CN115358784A - Method for extracting typical power consumption modes in distributed mode and related equipment - Google Patents
Method for extracting typical power consumption modes in distributed mode and related equipment Download PDFInfo
- Publication number
- CN115358784A CN115358784A CN202211010754.6A CN202211010754A CN115358784A CN 115358784 A CN115358784 A CN 115358784A CN 202211010754 A CN202211010754 A CN 202211010754A CN 115358784 A CN115358784 A CN 115358784A
- Authority
- CN
- China
- Prior art keywords
- load
- industrial
- power
- data
- typical
- 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
- 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
- G06Q30/00—Commerce
- G06Q30/02—Marketing; Price estimation or determination; Fundraising
- G06Q30/0201—Market modelling; Market analysis; Collecting market data
-
- 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—Systems or methods specially adapted for specific business sectors, e.g. utilities or tourism
- G06Q50/06—Electricity, gas 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
Abstract
The invention provides a method for extracting a distributed typical power consumption mode and related equipment, which relate to the field of industrial power big data processing and analysis and comprise the following steps: acquiring power load data of a plurality of industrial users in a target area; respectively processing the power load data of the industrial users aiming at the power load data of each industrial user to obtain a power utilization state matrix of the industrial users; generating a load curve set cluster by using the power utilization state matrixes of a plurality of industrial users, and classifying load curves in the load curve set cluster in the same state and in the same time length into one class; respectively extracting typical power consumption modes aiming at each type of load curve to obtain a typical power consumption mode set cluster corresponding to each industrial user; fusing typical electricity utilization mode set clusters of a plurality of industrial users to obtain a typical electricity utilization mode; the accuracy and the representativeness of typical electricity utilization mode extraction can be effectively improved, and the efficiency of mode extraction and data analysis is also improved.
Description
Technical Field
The invention relates to the field of industrial power big data processing and analysis, in particular to a method for extracting a typical power consumption mode in a distributed mode and related equipment.
Background
Owing to the rapid development of industrial manufacturing and new generation of high and new information technology industry, the electrification level of each industry is continuously improved, so that the demand of each industry on electric power energy is increased day by day. According to statistics of the national energy agency, the total electricity consumption of China in 2021 is 83128 hundred million kilowatt hours, wherein the industrial electricity consumption is 55090 million kilowatt hours, 66.27 percent of the total electricity consumption accounts for the second industry, and the proportion of the second industry electricity consumption is up to 98.15 percent. It follows that industrial power loads account for a very large proportion of the overall power demand of the whole society, especially in some areas of heavy industry development.
Industrial loads are already important components of power loads in China, and with the rapid development of smart power grids and the continuous progress of cash smart meter technology and power consumption information acquisition technology, more and more smart meters are deployed to each household and each industry. The frequent collection of the power load data of the users enables the data volume of the power load to present a remarkable climbing attitude. Meanwhile, under the influence of market environment and external factors, the electricity utilization behavior mode of industrial users has the characteristics of high uncertainty, high complexity, diversity and the like, so that the potential law of the electricity utilization of the users is difficult to find, and great challenges are brought to further data analysis and decision support. To some extent, it may even become one of the important factors hindering the future development of smart grids and smart power systems.
The data-driven extraction of the power utilization mode of the user is a key basis for supporting the mining of the application value of the big electric power data, the power utilization behavior habits of industrial users can be mined through the scientific and reasonable extraction of the power utilization mode, the potential rules of the industrial users can be found, and the efficiency of data analysis is further improved. On one hand, the power supplier can provide flexible and personalized accurate user service for the industrial users according to the power utilization modes of the industrial users, and can also generate positive promotion effect for intelligent resource management and real-time scheduling of the intelligent power system so as to improve the operation efficiency of the intelligent power system; on the other hand, industrial users can know the production mode and the power utilization condition of a factory according to the preference of own power utilization behaviors, and effective information is provided for further specifying a scientific and reasonable production plan and an energy-saving and emission-reducing strategy.
However, the existing user typical electricity pattern extraction method is facing the following challenges:
first, typical power usage pattern extraction at the industrial user level is rarely considered;
secondly, the drift effect of the electricity utilization behavior of the user is ignored in the mode extraction process, so that most typical electricity utilization modes have certain time-space similarity, and even the similar electricity utilization modes can be represented by the section curves of the typical electricity utilization modes with the drift effect;
third, the system is mainly designed for single user development, is rarely triggered from multiple user levels, and extracts typical power consumption patterns with common representativeness. Meanwhile, in the multi-user electricity consumption behavior pattern extraction process, the challenges of large data volume, low calculation efficiency and the like are faced.
In view of the above, a new typical power pattern extraction method is needed to solve the various challenges faced in the prior art.
Disclosure of Invention
The invention provides a method for extracting a typical power consumption mode in a distributed manner and related equipment, and aims to improve the accuracy of power consumption mode extraction of industrial users and the efficiency of data calculation and analysis in consideration of industrial users.
