CN112465385A - Demand response potential analysis method applying intelligent electric meter data - Google Patents

Demand response potential analysis method applying intelligent electric meter data Download PDF

Info

Publication number
CN112465385A
CN112465385A CN202011447244.6A CN202011447244A CN112465385A CN 112465385 A CN112465385 A CN 112465385A CN 202011447244 A CN202011447244 A CN 202011447244A CN 112465385 A CN112465385 A CN 112465385A
Authority
CN
China
Prior art keywords
user
energy consumption
temperature
data
hour
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
Application number
CN202011447244.6A
Other languages
Chinese (zh)
Other versions
CN112465385B (en
Inventor
侯慧
陈跃
李显强
张清勇
王建建
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.)
Wuhan University of Technology WUT
Original Assignee
Wuhan University of Technology WUT
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 Wuhan University of Technology WUT filed Critical Wuhan University of Technology WUT
Priority to CN202011447244.6A priority Critical patent/CN112465385B/en
Publication of CN112465385A publication Critical patent/CN112465385A/en
Application granted granted Critical
Publication of CN112465385B publication Critical patent/CN112465385B/en
Active legal-status Critical Current
Anticipated expiration legal-status Critical

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
    • 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
    • G06Q10/0631—Resource planning, allocation, distributing or scheduling for enterprises or organisations
    • G06Q10/06315—Needs-based resource requirements planning or analysis
    • 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
    • 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
    • G06Q10/0639—Performance analysis of employees; Performance analysis of enterprise or organisation operations
    • G06Q10/06393—Score-carding, benchmarking or key performance indicator [KPI] analysis
    • 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

Landscapes

  • Business, Economics & Management (AREA)
  • Engineering & Computer Science (AREA)
  • Human Resources & Organizations (AREA)
  • Economics (AREA)
  • Theoretical Computer Science (AREA)
  • Physics & Mathematics (AREA)
  • Strategic Management (AREA)
  • Entrepreneurship & Innovation (AREA)
  • General Physics & Mathematics (AREA)
  • General Business, Economics & Management (AREA)
  • Tourism & Hospitality (AREA)
  • Marketing (AREA)
  • Educational Administration (AREA)
  • Development Economics (AREA)
  • Data Mining & Analysis (AREA)
  • Operations Research (AREA)
  • Quality & Reliability (AREA)
  • Game Theory and Decision Science (AREA)
  • Health & Medical Sciences (AREA)
  • Water Supply & Treatment (AREA)
  • General Health & Medical Sciences (AREA)
  • Primary Health Care (AREA)
  • Probability & Statistics with Applications (AREA)
  • Public Health (AREA)
  • Life Sciences & Earth Sciences (AREA)
  • Artificial Intelligence (AREA)
  • Bioinformatics & Cheminformatics (AREA)
  • Bioinformatics & Computational Biology (AREA)
  • Computer Vision & Pattern Recognition (AREA)
  • Evolutionary Biology (AREA)
  • Evolutionary Computation (AREA)
  • General Engineering & Computer Science (AREA)
  • Air Conditioning Control Device (AREA)

Abstract

The invention provides a demand response potential analysis method applying intelligent electric meter data. The method comprises the steps of constructing energy consumption time sequence data and meteorological time sequence data of each user, and calculating a Pearson correlation coefficient; calculating the energy consumption of the temperature control load in hours and the temperature turning point of the user by a linear regression method, further constructing an energy consumption characteristic parameter vector of the user, clustering the screened user characteristic data vector set by an improved k-means method, and further calculating the user reference load of the temperature control load; calculating the real-time power of the temperature control load of each hour of the user according to the indoor temperature, and further calculating the real-time energy consumption cost of the temperature control load of each hour of the user; establishing an evaluation index and a comment subset corresponding to the evaluation index according to the indoor temperature of each hour of a user and the real-time energy consumption cost of the temperature control load of each hour of the user; by comparing and analyzing the benchmark load data, the response potential of the resident temperature control load demand is evaluated.

Description

Demand response potential analysis method applying intelligent electric meter data
Technical Field
The invention belongs to the technical field of intelligent power grids, and particularly relates to a demand response potential analysis method applying intelligent electric meter data.
