CN112070625A - User electricity consumption behavior pattern recognition method and system - Google Patents

User electricity consumption behavior pattern recognition method and system Download PDF

Info

Publication number
CN112070625A
CN112070625A CN202010800337.6A CN202010800337A CN112070625A CN 112070625 A CN112070625 A CN 112070625A CN 202010800337 A CN202010800337 A CN 202010800337A CN 112070625 A CN112070625 A CN 112070625A
Authority
CN
China
Prior art keywords
model
load
fhmm
behavior pattern
electricity consumption
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.)
Withdrawn
Application number
CN202010800337.6A
Other languages
Chinese (zh)
Inventor
张国防
张明松
Current Assignee (The listed assignees may be inaccurate. Google has not performed a legal analysis and makes no representation or warranty as to the accuracy of the list.)
Nanjing Hangling Information Technology Co ltd
Original Assignee
Nanjing Hangling Information Technology Co ltd
Priority date (The priority date is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the date listed.)
Filing date
Publication date
Application filed by Nanjing Hangling Information Technology Co ltd filed Critical Nanjing Hangling Information Technology Co ltd
Priority to CN202010800337.6A priority Critical patent/CN112070625A/en
Publication of CN112070625A publication Critical patent/CN112070625A/en
Withdrawn legal-status Critical Current

Links

Images

Classifications

    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06Q—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
    • G06Q50/00—Systems or methods specially adapted for specific business sectors, e.g. utilities or tourism
    • G06Q50/06—Electricity, gas or water supply
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06F—ELECTRIC DIGITAL DATA PROCESSING
    • G06F18/00—Pattern recognition
    • G06F18/20—Analysing
    • G06F18/21—Design or setup of recognition systems or techniques; Extraction of features in feature space; Blind source separation
    • G06F18/214—Generating training patterns; Bootstrap methods, e.g. bagging or boosting
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06F—ELECTRIC DIGITAL DATA PROCESSING
    • G06F18/00—Pattern recognition
    • G06F18/20—Analysing
    • G06F18/23—Clustering techniques

Abstract

The invention discloses a method and a system for identifying a user power consumption behavior pattern, which relate to the technical field of power systems and automation, and comprise the following steps: s1: establishing an HMM model of a single electric appliance load; s2: estimating parameters of an HMM model based on DBSCAN clustering; s3: establishing a total load model FHMM of the family; s4: the invention discloses a method and a system for identifying a user electricity consumption behavior pattern, which are based on FHMM model to carry out load decomposition on user electricity consumption, wherein a non-invasive decomposition method of low-frequency data based on the FHMM model is utilized, so that the method and the system have strong anti-interference capability and good clustering effect, and the trained HMM model has high accuracy; and the load decomposition problem is converted into an optimization problem for solving the maximum probability, the optimization problem is solved by using an integer programming method, the running state sequence of each electrical appliance is obtained by solving, the purpose of load decomposition is achieved, and a good decomposition recognition result is obtained.