In order to achieve the above object, the present invention provides a method for distributively extracting a typical electricity usage pattern, including:
step 1, collecting power load data of a plurality of industrial users in a target area;
step 2, processing the power load data of the industrial users respectively aiming at the power load data of each industrial user to obtain a power utilization state matrix of the industrial users;
step 3, generating a load curve set cluster by using the power utilization state matrixes of a plurality of industrial users, and classifying load curves in the load curve set cluster in the same state and in the same time length into one class;
step 4, respectively extracting typical power consumption modes aiming at each type of load curve to obtain a typical power consumption mode set cluster corresponding to each industrial user;
and 5, fusing the typical electricity utilization mode set clusters of the industrial users to obtain a typical electricity utilization mode.
Further, step 1 comprises:
respectively collecting power load data of N days from intelligent electric meters of industrial users aiming at each industrial user in a target area; the sampling frequency of the power load data sampling every day is T;
preprocessing the power load data of N days to obtain power load data X;
normalizing the power load data X to obtain normalized power load data X';
wherein, X = [ X ] 1 ，x 2 ，…，x d ，…，x N ]And x is d ＝[x d，1 ，x d，2 ，…，x d，t ，…，x d，T ]，x d，t Representing the daily load data collected by the industrial user in the t time period on the d day, wherein the value range of t is [1,T ]]。
Further, step 2 comprises:
respectively processing the normalized power load data X' aiming at the power load data of each industrial user to obtain the power load mean value mu of the industrial user;
dividing the daily load data into load states according to the power load mean value mu to obtain the division results of the load states, namely a ground state load and an excited state load;
and acquiring a load state matrix S of the industrial user according to the power load mean value mu and the division result.
Further, step 2 further comprises:
mean value of the electrical load mu of
Wherein N represents the number of sampling days, T represents the frequency of sampling the power load data per day,
x′ d，t represents the normalized power load value, x, of the t-th time period on the d-th day d，t The data of the daily load collected by the industrial user in the t period on the d day are represented, max (X) represents the maximum load value in the historical power load data of the industrial user, and min (X) represents the minimum load value in the historical power load data of the industrial user.
By passing
Dividing the load state of the power load data into a ground state load and an excited state load when S d，t When "= 0, S is the ground state load d，t And when =1, the load state is the excited load.
Based on the power load mean value mu and the load state of the industrial users, obtaining a load state matrix S of each industrial user as
Wherein S is d，t Represents the state of the electrical load of the industrial user at the time of day d, t, and S d，t E {0,1}, where N represents the number of sampling days and T represents the frequency of power load data sampling per day.
Further, step 3 comprises:
given the set of load states at day-d time-t in the load state matrix S of each industrial user as { S } d，t } t＝1，2，…T ；
When t =1 and S d，t If =0, the time t is the trigger time of the ground state load, and s is the time d，t When =1, the time t is taken as the baseThe end time of the dynamic load;
when t =1 and S d，t If =1, the time t is the trigger time of the excited load, and s is the time s d，t And if =0, the time t is set as the end time of the excited-state load.
Respectively aiming at the load state of each industrial user, generating an ith load curve Lbas of the industrial user about the ground state load in the d day according to the triggering time and the ending time of the load state d，i And a jth load curve Lexc for the excited state load d，j 。
Load curve Lbas d，i And load curve Lexc d，j Respectively packaging to obtain a Load curve set Load _ bas of the industrial user about the ground state Load on the day d d And a Load curve set Load _ exc with respect to the excited-state Load d 。
Load curve is collected into Load _ bas d And Load curve set Load _ exc d Packaging to obtain a Load curve set cluster Load _ model of the industrial user under all Load states on the d day d Collecting Load curve set cluster Load _ model under all Load states for N days d And packaging to obtain the Load curve set cluster Load _ model of the industrial user.
And classifying the Load curve sets with the same duration under the same Load state in the Load curve set cluster Load _ model into one type to obtain two types of Load curve set clusters, wherein the two types of Load curve set clusters are respectively a Load curve set cluster Lenload _ bas of an industrial user under the ground state Load and a Load curve set cluster Lenload _ exc of the industrial user under the excited state Load.
Further, step 4 comprises:
and respectively aiming at each type of load curve, acquiring the distance between any two load curves X and Y with the same time length as DTW (X, Y) by using a DTW-based distance measurement method.
Clustering load curve set cluster LenLoad _ bas and load curve set cluster LenLoad _ exc based on load curve AP clustering algorithm of DTW (X, Y), and extracting typical load conditions of the pth industrial userElectricity consumption mode set RepLoad _ Model p 。
Calling a textFile () function in Spark context, creating RDD (resource description device) related to the power load data of each industrial user, and performing mode extraction from the RDD through a Spark distributed parallel computing framework to obtain a typical power consumption mode set cluster RePlad _ Model corresponding to each industrial user.