Background
With the rapid development of economy, the energy consumption demand of residential users is also getting larger and larger, and meanwhile, the peak shaving potential of the user side cannot be developed. Therefore, analyzing and utilizing the demand-side resource and performing demand response become an important approach to solving the problem of power supply and demand imbalance. Meanwhile, the smart meter plays a crucial role in demand response. Through the intelligent electric meter, a user can conveniently access to the electric power system, and bidirectional flow of information and energy with an electric power company is realized. According to relevant data statistics, the proportion of the resident users participating in the demand response to reduce the load is respectively 27%, but the demand response is not popularized among the resident users at present, and the response potential of the resident load is not completely developed.
In the related technical field of data analysis and application of intelligent electric meters, at present, the influence of user behaviors on load groups is mostly ignored in load prediction or demand response analysis work, and the great influence of the user behaviors on response results in demand response activities is rarely considered. Meanwhile, in the aspect of response potential analysis, the prior art focuses on predictive analysis of a load curve after response analysis, and lacks determination and comparative analysis of user reference load before demand response. Based on the method, the problem that the user is uncertain in the demand response activity is fully considered, the user decision is simulated, and the regional response potential is obtained through the comparison and analysis of the basic load.
Disclosure of Invention
The technical scheme of the invention is a demand response potential analysis method applying intelligent electric meter data, which comprises the following steps:
step 1: constructing energy consumption time sequence data and meteorological time sequence data of each user, calculating a Pearson correlation coefficient between the energy consumption time sequence data and each meteorological factor time sequence data, calculating temperature control load hour energy consumption by a linear regression method, calculating total energy consumption of the user according to the temperature control load hour energy consumption, calculating a temperature turning point of the user, further constructing an energy consumption characteristic parameter vector of the user, and screening according to the energy consumption characteristic parameter vector of the user to obtain a screened user characteristic data vector set;
step 2, dividing users with similar load modes into the same cluster by adopting an improved k-means clustering method for the screened user characteristic data vector set, and further calculating the reference load of the temperature control load user;
and step 3: calculating the refrigerating capacity of the user in each hour under the unit power of the temperature control load, further calculating the indoor temperature of the user in each hour, calculating the real-time power of the temperature control load of the user in each hour, further calculating the real-time energy consumption cost of the temperature control load of the user in each hour, and constructing an evaluation index and a comment subset corresponding to the evaluation index through the indoor temperature of the user in each hour and the real-time energy consumption cost of the temperature control load of the user in each hour;
and 4, step 4: by comparing and analyzing the benchmark load data, the response potential of the resident temperature control load demand is evaluated.
Preferably, in step 1, energy consumption time series data and meteorological time series data of each user are constructed, and a pearson correlation coefficient between the energy consumption time series data and each meteorological factor time series data is calculated as follows:
the user energy consumption time sequence data is as follows:
Xm＝[xm,1,xm,2,…xm,i,…,xm,I]
wherein M is 1,2, …, M is the total number of users, XmThe energy consumption time sequence Data of the mth user comprises user load Data per hour, and the total Data is mod [ I/24 ]]Day, xm,iThe energy consumption data of the mth user in the ith hour is represented, and I is the number of the hours;
the weather time sequence data of the user is as follows:
Ym＝[ym,1,ym,2,…,ym,n]
ym,i＝{vism,i,Toutm,i,dpm,i,prem,i,humm,i,spm,i}
wherein, YmIs the m < th > oneWeather time series data of the user; y ism,iIs the meteorological data set, Tout, of the ith hour in the meteorological time series data of the mth userm,iIs the outdoor temperature, vis, in the meteorological data set of the ith hour in the meteorological time sequence data of the mth userm,iIs visibility dp in meteorological data set of ith hour in meteorological time sequence data of mth userm,iIs the dew point, pre in the meteorological data set of the ith hour in the meteorological time sequence data of the mth userm,iIs the air pressure, hum in the meteorological data set of the ith hour in the meteorological time sequence data of the mth userm,iHumidity, sp, in the meteorological data set of the ith hour in the meteorological time series data of the mth userm,iWind speed in the meteorological data set of the ith hour in the meteorological time series data of the mth user;
