737993A (en) | Method for extracting health state of equipment from fault defect text of power distribution network equipment | |
CN116070458A (en) | New wind power plant scene generation method based on RAC-GAN | |
CN116470491A (en) | Photovoltaic power probability prediction method and system based on copula function | |
CN110634294A (en) | Time-space traffic flow prediction method driven by reinforced hierarchical learning | |
CN114638284A (en) | Power utilization behavior characterization method considering external influence factors | |
CN113705229A (en) | High-risk enterprise illegal production identification method based on electric power big data | |
CN113435494A (en) | Low-voltage resident user abnormal electricity utilization identification method and simulation system | |
Peiyi et al. | Analysis and research on enterprise resumption of work and production based on K-means clustering | |
Liu et al. | Short-term Load Forecasting Approach with SVM and Similar Days Based on United Data Mining Technology | |
CN118133051B (en) | Construction method and device of element evaluation model |
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 | ||
GR01 | Patent grant | ||
GR01 | Patent grant |