Further, step 5 is followed by:
classifying a plurality of industrial users according to a typical power consumption mode to obtain an industrial user cluster;
and carrying out big data analysis and visualization processing on the typical power consumption mode and the industrial user cluster to obtain a visualized data analysis report.
The invention also provides a system for extracting the typical power consumption mode in a distributed manner, which comprises the following steps:
the data acquisition module is used for acquiring power load data of a plurality of industrial users in a target area;
the data processing module is used for processing the power load data of the industrial users respectively aiming at the power load data of each industrial user to obtain a power utilization state matrix of the industrial users;
the data classification module is used for generating a load curve set cluster by utilizing the power utilization state matrixes of a plurality of industrial users and classifying the load curves in the load curve set cluster in the same state and in the same time length into one class;
the data extraction module is used for extracting typical power consumption modes aiming at each type of load curve respectively to obtain a typical power consumption mode set cluster corresponding to each industrial user;
and the data fusion module is used for fusing the typical electricity utilization mode set clusters of the industrial users to obtain typical electricity utilization modes.
The present invention also provides a computer-readable medium comprising: for storing a computer program for implementing the above-described method for distributed extraction of typical electricity usage patterns by executing the computer program.
The invention also provides a device for extracting the typical power consumption mode in a distributed manner, which is used for realizing the method for extracting the typical power consumption mode in the distributed manner, and is characterized by comprising the following steps: a memory and a processor;
the memory is used for storing a computer program;
the processor is for executing the computer program stored by the memory.
The scheme of the invention has the following beneficial effects: the method and the system have the advantages that in consideration of industrial user level, the power load data of a plurality of industrial users are collected, the power load data of the industrial users are processed, extracted and fused to obtain the typical power utilization mode, the industrial users are classified according to the typical power utilization mode to obtain the industrial user cluster, the accuracy and the representativeness of the typical power utilization mode extraction can be effectively improved, the efficiency of the mode extraction and the data analysis can be improved, the large data analysis and the visualization processing are carried out on the typical power utilization mode and the industrial user cluster to obtain the visualized data analysis report to be recommended to a power supplier or the industrial user, and strategies such as decision support, demand response, energy conservation, emission reduction and the like can be realized.
Other advantages of the present invention will be described in detail in the detailed description that follows.
Drawings
FIG. 1 is a flow chart of an embodiment of the present invention.
Detailed Description
In order to make the technical problems, technical solutions and advantages of the present invention more apparent, the following detailed description is given with reference to the accompanying drawings and specific embodiments. It is to be understood that the embodiments described are only a few embodiments of the present invention, and not all embodiments. All other embodiments, which can be derived by a person skilled in the art from the embodiments given herein without making any creative effort, shall fall within the protection scope of the present invention.
In the description of the present invention, it should be noted that the terms "center", "upper", "lower", "left", "right", "vertical", "horizontal", "inner", "outer", etc., indicate orientations or positional relationships based on the orientations or positional relationships shown in the drawings, and are only for convenience of description and simplicity of description, but do not indicate or imply that the device or element being referred to must have a particular orientation, be constructed and operated in a particular orientation, and thus, should not be construed as limiting the present invention. Furthermore, the terms "first," "second," and "third" are used for descriptive purposes only and are not to be construed as indicating or implying relative importance.
In the description of the present invention, it should be noted that, unless otherwise explicitly specified or limited, the terms "mounted", "connected" and "connected" are to be construed broadly, e.g., as being either a locked connection, a detachable connection, or an integral connection; can be mechanically or electrically connected; they may be connected directly or indirectly through intervening media, or they may be interconnected between two elements. The specific meanings of the above terms in the present invention can be understood in specific cases to those skilled in the art.
In addition, the technical features involved in the different embodiments of the present invention described below may be combined with each other as long as they do not conflict with each other.
The invention provides a method for extracting typical power consumption modes in a distributed mode and related equipment aiming at the existing problems.
As shown in fig. 1, an embodiment of the present invention provides a method for extracting typical power consumption patterns in a distributed manner, which is oriented to industrial power big data, and includes:
step 1, collecting power load data of a plurality of industrial users in a target area;
step 2, processing the power load data of the industrial users respectively aiming at the power load data of each industrial user to obtain a power utilization state matrix of the industrial users;
step 3, generating a load curve set cluster by using the power utilization state matrixes of a plurality of industrial users, and classifying load curves in the load curve set cluster in the same state and in the same time length into one class;
step 4, respectively extracting typical power consumption modes aiming at each type of load curve to obtain a typical power consumption mode set cluster corresponding to each industrial user;
and 5, fusing the typical electricity utilization mode set clusters of the industrial users to obtain a typical electricity utilization mode.