the visibility time series data of the mth user is as follows: vism＝{vism,1,vism,2,…,vism,I}；
The Pearson correlation coefficient between the energy consumption time series data and the visibility factor time series data is as follows:
Figure BDA0002824959820000021
wherein, cov (X)m,vism) Is the covariance, vis, between the energy consumption time series data and the visibility time series data of the mth usermIn order for the visibility time series data to be visible,
Figure BDA0002824959820000022
for the standard deviation of the energy consumption time series data,
Figure BDA0002824959820000023
is the standard deviation of the visibility time series data;
Figure BDA0002824959820000024
Figure BDA0002824959820000025
wherein the content of the first and second substances,
Figure BDA0002824959820000026
is the average of the energy consumption of the mth user,
Figure BDA0002824959820000027
is the average value, x, of the visibility time series data of the mth userm,iIs the energy consumption data point, vis, of the m-th user in the ith hourm,iThe visibility in the meteorological data set of the ith hour in the meteorological time series data of the mth user is shown, wherein I is the number of hours;
step 1, calculating the energy consumption of the temperature control load hour by a linear regression method:
the energy consumption when the temperature control load is calculated is as follows:
Wtclm,i＝[βm+γm*(Toutm,i-cpm)-αm]*Sm,i
Figure BDA0002824959820000028
wherein, Wtclm,iThe energy consumption of the temperature control load of the mth user within the ith hour;
Toutm,ii hours outdoor temperature for the mth user;
αmthe base load per hour for the mth user;
cpmthe temperature turning point of the mth user is the variable to be solved;
βmintercept of regression model for mth user;
γmthe slope of the regression model for the mth user;
Sm,ithe temperature control load state parameter is the i hour of the mth user;
step 1, further calculating the total energy consumption of the user according to the temperature control load hour energy consumption as follows:
Figure BDA0002824959820000031
wherein, Wm,Sigma is the total energy consumption of the mth user in the I period;
step 1, calculating the temperature turning point of the user, wherein the specific calculation method comprises the following steps:
according to energy consumption time sequence data Xm＝[xm,1,xm,2,…,xm,n]Outdoor temperature time sequence data Toutm＝[Toutm,1,Toutm,2,…,Toutm,n]Traversing the temperature points in the interval according to the division value of 0.1 ℃, and totaling Lm＝(Max(Toutm)-Min(Toutm) 0.1 temperature values) and as turning point cpm,lThe regression model was introduced, i.e.:
cpm,l＝Min(Toutm)+0.1*l
wherein cpm,lThe temperature turning point of the first regression model of the mth user, wherein L is 0, 1.. and L;
Min(Toutm) The minimum value of the outdoor temperature of the mth user;
solving through a regression model to obtain an energy consumption regression sequence under the model:
fm,l,i＝βm,l+γm,l*(Toutm,i-cpm,l)
wherein f ism,l,iIs the energy consumption regression value, beta, of the mth user at the ith hour in the ith modelm,lIs the intercept, γ, in the first regression model of the mth userm,lSlope in the l regression model for the m user; toutm,iIs the outdoor temperature value, cp, of the mth user at the ith hourm,lIs the turning point in the first regression model of the mth user;
and calculating the regression correlation coefficient under the regression model as follows:
Figure BDA0002824959820000041
wherein R ism,l 2Regression correlation coefficient in the l model for the m-th user, fm,l,iFor the ith user, the ith energy consumption value in the ith model, i.e., energy consumption regression value, xm,iThe energy consumption data of the mth user in the ith hour;
thus forming a set of l model regression correlation coefficients for the mth user as:
Figure BDA0002824959820000042
obtaining the following results by taking the maximum value:
Figure BDA0002824959820000043
i.e. at*The maximum regression coefficient of the mth user is obtained from the model, and the l*The model is an optimal regression model of the mth user, and the load energy consumption temperature turning point of the mth user is as follows:
Figure BDA0002824959820000044
the corresponding intercept, slope, and base load correspond to the ith*Individual model values, i.e. the resulting regression model parameters, were:
Figure BDA0002824959820000045
at the moment, the regression correlation coefficient is maximum, the regression error is minimum, and the regression model effect is optimal;
carrying out regression analysis through the optimal regression model of the single user, and establishing a temperature control load equation of the single user:
Figure BDA0002824959820000046
Figure BDA0002824959820000047
wherein, Wtclm,iIs the energy consumption regression value, cp, of the mth user at the ith hour* mThe energy consumption temperature inflection point of the mth user;
step 1, the energy consumption characteristic parameter vector of the user is as follows:
Figure BDA0002824959820000048
wherein M is ∈ [1, M ∈]And M is the total number of users,
Figure BDA0002824959820000049
is the energy consumption characteristic parameter vector, cp, of the mth user* mThe energy consumption temperature inflection point of the mth user;