Description

User electricity consumption behavior pattern recognition method and system
Technical Field
The invention belongs to the technical field of power systems and automation, and particularly relates to a user electricity consumption behavior pattern recognition method and system.
Background
With the development of intelligent measurement and demand-side management, the sampling frequency required by load decomposition is low, and the method is more suitable for the existing intelligent measurement device, so that the method becomes a very popular decomposition method in recent years. The low-frequency data can be data such as current amplitude, active power and reactive power sampled once per minute, and the intelligent measuring equipment can well meet the data requirements. With the upgrading of industrial manufacturing industry, the kinds of household appliances are more and more, the structural functions are more and more complex, and the load decomposition becomes more difficult.
Disclosure of Invention
The invention aims to provide a method and a system for identifying a user electricity consumption behavior pattern, so as to solve the problems in the background technology.
In order to achieve the purpose, the invention provides the following technical scheme: a method and a system for identifying a user electricity consumption behavior pattern comprise the following steps:
s1: establishing an HMM model of a single electric appliance load;
s2: estimating parameters of an HMM model based on DBSCAN clustering;
s3: establishing a total load model FHMM of the family;
s4: and carrying out load decomposition on the power consumption of the user based on the FHMM model.
Further, the establishment of the HMM model can use DBSCAN clustering parameters
Figure BDA0002627135960000011
Describing that building an HMM for a single appliance actually uses a sequence of observations to estimate three matrix parameters.
Further, the
Figure BDA0002627135960000012
The estimation method of the three matrix parameters is as follows:
s1: firstly, carrying out preliminary analysis on input observation data, and calculating input parameters EPS and MinPts of the DBSCAN by using a self-adaptive method;
s2: then inputting the observation data, the parameters EPS and MinPts into a DBSCAN algorithm for clustering to obtain a clustering result with a label, wherein the value of the label is-1, namely a noise point, and after the noise point is eliminated, calculating three parameter matrixes Pi, A and e;
s3: finally, HMM parameters are returned
Figure BDA0002627135960000021
Further, the total load model FHMM of the family can be established by three parameters
Figure BDA0002627135960000022
The description is that the total load is a result of the superposition of a plurality of power loads, and is specifically shown as follows;
ii denotes the initial state probability, which is determined by the initial states of a plurality of electric appliances, i.e. P (Z)1)＝P(S1 1,S1 2,…S1 T)；
A is also a state transition matrix describing the previous state z of the plurality of appliancest-1Transition to the next state ztProbability of (a), P (z)t|zt-1)＝P(St 1,St 2...St N|St-1 1,St-1 2,…St-1 N,)；
e represents the output probability of a state, in a certain state ztThe probability of the total output P ((y) shown belowt|zt))＝P((yt|St 1,St 2...St N))。
Furthermore, the FHMM model based on the FHMM model is used for load decomposition of user electricity utilization, namely, the FHMM model of family load is established by using low-frequency data
Figure BDA0002627135960000023
Observing the obtained current or power data at the power supply inlet to obtain the most possible operation state Z of each electric appliance in the time periodt＝{St 1,St 2，...，St NDecomposing the user power load into an optimization problem for solving the maximum probability.
In another aspect, the invention also provides a system for identifying the user electricity consumption behavior pattern.
A user electricity consumption behavior pattern recognition system comprises an input observation sequence Y, HMM model, an observation sequence segmentation Y ═ Y1, Y2, …, YN }, a smoothing sequence of each segment and an integer programming solution according to an objective function, an implicit state sequence is returned, and finally the state sequence of each electrical appliance is obtained.
Compared with the prior art, the invention has the beneficial effects that:
1. the non-invasive decomposition method based on the FHMM model and based on the low-frequency data has strong anti-jamming capability and good clustering effect, so that the trained HMM model has high accuracy.
2. The load decomposition problem is converted into an optimization problem for solving the maximum probability, the optimization problem is solved by using an integer programming method, the running state sequence of each electric appliance is obtained by solving, the purpose of load decomposition is achieved, and a good decomposition recognition result is achieved.
Drawings
FIG. 1 is a flow chart of HMM model parameter estimation based on DBSCAN according to the present invention.
FIG. 2 is a flow chart of the present invention for load decomposition based on FHMM model.
Detailed Description
As shown in fig. 1-2, a method for identifying a user electricity consumption behavior pattern includes the following steps:
s1: establishing an HMM model of a single electric appliance load;
s2: estimating parameters of an HMM model based on DBSCAN clustering;
s3: establishing a total load model FHMM of the family;
s4: and carrying out load decomposition on the power consumption of the user based on the FHMM model.
The HMM model can be established by using DBSCAN clustering parameters
Figure BDA0002627135960000033
Describing that building an HMM for a single appliance actually uses a sequence of observations to estimate three matrix parameters.
Wherein the content of the first and second substances,
Figure BDA0002627135960000031
the estimation method of the three matrix parameters is as follows:
s1: firstly, carrying out preliminary analysis on input observation data, and calculating input parameters EPS and MinPts of the DBSCAN by using a self-adaptive method;
s2: then inputting the observation data, the parameters EPS and MinPts into a DBSCAN algorithm for clustering to obtain a clustering result with a label, wherein the value of the label is-1, namely a noise point, and after the noise point is eliminated, calculating three parameter matrixes Pi, A and e;
s3: finally, HMM parameters are returned
Figure BDA0002627135960000032
Wherein, the establishment of the total family load model FHMM can also be composed of three parameters
Figure BDA0002627135960000041
The description is that the total load is a result of the superposition of a plurality of power loads, and is specifically shown as follows;
ii denotes the initial state probability, which is determined by the initial states of a plurality of electric appliances, i.e. P (Z)1)＝P(S1 1,S1 2,…S1 T)；
A is also a state transition matrix describing the previous state z of the plurality of appliancest-1Transition to the next state ztProbability of (a), P (z)t|zt-1)＝P(St 1,St 2...St N|St-1 1,St-1 2,…St-1 N,)；
e represents the output probability of a state, in a certain state ztThe probability of the total output P ((y) shown belowt|zt))＝P((yt|St 1,St 2...St N))。
The FHMM model based on the FHMM model is used for load decomposition of user power utilization, wherein the FHMM model of family load is established by using low-frequency data
Figure BDA0002627135960000042
Observing the obtained current or power data at the power supply inlet to obtain the most possible operation state Z of each electric appliance in the time periodt＝{St 1,St 2，...，St NDecomposing the user power load into an optimization problem for solving the maximum probability.
Wherein, the model of the input observation sequence Y, HMM and the observation sequence segment Y ═ Y1,Y2,…,YNAnd solving each section of the flattened sequence by using integer programming according to a target function, and returning a hidden state sequence to finally obtain the state sequence of each electric appliance.
The working principle and the using process of the invention are as follows: the user power consumption behavior pattern recognition uses a BDSCAN clustering algorithm to train an HMM model of a single load, and has strong anti-interference capability and good clustering effect, so that the trained HMM model has high accuracy; for the total load model, as the total load is the result of the common superposition of a plurality of power loads, the establishment can also be carried out by three parameters
Figure BDA0002627135960000043
The method comprises the steps of utilizing low-frequency data to carry out load decomposition of user electricity consumption based on the FHMM model, and establishing the FHMM model of family load
Figure BDA0002627135960000044
The most possible operation state Zt of each electric appliance in the time period is obtained by observing the obtained current or power data at the power supply inlet { St1, St2.., StN }, the user power load is decomposed into an optimization problem for solving the maximum probability, and then the operation state sequence of each electric appliance is solved, so that the purpose of load decomposition is achieved, and a good decomposition recognition result is obtained.
Although embodiments of the present invention have been shown and described, it will be appreciated by those skilled in the art that changes, modifications, substitutions and alterations can be made in these embodiments without departing from the principles and spirit of the invention, the scope of which is defined in the appended claims and their equivalents.