Specifically, step 1 includes:
respectively collecting power load data of N days from intelligent electric meters of industrial users aiming at each industrial user in a target area; the sampling frequency of the power load data sampling every day is T;
preprocessing the power load data of N days to obtain power load data X;
wherein, X = [ X ] 1 ，x 2 ，…，x d ，…，x N ]And x is d ＝[x d，1 ，x d，2 ，…，x d，t ，…，x d，T ]，x d，t Representing the daily load data collected by the industrial user in the t time period on the d day, wherein the value range of t is [1,T ]]；
And carrying out maximum-minimum normalized processing on the power load data X to obtain normalized power load data X'.
In the embodiment of the invention, the industrial users in the target area refer to all industrial users with large power consumption and obvious power load consumption in a certain industrial park or a certain range, and can have the same power consumption behavior mode or multiple different power consumption behavior modes according to the conditions that the industrial users in the target area have the same power consumption behavior mode or the different power consumption behavior modes
Carrying out maximum-minimum normalization processing on the power load data X, acquiring normalized industrial user power load data X ', and transmitting the normalized power load data X ' to a distributed data storage system for storage, wherein X ' d，t And the normalized power load value of the t period on the day d is represented, max (X) represents the maximum load value in the historical power load data of the industrial user, and min (X) represents the minimum load value in the historical power load data of the industrial user.
Specifically, step 2 includes:
respectively processing the normalized power load data set X' aiming at the power load data of each industrial user to obtain the power load mean value mu of the industrial user;
dividing the daily load data into load states according to the power load mean value mu to obtain the division results of the load states, namely a ground state load and an excited state load;
and acquiring a load state matrix S of the industrial user according to the power load mean value mu and the division result.
Specifically, according to the power load data of a plurality of industrial users, the power load mean value μ of each industrial user is obtained, the load level of the industrial user is judged by taking the power load mean value μ as a standard, and the calculation formula of the power load mean value μ is as follows:
the embodiment of the invention is triggered from the aspect of energy consumption in the actual industrial power utilization process by
Dividing the load state of the power load data into a ground state load and an excited state load, wherein the ground state load is the basic load energy consumption generated when a factory or a workshop uses some low-energy electric equipment under the condition that no operation is carried out, and the excited state load is the high load energy consumption generated when the factory or the workshop uses high-energy electric equipment under the condition that the operation is carried out; when S is d，t When "= 0, S is the ground state load d，t When =1, the load state is the excited load;
based on the power load mean value mu and the power utilization state of the industrial users, obtaining a load state matrix S of each industrial user as
Wherein S is d，t Represents the state of the electrical load of the industrial user at the time of day d, t, and S d，t E {0,1}, where N represents the number of sampling days and T represents the frequency of sampling power load data per day.
Specifically, step 3 includes:
given the load state set of the day d in the electricity utilization state matrix S of each industrial user as { S d，t } t＝1，2，…T ；
When t =1 and S d，t When =0 or when t ≠ 1 and S d，t ＝0、S d，t-1 If =1, the time t is the trigger time of the ground state load, and S is the time d，t If =1, the time t is set as the end time of the ground state load;
when t =1 and S d，t When =1 or when t ≠ 1 and S d，t ＝1、S d，t-1 If =0, the time t is the trigger time of the excited load, and when S is d，t If =0, time t is set as the end time of the excited load.
Respectively aiming at the load state of each industrial user, generating an ith load curve Lbas of the industrial user about the ground state load on the d day according to the triggering time and the ending time of the load state d，i And a jth loading curve Lexc for the excited state loading d，j ；
Load curve Lbas d，i And load curve Lexc d，j Respectively packaging to obtain a Load curve set Load _ bas of the industrial user about the ground state Load on the day d d And a Load curve set Load _ exc with respect to the excited-state Load d ；
Load curve is collected into Load _ bas d And Load curve set Load _ exc d Packaging to obtain a Load curve set cluster Load _ model of the industrial user under all Load states on the d day d Collecting Load curve set cluster Load _ model under all Load states for N days d And packaging to obtain the Load curve set cluster Load _ model of the industrial user.
In the embodiment of the invention, the trigger time and the end time of the load state can be [ ts1, te1 ]]Time sum [ ts2, te2]The load curves at the moment are respectively identified as a ground state load and an excited state load, and the ground state load and the excited state load are taken as the ith load curve Lbas of the industrial user relative to the ground state load on the d day d，i And the j-th electronic die Lexc related to the excited state load d，j Respectively, as follows:
according to the identification method of the Load state, a plurality of Load curves of the industrial user in different Load states in the d-th day are obtained, and the obtained Load curves are packaged into a Load curve set cluster Load _ bas of the industrial user in the d-th day in relation to the ground state Load according to the different Load states d ＝{Lbas d，1 ，Lbas d，2 … and Load curve set cluster Load _ exc for excited state Load d ＝{Lexc d，1 ，Lexc d，2 ，…}。
Based on the identification method of the Load state and the packaging strategy of the Load curve set, the Load curve set cluster Load _ bas of the industrial user about the ground state Load d ＝{Lbas d，1 ，Lbas d，2 … and Load curve for excited state Load Load _ exc d ＝{Lexc d，1 ，Lexc d，2 …) to obtain a Load curve set cluster Load _ model of the industrial user under all Load states on day d d ＝ {Load_bas d ，Load_exc d Collecting Load curve set cluster Load _ model under all Load states for N days d Performing further packaging, and acquiring a Load curve set cluster of the industrial user Load _ model = { Load _ bas, load _ exc }, and Load _ bas = { Load _ bas }, wherein the Load _ bas = 1 ，…，Load_bas N }、 Load_exc＝{Load_exc 1 ，…，Load_exc N }. And classifying the Load curve sets with the same duration under the same Load state in the Load curve set cluster Load _ model into one class to obtain a Load curve set cluster LenLoad _ bas of the industrial user under the ground state Load and a Load curve set cluster LenLoad _ exc of the industrial user under the excited state Load.