Figure BDA00028249598200000410
the base energy consumption for the mth user;
Figure BDA00028249598200000411
the energy consumption regression intercept of the mth user represents the basic energy consumption level of the mth user for starting the temperature control load;
Figure BDA00028249598200000412
the energy consumption regression slope of the mth user represents the temperature sensitivity,
Figure BDA00028249598200000413
the regression correlation coefficient of the mth user;
step 1, screening according to the energy consumption characteristic parameter vector of the user to obtain a screened user characteristic data vector set which is:
Figure BDA0002824959820000051
obtaining N screened user characteristic data vectors, and constructing a screened user characteristic data vector set;
the set of the screened user characteristic data vectors is as follows:
Figure BDA0002824959820000052
wherein the content of the first and second substances,
Figure BDA0002824959820000053
for the z-th filtered user feature data vector in the filtered user feature data vector set, i.e. the m-thzEnergy consumption characteristic parameter vector of each user, z belongs to [1, N ∈]N is the number of the screened user characteristic data vectors;
preferably, the improved k-means clustering method in step 2 has a clustering objective function of k-means as follows:
Figure BDA0002824959820000054
wherein SSE is the sum of squares of the minimum errors and is an objective function; numkRepresenting the number of sample set data points of the kth cluster;
Figure BDA0002824959820000055
is the n-th in the k-th classkUser feature data vector K1, 2, …, K, nk＝1,2,…,numkK is the number of clusters; u. ofkFor the kth cluster center, the calculation formula is:
Figure BDA0002824959820000056
the symbol "| | |" indicates taking the vector modulo;
step 2, the classification of users with similar load patterns into the same cluster is:
step 2.1, initialize clustering center u ═ u1,u2,…,uK,}；
Wherein u iskFor the kth cluster center, K ═ 1,2, …, K, the initialization process is as follows:
step 2.2, randomly selecting one from the input data point set DThe point is taken as the first clustering center u1；
Step 2.3, for each point in the dataset
Figure BDA0002824959820000057
Calculating the distance to the nearest cluster center in the selected cluster centers:
Figure BDA0002824959820000058
wherein the content of the first and second substances,
Figure BDA0002824959820000059
are data points
Figure BDA00028249598200000510
Distance to the nearest cluster center; u. ofkIs the k-th cluster center;
step 2.4, selecting a new data point as a new clustering center according to the following selection principle:
Figure BDA00028249598200000511
wherein u isk+1To find the (k + 1) th cluster center,
Figure BDA00028249598200000512
representing the data point corresponding to the maximum cluster;
step 2.5, repeating the step 2.3 and the step 2.4 until K clustering centers are selected;
the initial clustering center can be optimized through the steps;
step 2.6, calculate all sample points
Figure BDA0002824959820000061
To each cluster center u ═ u { [ u ]1,u2,…,uKDividing the distance into K clusters according to the nearest principle;
step 2.7, according to the formula
Figure BDA0002824959820000062
Updating K clustering centers;
wherein u iskIs the k-th cluster center; numkRepresenting the number of sample set data points of the kth cluster;
Figure BDA0002824959820000063
is the n-th in the k-th classkUser feature data vector K1, 2, …, K, nk＝1,2,…,numkK is the number of clusters;
step 2.8, repeating steps 2.6 and 2.7 until the maximum Iteration number Iteration is 500 or the clustering center u is not changed any more; at this time, updated K clusters c are obtained1、c2、…、cKThe cluster centers are respectively u1、u2、…、uKThe number of data points in each cluster is num1、num2、...、numK
Defining the cluster center of the cluster with the maximum data point number as a temperature control load user reference point:
Figure BDA0002824959820000064
wherein: kbasic＝arg Max(numk) Indicating the cluster number containing the most data points; cp (p)basicIs the temperature turning point of the reference load model; alpha is alphabasicIs the base load value of the reference load model; beta is abasicIs the regression intercept of the reference load model; gamma raybasicIs the regression slope of the reference load model; rbasicIs a reference load model regression coefficient;
step 2, calculating the user reference load of the temperature control load as follows:
user temperature control user reference load may be defined as:
Wtclbasic,i＝[βbasic+γbasic*(Toutbasic,i-cpbasic)-αbasic]*Sbasic,i
wherein:
Figure BDA0002824959820000065
Toutbasic,ithe average temperature of the area where the temperature control load user is located in i hours is represented; n is the number of users with temperature control load;
Figure BDA0002824959820000066
is m atzOutdoor temperature of the ith hour of the individual user.
Sbasic,iThe definition is as follows:
Figure BDA0002824959820000067
wherein Toutbasic,iThe average temperature of the area where the temperature control load user is located in the ith hour is represented; cp (p)basicIs the temperature turning point of the reference load model.
And obtaining the user reference load of the temperature control load according to the parameters and the steps.