Claims (6)

1. A user electricity consumption behavior pattern recognition method is characterized by comprising the following steps: the method comprises the following steps:
s1: establishing an HMM model of a single electric appliance load;
s2: estimating parameters of an HMM model based on DBSCAN clustering;
s3: establishing a total load model FHMM of the family;
s4: and carrying out load decomposition on the power consumption of the user based on the FHMM model.
2. The method for recognizing the user electricity consumption behavior pattern according to claim 1, characterized in that: the establishment of the HMM model can use DBSCAN clustering parameters
Figure FDA0002627135950000011
Describing that building an HMM for a single appliance actually uses a sequence of observations to estimate three matrix parameters.
3. The method for recognizing the user electricity consumption behavior pattern according to claim 2, characterized in that: the above-mentioned
Figure FDA0002627135950000012
The estimation method of the three matrix parameters is as follows:
s1: firstly, carrying out preliminary analysis on input observation data, and calculating input parameters EPS and MinPts of the DBSCAN by using a self-adaptive method;
s2: then inputting the observation data, the parameters EPS and MinPts into a DBSCAN algorithm for clustering to obtain a clustering result with a label, wherein the value of the label is-1, namely a noise point, and after the noise point is eliminated, calculating three parameter matrixes Pi, A and e;
s3: finally, HMM parameters are returned
Figure FDA0002627135950000013
4. The method for recognizing the user electricity consumption behavior pattern according to claim 1, characterized in that: the total load model FHMM of the family can be established by three parameters
Figure FDA0002627135950000014
The description is that the total load is a result of the superposition of a plurality of power loads, and is specifically shown as follows;
ii denotes an initial state probability, which is determined by the initial states of a plurality of electric appliances, i.e., P (Z1) ═ P (S11, S12, … S1T);
a is also a state transition matrix describing the probability of a previous state zt-1 of multiple appliances transitioning to a next state zt, P (zt | zt-1) ═ P (St1, St2.. StN | St-11, St-12, … St-1N');
e represents the output probability of a state, and the probability P of the total output expressed in a certain state zt ((yt | zt)) ═ P ((yt | St1, St2.. StN)).
5. The method and system for recognizing the user electricity consumption behavior pattern according to claim 1, wherein: the FHMM model based on the FHMM model is used for carrying out the load decomposition of the electricity consumption of the user, namely, the FHMM model of the family load is established by using low-frequency data
Figure FDA0002627135950000021
By observing the obtained current or power data at the power supply inlet to obtain the most possible operation state Zt of each electric appliance in the time period { St1, St2.., StN }, the decomposition of the user power load becomes an optimization problem for solving the maximum probability.
6. A user power consumption behavior pattern recognition system is characterized in that: the method comprises the steps of inputting an observation sequence Y, HMM model, segmenting an observation sequence Y to be { Y1, Y2, …, YN }, solving each segment of the flattened sequence by using integer programming according to a target function, returning an implicit state sequence, and finally obtaining the state sequence of each electrical appliance.
CN202010800337.6A 2020-08-11 2020-08-11 User electricity consumption behavior pattern recognition method and system Withdrawn CN112070625A (en)