Specifically, step 4 includes:
respectively aiming at each type of load curve, acquiring the distance between any two load curves X and Y with the same time length as DTW (X, Y) by using a DTW-based distance measurement method;
clustering load curve set LenLoad _ bas and load curve set LenLoad _ exc based on a load curve AP clustering algorithm of DTW (X, Y), and extracting a typical power consumption mode set RepLoad _ Model of the pth industrial user in different load states p ；
Calling a textFile () function in Spark context, creating RDD (resource description device) related to the power load data of each industrial user, and performing mode extraction from the RDD through a Spark distributed parallel computing framework to obtain a typical power consumption mode set cluster RePlad _ Model corresponding to each industrial user.
In the embodiment of the invention, the Load curves with the same duration under the same Load state in the Load _ model of the Load curve set cluster are classified into one class, so that the Load curve set LenLoad _ bas of the industrial user under the ground state Load and related to the Load duration is obtained i And a set of load curves LenLoad _ exc for load duration under excited state load i The calculation formula is as follows:
wherein LenLoad _ bas i And LenLoad _ exc i Respectively representing reference to the ground stateLoad curve set clusters of the load and the excited state load, | · | represents the sum of the number of elements in the set clusters, namely the cardinality.
Based on the load curve sets classified with respect to duration under different load states, acquiring a load curve set cluster LenLoad _ bas = { LenLoad _ bas = of the industrial user under the ground state load 1 ，LenLoad_bas 2 … and load curve set under excited-state loading, lenLoad _ exc = { LenLoad _ exc = LenLoad _ exc } 1 ，LenLoad_exc 2 ，…}。
Specifically, in the embodiment of the present invention, the load curve clustering algorithm based on dynamic time warping DTW is an AP clustering algorithm based on dynamic time warping DTW, and includes the following steps:
step 41, obtaining the LenLoad _ bas of the electricity utilization mode set of the industrial user based on the distance measurement method of DTW i A similarity matrix S;
step 42, setting an initial reference degree P and an iteration number M, generally setting the reference degree P = mean (S), and setting the iteration number M =50;
step 43, acquiring a LenLoad _ bas electricity consumption mode set of the industrial user i An attraction degree matrix R and an attribution degree matrix A;
step 44, iteratively updating the attraction degree matrix R and the attribution degree matrix A until the clustering center does not change in a plurality of iterative processes or reaches a set iterative number;
step 45, outputting the clustering centers and the power consumption patterns included by each clustering center, and adding the clustering centers of each clustering to the typical power consumption pattern RepLoad _ Model of the p-th industrial user p Until all power consumption pattern set clusters LenLoad _ bas and LenLoad _ exc of the industrial user are clustered.
Describing the similarity degree between the load curves having the same time length in the same state using the DTW-based distance measurement method, for any two load curves X and Y having the same time length, a distance measurement formula DTW (X, Y) with respect to the load curves X and Y may be expressed as follows:
DTW(X，Y)＝minC(X，Y)
wherein the content of the first and second substances,
representing the euclidean distance between load points x and y.
Extracting a load curve RePlad _ Model of the pth industrial user under different load states according to a load curve assembly cluster LenLoad _ bas under a ground state load, a load curve assembly cluster LenLoad _ exc under an excited state load and a load curve clustering algorithm based on DTW (X, Y) p To characterize similar power usage patterns.
In the embodiment of the invention, a Spark Application running environment is constructed, spark context is started, a textFile () function in Spark context is called to create load data RDD for the electricity load data of each industrial user stored in a distributed data storage system, and the related algorithms such as the load state identification method, the load curve extraction algorithm and the load curve AP clustering algorithm based on DTW (dynamic time warping) proposed in the foregoing are packaged into a machine learning algorithm library MLlib in Spark; the SparkContext sends a request to the resource manager to apply for extracting the typical load curve in the execution Executor; based on Spark distributed parallel computing framework, an executive Executor calls a related algorithm for typical load curve extraction packaged in an MLlib algorithm library, a distributed computing mode is adopted for load data RDD files of a plurality of users, the typical load curves of the plurality of industrial users are extracted and subjected to cluster analysis, and therefore typical electricity utilization mode set cluster RePad _ Model = { RePad _ Model = for the plurality of industrial users is obtained 1 ，…，RepLoad_Model p …, wherein RepLoad _ Model p Represented as a collection of typical power usage patterns for the pth industrial user.