Preferably, in step 3, the calculation of the refrigerating capacity per hour of the temperature-controlled load per unit power of the user is as follows:
Figure BDA0002824959820000071
wherein z ∈ [1, N ]]N is the number of users with temperature control load, I belongs to [1, I ∈]And I is the total number of hours,
Figure BDA0002824959820000072
is m atzThe refrigerating capacity of the temperature control load per unit power at the ith hour of each user,
Figure BDA0002824959820000073
for user mzTemperature-controlled load energy efficiency ratio of (S)mz,jRepresentsUser mzThe temperature controlled load operation state at the ith hour,
Figure BDA0002824959820000074
for user mzThe temperature control load rated power of (1);
step 3, further calculating the indoor temperature of the user per hour as follows:
Figure BDA0002824959820000075
wherein the content of the first and second substances,
Figure BDA0002824959820000076
is m atzThe indoor temperature of the i-th hour of the individual user,
Figure BDA0002824959820000077
is m atzThe outdoor temperature of the user for the ith hour,
Figure BDA0002824959820000078
is m atzThe equivalent impedance of the house heat exchange of an individual user,
Figure BDA0002824959820000079
is m atzThe house equivalent specific heat capacity of each user,
Figure BDA00028249598200000710
is m atzRefrigerating capacity of the temperature control load at the ith hour of each user under unit power;
and 3, calculating the real-time power of the temperature control load of each hour of the user as follows:
Figure BDA00028249598200000711
wherein the content of the first and second substances,
Figure BDA00028249598200000719
m thzFor one to useThe real-time power of the temperature control load in the ith hour,
Figure BDA00028249598200000712
representative user mzThe temperature controlled load operation state at the ith hour,
Figure BDA00028249598200000713
for user mzThe temperature control load rated power of (1);
step 3, further calculating the real-time energy consumption cost of the temperature control load per hour of the user as follows:
Figure BDA00028249598200000714
wherein the content of the first and second substances,
Figure BDA00028249598200000715
is m atzThe real-time energy consumption cost of the temperature control load of the ith hour of each user,
Figure BDA00028249598200000720
m thzTemperature controlled load real time power, ToU, for the ith hour of an individual useriThe time-of-use electricity price of the ith hour;
step 3, selecting the indoor temperature of each hour of the user and the real-time energy consumption cost of the temperature control load of each hour of the user as evaluation indexes, and defining a comment subset and the weight of the evaluation indexes;
through the m < th > mzIndoor temperature of individual user, mthzEstablishing an evaluation index of the electric charge of each user;
the evaluation indexes are as follows:
Figure BDA00028249598200000716
wherein, UmzIs m atzSet of fuzzy variables of individual users, first fuzzy variable
Figure BDA00028249598200000717
Is m atzIndoor temperature of individual user, second fuzzy variable
Figure BDA00028249598200000718
Is m atzThe electricity charge of the individual user;
step 3, the comment subset is:
v1{ "cold", "comfort", "hot" }
v2As { "cheap", "moderate", "expensive" }
Wherein v is1Denotes the m-thzSubset of comments on the indoor temperature of an individual user, v2Denotes the m-thzA comment subset of the individual user's electricity charge;
and 3, the weight vector of the evaluation index is as follows:
A＝{a1,a2}
wherein, a1Denotes the m-thzComment weight of indoor temperature of individual user, a2Represents the m-thzThe comment weight of the electric charge of each user;
the human thermal comfort index is calculated as follows:
Figure BDA0002824959820000081
wherein Tin is the ambient temperature, and PMV is the index of human thermal comfort;
the text takes the elasticity range value of the human comfort level index: PMVa＞PMVb＞PMVc＞PMVdAt this time, the corresponding indoor ambient temperature is Ta,Tb,Tc,Td；
In the blurring process, for a temperature at Tb～TcIs absolutely comfortable; and for a temperature range of Ta～Tb,Tc～TdFor relative comfort, i.e. not completely belonging to the "comfort" concept, the model uses a trapezoidal function to quantify the degree of membership:
the fuzzy membership function model is as follows:
Figure BDA0002824959820000082
Figure BDA0002824959820000083
Figure BDA0002824959820000084
wherein the content of the first and second substances,
Figure BDA0002824959820000085
is m atzH fuzzy variable for each user to comment set vhThe membership degree of the g-th comment, h is 1,2, g is 1,2, 3;
Figure BDA0002824959820000091
is m atzH fuzzy variable of each user; a, b, c and d represent the upper and lower limits of the membership function;
when the fuzzy variable is the indoor temperature
Figure BDA0002824959820000092
The upper and lower bounds of the fuzzy membership function respectively correspond to Ta,Tb,Tc,Td(ii) a At this point, the comment subset v can be accessed1{ "Cold", "comfort", "Heat" } and fuzzy variables
Figure BDA0002824959820000093
And establishing a membership relationship.