Priority Applications (1)

Application Number Priority Date Filing Date Title
CN202010800337.6A CN112070625A (en) 2020-08-11 2020-08-11 User electricity consumption behavior pattern recognition method and system

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
CN202010800337.6A CN112070625A (en) 2020-08-11 2020-08-11 User electricity consumption behavior pattern recognition method and system

Publications (1)

Publication Number Publication Date
CN112070625A true CN112070625A (en) 2020-12-11

Family

ID=73661112

Family Applications (1)

Application Number Title Priority Date Filing Date
CN202010800337.6A Withdrawn CN112070625A (en) 2020-08-11 2020-08-11 User electricity consumption behavior pattern recognition method and system

Country Status (1)

Country Link
CN (1) CN112070625A (en)

Cited By (2)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN112613542A (en) * 2020-12-14 2021-04-06 国网甘肃省电力公司营销服务中心 Bidirectional LSTM-based enterprise decontamination equipment load identification method
CN113627661A (en) * 2021-08-02 2021-11-09 深圳供电局有限公司 Method for predicting charging load of electric automobile

Cited By (3)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN112613542A (en) * 2020-12-14 2021-04-06 国网甘肃省电力公司营销服务中心 Bidirectional LSTM-based enterprise decontamination equipment load identification method
CN112613542B (en) * 2020-12-14 2024-01-12 国网甘肃省电力公司营销服务中心 Bidirectional LSTM-based load identification method for enterprise decontamination equipment
CN113627661A (en) * 2021-08-02 2021-11-09 深圳供电局有限公司 Method for predicting charging load of electric automobile

Similar Documents

Publication Publication Date Title
CN109685314B (en) Non-intrusive load decomposition method and system based on long-term and short-term memory network
Zhou et al. Sequence-to-sequence load disaggregation using multiscale residual neural network
CN110580502A (en) Factor hidden Markov load decomposition method based on Gaussian mixture
CN113902104A (en) Non-invasive load monitoring method combining unsupervised domain self-adaptive strategy and attention mechanism
CN112070625A (en) User electricity consumption behavior pattern recognition method and system
CN109633301B (en) Non-invasive electrical appliance load identification method based on quantum genetic optimization
CN112215405A (en) Non-invasive type residential electricity load decomposition method based on DANN domain adaptive learning
Chen et al. An intelligent system for mining usage patterns from appliance data in smart home environment
CN105225000A (en) A kind of wind power probability model nonparametric probability method optimized based on Fuzzy Ordered
Chen et al. Nonintrusive load monitoring based on self-supervised learning
CN113970667A (en) Non-invasive load monitoring method based on midpoint value of prediction window
CN113762355B (en) User abnormal electricity behavior detection method based on non-invasive load decomposition
CN109214545B (en) Load decomposition algorithm based on human behaviors
CN117272230A (en) Non-invasive load monitoring method and system based on multi-task learning model
CN106340874A (en) Identification decision-making method and system for power load decomposition
CN114091338B (en) Method and device for establishing power load decomposition model
CN115099522A (en) Active and reactive carbon emission prediction method for special transformer user based on BP neural network
CN112016684B (en) Electric power terminal fingerprint identification method of deep parallel flexible transmission network
CN111598145B (en) Non-invasive load monitoring method based on mixed probability label time-varying constraint distribution
Wang et al. Non-intrusive load monitoring based on residual U-Net and conditional generation adversarial networks
CN115392318A (en) Intelligent building non-invasive load decomposition method based on parallel connection network
CN113158446A (en) Non-invasive electric load identification method
Li et al. A data augmentation method for distributed photovoltaic electricity theft using wasserstein generative adversarial network
Gan et al. Non-intrusive Load Monitoring Based on Self-attention
CN115859452B (en) Data-driven-based transferable load modeling method, device, equipment and medium

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
WW01 Invention patent application withdrawn after publication
WW01 Invention patent application withdrawn after publication

Application publication date: 20201211