The load data RDD is an elastic distributed data set format and is mainly used for distributed parallel operation of Spark platforms. RDD elastic Distributed data sets (resource Distributed data sets), a read-only object set Distributed in a cluster, are composed of a plurality of partitions, and are constructed through conversion operation, after failure, the read-only object set is automatically reconstructed to realize elasticity, the RDD can be regarded as an object of Spark, and the RDD is stored in a memory or a disk. The memory distribution data set started by the spark distributed parallel computing framework can provide interactive query, optimize iterative workload and be used for constructing a large-scale and low-delay data analysis application program.
Specifically, step 5 includes:
and obtaining representative typical electricity consumption mode RepLoad _ Agg from the typical electricity consumption mode set cluster RepLoad _ Model to represent the electricity consumption behavior mode of the whole industrial users, and feeding back the representative typical electricity consumption mode RepLoad _ Agg to the distributed data storage system.
Specifically, step 5 is followed by:
classifying the industrial users based on representative typical electricity consumption mode RepLoad _ Agg to obtain an industrial user cluster with a similar electricity consumption behavior mode, and feeding the industrial user cluster and the industrial user cluster back to the distributed data storage system;
and carrying out big data analysis and visual processing on a typical power utilization mode and an industrial user cluster in the distributed data storage system to obtain a visual data analysis report.
In the embodiment of the invention, technologies such as big data analysis and visualization are utilized to process the typical power utilization mode and the industrial user cluster fed back to the distributed data storage system to obtain the visualized data analysis report, and the visualized data analysis report is transmitted to the power supplier and the industrial user to assist the power supplier or the industrial user in making a power strategy and an energy-saving emission-reducing scheme, so that decision support and demand response are realized.
The embodiment of the present invention further provides a system for extracting a typical power consumption pattern in a distributed manner, including:
the data acquisition module is used for acquiring power load data of a plurality of industrial users in a target area;
the data processing module is used for processing the power load data of the industrial users aiming at the power load data of each industrial user to obtain a power utilization state matrix of the industrial users;
the data classification module is used for generating a load curve set cluster by utilizing the power utilization state matrixes of a plurality of industrial users and classifying the load curves in the load curve set cluster in the same state and in the same time length into one class;
the data extraction module is used for extracting typical power consumption modes aiming at each type of load curve respectively to obtain a typical power consumption mode set cluster corresponding to each industrial user;
and the data fusion module is used for fusing the typical electricity utilization mode set clusters of a plurality of industrial users to obtain a typical electricity utilization mode.
Embodiments of the present invention also provide a computer-readable medium, comprising: for storing a computer program for implementing the above-described method for distributed extraction of typical electricity usage patterns by executing the computer program.
The embodiment of the present invention further provides a device for extracting a typical power consumption pattern in a distributed manner, which is used for implementing the method for extracting a typical power consumption pattern in a distributed manner, and is characterized by including: a memory and a processor;
the memory is used for storing a computer program;
the processor is for executing the computer program stored by the memory.
In summary, in the embodiment of the present invention, the power load data of the multiple industrial users is collected first, the power load data of the multiple industrial users is processed, extracted, and fused to obtain the typical power consumption mode, the multiple industrial users are classified according to the typical power consumption mode to obtain the industrial user cluster, the accuracy and the representativeness of the typical power consumption mode extraction can be effectively improved, the efficiency of the mode extraction and the data analysis can be further improved, the large data analysis and the visualization processing are performed on the typical power consumption mode and the industrial user cluster to obtain the visualized data analysis report which is recommended to the power supplier or the industrial user, and the strategies such as decision support, demand response, energy saving, emission reduction, and the like can be implemented.
While the foregoing is directed to the preferred embodiment of the present invention, it will be understood by those skilled in the art that various changes and modifications may be made without departing from the spirit and scope of the invention as defined in the appended claims.
Claims (10)
1. A method for distributively extracting a typical electricity consumption mode is characterized by comprising the following steps:
step 1, collecting power load data of a plurality of industrial users in a target area;
step 2, processing the power load data of the industrial users respectively aiming at the power load data of each industrial user to obtain a power utilization state matrix of the industrial users;
step 3, generating a load curve set cluster by using the electricity utilization state matrixes of the industrial users, and classifying load curves in the load curve set cluster in the same state and in the same time length into one class;
step 4, respectively extracting typical power consumption modes aiming at each type of load curve to obtain a typical power consumption mode set cluster corresponding to each industrial user;
and 5, fusing the typical electricity utilization mode set clusters of the industrial users to obtain a typical electricity utilization mode.