Similarly, when the fuzzy variable is the electricity charge, the comment subset v can be selected2Membership of { "cheap", "moderate", "expensive" } to the fuzzy variable electricity rate, i.e.
Figure BDA0002824959820000094
Evaluation vector calculation processing
Using fuzzy synthesis operator to make the weight vector A of evaluation index and evaluation matrix RmSynthesizing to obtain an evaluation result vector B:
Figure BDA0002824959820000095
wherein
Figure BDA00028249598200000916
Represents the fuzzy synthesis operator and the fuzzy synthesis operator,
Figure BDA0002824959820000096
in order to take the value of the weight,
Figure BDA0002824959820000097
is the fuzzy degree of membership.
Processing the fuzzy comprehensive evaluation vector by a weighted average principle to obtain a user satisfaction quantification result:
Figure BDA0002824959820000098
wherein the content of the first and second substances,
Figure BDA0002824959820000099
is m atzSatisfaction of the individual user; bm,gIs m atzThe g membership value of each user; num is a natural number, and the purpose is to control the weight distribution caused by larger membership;
and 3.3, optimizing the comprehensive satisfaction degree of the user by taking one day as a time unit, and outputting an optimization result of each day.
The objective function is:
Figure BDA00028249598200000910
wherein: j ═ D × 24+1, D × 24+2, …, D × 24+ 24; d-1, 2, …, Mod [ I/24 ]]Represents the jth time in 1-24 of day D, and the total Data is mod [ I/24 ]]Day;
Figure BDA00028249598200000911
Figure BDA00028249598200000912
for user mzAt time j on day D the room temperature was
Figure BDA00028249598200000913
The satisfaction of (2);
Figure BDA00028249598200000914
for user mzThe indoor temperature possible value at time j on day D,
al＝1,2,…,(Tinm,j-1-Toutm,j) 0.1, i.e. all possible room temperatures are traversed by a division value of 0.1, i.e.:
Figure BDA00028249598200000915
for the user, only the above sum (Tin)m,j-1-Toutm,j) The/0.1 set temperature can be selected, so the optimization problem is to traverse all the selected temperatures, and the aim of simulating the decision of the user is achieved by optimizing the comprehensive satisfaction pair.
Through the satisfaction optimizing model, the user mzThe decision result is:
Figure BDA0002824959820000101
i.e. user mzIndoor temperature was set at time j on day D
Figure BDA0002824959820000102
And the comprehensive satisfaction is optimal.
Bring this temperature value into user mzTemperature controlled load thermoelectric model, available user mzEnergy consumption at time j on day D, recorded as
Figure BDA0002824959820000103
Preferably, the user of the temperature-controlled load obtained in step 1 has a reference load at time i of Wtclbasic,i(ii) a From step 2.3, user m is availablezEnergy consumption at time j on day D of
Figure BDA0002824959820000104
The temperature control load user demand response potential is evaluated herein in terms of time periods of days. User mzThe response potential on day D was:
Figure BDA0002824959820000105
wherein the content of the first and second substances,
Figure BDA0002824959820000106
for user mzCorresponding potential at day D; wtclbasic,(D-1)*24+jThe reference load at the j moment on the D day;
Figure BDA0002824959820000107
) For user mzDemand response optimized energy consumption at day D, j.
The response potential of the temperature control load user group on day D is:
Figure BDA0002824959820000108
wherein, DRP∑,DResponse potential for the temperature controlled load user population on day D;
Figure BDA0002824959820000109
for user mzCorresponding potential at day D.
The invention fully considers the problem that the user is uncertain in the demand response activity, simulates the user decision and obtains the regional response potential through the comparison and analysis of the basic load.
Drawings
FIG. 1: the method of the invention is a flow chart.
FIG. 2: energy consumption weather correlation coefficient chart.
FIG. 3: and (4) a user energy consumption temperature regression schematic diagram.
FIG. 4: and (4) regression parameter distribution.
FIG. 5: the resident TCL.
FIG. 6: and clustering results of the TCL power curve.
FIG. 7: a single user decision result.
FIG. 8: population TCL profiles before and after response.
Detailed Description
For the purpose of facilitating the understanding and practice of the present invention, as will be described in further detail below with reference to the accompanying drawings and examples, it is to be understood that the examples described herein are for purposes of illustration and explanation, and are not intended to limit the invention.