2. The distributed method for extracting typical electricity usage patterns according to claim 1, wherein the step 1 comprises:
respectively aiming at each industrial user in a target area, acquiring power load data of N days from an intelligent ammeter of the industrial user; the sampling frequency of the power load data sampling every day is T;
preprocessing the power load data of N days to obtain power load data X;
normalizing the power load data X to obtain normalized power load data X';
wherein, X = [ X ] 1 ，x 2 ，…，x d ，…，x N ]And x is d ＝[x d，1 ，x d，2 ，…，x d，t ，…，x d，T ]，x d，t Representing the daily load data collected by the industrial user in the t time period on the d day, wherein the value range of t is [1,T ]]。
3. The distributed method for extracting typical electricity usage patterns according to claim 2, wherein the step 2 includes:
respectively processing the normalized power load data X' aiming at the power load data of each industrial user to obtain the power load mean value mu of the industrial user;
dividing the daily load data into load states according to the power load mean value mu to obtain the division results of the load states, namely a ground state load and an excited state load;
and acquiring a load state matrix S of the industrial user according to the power load mean value mu and the division result.
4. The distributed method for extracting typical electricity usage patterns according to claim 3, wherein the step 2 further comprises:
the average value mu of the power load is
Wherein N represents the number of sampling days, T represents the frequency of sampling the power load data per day,
x′ d，t represents the normalized power load value, x, of the t-th time period on the d-th day d，t Representing daily load data collected by the industrial user in the t period on the d day, and max (X) representing the most historical power load data of the industrial userA large load value, min (X) represents a minimum load value in historical power load data of an industrial user;
by passing
Dividing the load state of the power load data into a ground state load and an excited state load when S is d，t When "= 0, S is the ground state load d，t When =1, the load state is the excited load;
based on the power load mean value mu and the load state of the industrial users, acquiring the load state matrix S of each industrial user as
Wherein S is d，t Represents the state of the electrical load of the industrial user at the time of day d, t, and S d，t E {0,1}, where N represents the number of sampling days and T represents the frequency of power load data sampling per day.
5. The distributed method for extracting typical electricity usage patterns according to claim 4, wherein the step 3 includes:
defining the load state set at the tth day time in the load state matrix S of each industrial user as { S d，t } t＝1，2，...T ；
When t =1 and S d，t If =0, the time t is the trigger time of the ground state load, S d，t If =1, the time t is set as the end time of the ground state load;
when t =1 and S d，t If =1, the time t is taken as the trigger time of the excited load, S d，t If =0, the time t is set as the end time of the excited load;
respectively aiming at the load state of each industrial user, and generating the ground state load of the industrial user in the d day according to the triggering time and the ending time of the load stateIth load curve Lbas d，i And a jth load curve Lexc for the excited state load d，j ；
Apply the load curve Lbas d，i And the load curve Lexc d，j Respectively packaging to obtain a Load curve set Load _ bas of the industrial user about the ground state Load on the day d d And a set of Load curves for the excited state Load, load _ exc d ；
Collecting the Load curve to Load _ bas d And the Load curve set Load _ exc d Packaging to obtain a Load curve set cluster Load _ model of the industrial user under all Load states on the d day d Collecting Load curve set cluster Load _ model under all Load states for N days d Packaging to obtain a Load curve set cluster Load _ model of the industrial user;
and classifying the Load curve sets with the same duration time in the same Load state in the Load curve set cluster Load _ model into one type to obtain two types of Load curve set clusters, wherein the two types of Load curve set clusters are respectively a typical Load curve set cluster LenLoad _ bas of the industrial user under the ground state Load and a typical Load curve set cluster LenLoad _ exc of the industrial user under the excited state Load.
6. The distributed method for extracting typical electricity usage patterns according to claim 5, wherein the step 4 comprises:
respectively aiming at each type of load curve, acquiring the distance between any two load curves X and Y with the same time length as DTW (X, Y) by using a DTW-based distance measurement method;
clustering load curve set cluster LenLoad _ bas and load curve set cluster LenLoad _ exc based on DTW (X, Y) load curve AP clustering algorithm, and extracting typical power consumption mode set ReLoad _ Model of the pth industrial user in different load states p ；
Calling a textFile () function in the SparkContext, creating an RDD (resource description device) related to the electricity load data of each industrial user, and performing mode extraction from the RDD through a Spark distributed parallel computing framework to obtain a typical electricity consumption mode set cluster ReLoad _ Model corresponding to each industrial user.
7. The distributed method for extracting typical electricity usage patterns according to claim 1, wherein the step 5 is further followed by:
classifying a plurality of industrial users according to the typical power consumption mode to obtain an industrial user cluster;
and carrying out big data analysis and visual processing on the typical power consumption mode and the industrial user cluster to obtain a visual data analysis report.