The following describes an embodiment of the present invention with reference to fig. 1 to 8, which is a method for analyzing demand response potential using smart meter data, and includes the following steps:
the invention mainly constructs a resident TCL demand response potential evaluation model based on the data of the resident intelligent electric meter. In the model, a TCL energy consumption reference curve of the residential user is obtained through correlation analysis, linear regression analysis and k-means cluster analysis of the residential user energy consumption data on meteorological factors; through a fuzzy comprehensive evaluation method, a TCL scheduling strategy of comprehensive satisfaction considering user environment comfort and cost is provided; under the scene of time-of-use electricity price, the TCL demand response potential is evaluated through the comparative analysis of energy consumption data before and after the TCL of a resident user participates in response. And finally, simulating a single TCL user and a TCL user group in the region, and calculating to prove the effectiveness of the scheduling strategy and the evaluation model. The main technical route of the invention is shown in figure 1.
The method comprises the following steps:
step 1: constructing energy consumption time sequence data and meteorological time sequence data of each user, calculating a Pearson correlation coefficient between the energy consumption time sequence data and each meteorological factor time sequence data, calculating temperature control load hour energy consumption by a linear regression method, calculating total energy consumption of the user according to the temperature control load hour energy consumption, calculating a temperature turning point of the user, further constructing an energy consumption characteristic parameter vector of the user, and screening according to the energy consumption characteristic parameter vector of the user to obtain a screened user characteristic data vector set;
step 1, constructing energy consumption time sequence data and meteorological time sequence data of each user, and calculating a Pearson correlation coefficient between the energy consumption time sequence data and each meteorological factor time sequence data as follows:
the user energy consumption time sequence data is as follows:
Xm＝[xm,1,xm,2,…xm,i,…,xm,I]
wherein M is 1,2, …, M is the total number of users, XmThe energy consumption time sequence Data of the mth user comprises user load Data per hour, and the total Data is mod [ I/24 ]]Day, xm,iThe energy consumption data of the mth user in the ith hour is represented, and I is the number of the hours;
the weather time sequence data of the user is as follows:
Ym＝[ym,1,ym,2,…,ym,n]
ym,i＝{vism,i,Toutm,i,dpm,i,prem,i,humm,i,spm,i}
wherein, YmWeather time sequence data of the mth user; y ism,iIs the meteorological data set, Tout, of the ith hour in the meteorological time series data of the mth userm,iIs the outdoor temperature, vis, in the meteorological data set of the ith hour in the meteorological time sequence data of the mth userm,iIs visibility dp in meteorological data set of ith hour in meteorological time sequence data of mth userm,iIs the dew point, pre in the meteorological data set of the ith hour in the meteorological time sequence data of the mth userm,iIs the air pressure, hum in the meteorological data set of the ith hour in the meteorological time sequence data of the mth userm,iIs as followsHumidity, sp, in the meteorological data set of the ith hour in the meteorological time series data of m usersm,iWind speed in the meteorological data set of the ith hour in the meteorological time series data of the mth user;
the visibility time series data of the mth user is as follows: vism＝{vism,1,vism,2,…,vism,I}；
The Pearson correlation coefficient between the energy consumption time series data and the visibility factor time series data is as follows:
Figure BDA0002824959820000121
wherein, cov (X)m,vism) Is the covariance, vis, between the energy consumption time series data and the visibility time series data of the mth usermIn order for the visibility time series data to be visible,
Figure BDA0002824959820000127
for the standard deviation of the energy consumption time series data,
Figure BDA0002824959820000128
is the standard deviation of the visibility time series data;
Figure BDA0002824959820000122
Figure BDA0002824959820000123
wherein the content of the first and second substances,
Figure BDA0002824959820000124
is the average of the energy consumption of the mth user,
Figure BDA0002824959820000125
is the average value, x, of the visibility time series data of the mth userm,iIs the energy consumption data point of the ith hour of the mth user, vism,iThe visibility in the meteorological data set of the ith hour in the meteorological time series data of the mth user is shown, wherein I is the number of hours;
pearson correlation coefficients among seven items of data of power, temperature, visibility, dew point, air pressure, humidity and wind speed are formed into a matrix and are subjected to visualization processing, and the change of the size of the matrix elements is represented by an asymptotic color band, so that a heat map can be obtained, as shown in FIG. 2. Deeper in the heat map indicates a larger pearson correlation coefficient, and a stronger positive correlation.