8. A system for distributed extraction of typical power usage patterns, comprising:
the data acquisition module is used for acquiring power load data of a plurality of industrial users in a target area;
the data processing module is used for processing the power load data of the industrial users respectively aiming at the power load data of each industrial user to obtain a power utilization state matrix of the industrial users;
the data classification module is used for generating a load curve set cluster by utilizing the electricity utilization state matrixes of the industrial users and classifying load curves with the same time length in the same state in the load curve set cluster into one class;
the data extraction module is used for extracting typical power consumption modes aiming at each type of load curve respectively to obtain a typical power consumption mode set cluster corresponding to each industrial user;
and the data fusion module is used for fusing the typical electricity utilization mode set clusters of the industrial users to obtain a typical electricity utilization mode.
9. A computer-readable medium, comprising: for storing a computer program, characterized in that it is adapted to carry out the method of distributed extraction of typical electricity usage patterns of any of the preceding claims 1-7 by executing said computer program.
10. A device for distributed extraction of typical electricity usage patterns, characterized in that the method for realizing distributed extraction of typical electricity usage patterns according to any one of claims 1 to 7 comprises: a memory and a processor;
the memory is used for storing a computer program;
the processor is configured to execute the computer program stored by the memory.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202211010754.6A CN115358784A (en) | 2022-08-23 | 2022-08-23 | Method for extracting typical power consumption modes in distributed mode and related equipment |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202211010754.6A CN115358784A (en) | 2022-08-23 | 2022-08-23 | Method for extracting typical power consumption modes in distributed mode and related equipment |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN115358784A true CN115358784A (en) | 2022-11-18 |
Family
ID=84002688
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202211010754.6A Pending CN115358784A (en) | 2022-08-23 | 2022-08-23 | Method for extracting typical power consumption modes in distributed mode and related equipment |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN115358784A (en) |
Cited By (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN115543991A (en) * | 2022-12-02 | 2022-12-30 | 湖南工商大学 | Data restoration method and device based on feature sampling and related equipment |
| CN117195066A (en) * | 2023-08-21 | 2023-12-08 | 中南大学 | Distributed power equipment fault detection method, system, storage medium and processor |
-
2022
- 2022-08-23 CN CN202211010754.6A patent/CN115358784A/en active Pending
Cited By (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN115543991A (en) * | 2022-12-02 | 2022-12-30 | 湖南工商大学 | Data restoration method and device based on feature sampling and related equipment |
| CN115543991B (en) * | 2022-12-02 | 2023-03-10 | 湖南工商大学 | Data restoration method and device based on feature sampling and related equipment |
| CN117195066A (en) * | 2023-08-21 | 2023-12-08 | 中南大学 | Distributed power equipment fault detection method, system, storage medium and processor |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN115358784A (en) | Method for extracting typical power consumption modes in distributed mode and related equipment | |
| CN112561156A (en) | Short-term power load prediction method based on user load mode classification | |
| Vercamer et al. | Predicting consumer load profiles using commercial and open data | |
| CN112330078B (en) | Power consumption prediction method and device, computer equipment and storage medium | |
| CN107248031B (en) | Rapid power consumer classification method aiming at load curve peak-valley difference | |
| CN111612275A (en) | Method and device for predicting load of regional user | |
| Cai et al. | Efficient time series clustering by minimizing dynamic time warping utilization | |
| Himeur et al. | On the applicability of 2d local binary patterns for identifying electrical appliances in non-intrusive load monitoring | |
| CN109657705A (en) | A kind of automobile user clustering method and device based on random forests algorithm | |
| Miraftabzadeh et al. | Knowledge Extraction From PV Power Generation With Deep Learning Autoencoder and Clustering-Based Algorithms | |
| CN113094448B (en) | Analysis method and analysis device for residence empty state and electronic equipment | |
| CN113743977A (en) | User behavior-based electricity consumption data feature extraction method and system | |
| CN115146744B (en) | Electric energy meter load real-time identification method and system integrating time characteristics | |
| Durairaj et al. | Random forest based power sustainability and cost optimization in smart grid | |
| CN114330440B (en) | Distributed power supply load abnormality identification method and system based on simulation learning discrimination | |
| Miyasawa et al. | Energy disaggregation based on semi-supervised matrix factorization using feedback information from consumers | |
| CN112241922B (en) | Power grid asset comprehensive value assessment method based on improved naive Bayesian classification | |
| Guo et al. | Personalized home BESS recommender system based on neural collaborative filtering | |
| CN111222688B (en) | Daily load prediction method for commercial building | |
| CN111062502B (en) | User electricity consumption behavior subdivision method and fault analysis method thereof | |
| Thangavel et al. | Forecasting energy demand using conditional random field and convolution neural network | |
| Xia et al. | Energy-saving analysis of cloud workload based on K-means clustering | |
| Iskandarnia et al. | Load forecasting in different scale and horizon-a review | |
| Papadopoulos et al. | Efficient design under uncertainty of renewable power generation systems using partitioning and regression in the course of optimization | |
| CN117834455B (en) | Electric power Internet of things data transmission simulation method and system |
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 |