Step 1, calculating the energy consumption of the temperature control load hour by a linear regression method:
the energy consumption when the temperature control load is calculated is as follows:
Wtclm,i＝[βm+γm*(Toutm,i-cpm)-αm]*Sm,i
Figure BDA0002824959820000126
wherein, Wtclm,iThe energy consumption of the temperature control load of the mth user within the ith hour;
Toutm,ii hours outdoor temperature for the mth user;
αmthe base load per hour for the mth user;
cpmthe temperature turning point of the mth user is the variable to be solved;
βmintercept of regression model for mth user;
γmthe slope of the regression model for the mth user;
Sm,ithe temperature control load state parameter is the i hour of the mth user;
regression was performed on the residential energy consumption data and the temperature data to obtain the results shown in fig. 3. For the above regression results, the user TCL regression day curve is obtained as shown in fig. 5, and the regression results show that TCL was at 10 a.m.: 00 reaches peak load, and thereafter continues until 20: 00, with a downward trend thereafter. At night 0: 00-5: and 00 is a valley.
Step 1, further calculating the total energy consumption of the user according to the temperature control load hour energy consumption as follows:
Figure BDA0002824959820000131
wherein, Wm,∑Total energy consumption for the mth user during the I period;
step 1, calculating the temperature turning point of the user, wherein the specific calculation method comprises the following steps:
according to energy consumption time sequence data Xm＝[xm,1,xm,2,…,xm,n]Outdoor temperature time sequence data Toutm＝[Toutm,1,Toutm,2,…,Toutm,n]Traversing the temperature points in the interval according to the division value of 0.1 ℃, and totaling Lm＝(Max(Toutm)-Min(Toutm) 0.1 temperature values) and as turning point cpm,lThe regression model was introduced, i.e.:
cpm,l＝Min(Toutm)+0.1*l
wherein cpm,lThe temperature turning point of the first regression model of the mth user, wherein L is 0,1, …, L;
Min(Toutm) The minimum value of the outdoor temperature of the mth user;
solving through a regression model to obtain an energy consumption regression sequence under the model:
fm,l,i＝βm,l+γm,l*(Toutm,i-cpm,l)
wherein f ism,l,iIs the energy consumption regression value, beta, of the mth user at the ith hour in the ith modelm,lIs the intercept, γ, in the first regression model of the mth userm,lSlope in the l regression model for the m user; toutm,iIs the outdoor temperature value, cp, of the mth user at the ith hourm,lIs the turning point in the first regression model of the mth user;
and calculating the regression correlation coefficient under the regression model as follows:
Figure BDA0002824959820000132
wherein R ism,l 2Regression correlation coefficient in the l model for the m-th user, fm,l,iFor the ith user, the ith energy consumption value in the ith model, i.e., energy consumption regression value, xm,iThe energy consumption data of the mth user in the ith hour;
thus forming a set of l model regression correlation coefficients for the mth user as:
Figure BDA0002824959820000133
obtaining the following results by taking the maximum value:
Figure BDA0002824959820000134
i.e. at*The maximum regression coefficient of the mth user is obtained from the model, and the l*The model is an optimal regression model of the mth user, and the load energy consumption temperature turning point of the mth user is as follows:
Figure BDA0002824959820000141
the corresponding intercept, slope, and base load correspond to the ith*Individual model values, i.e. the resulting regression model parameters, were:
Figure BDA0002824959820000142
at the moment, the regression correlation coefficient is maximum, the regression error is minimum, and the regression model effect is optimal;
carrying out regression analysis through the optimal regression model of the single user, and establishing a temperature control load equation of the single user:
Figure BDA0002824959820000143
Figure BDA0002824959820000144
wherein, Wtclm,iIs the energy consumption regression value, cp, of the mth user at the ith hour* mThe energy consumption temperature inflection point of the mth user;
step 1, the energy consumption characteristic parameter vector of the user is as follows:
Figure BDA0002824959820000145
wherein M is ∈ [1, M ∈]And M is the total number of users,
Figure BDA0002824959820000146
is the energy consumption characteristic parameter vector, cp, of the mth user* mThe energy consumption temperature inflection point of the mth user;
Figure BDA0002824959820000147
the base energy consumption for the mth user;
Figure BDA0002824959820000148
the energy consumption regression intercept of the mth user represents the basic energy consumption level of the mth user for starting the temperature control load;
Figure BDA0002824959820000149
the energy consumption regression slope of the mth user represents the temperature sensitivity,
Figure BDA00028249598200001410
the regression correlation coefficient of the mth user;
fig. 4 shows the distribution of regression parameters of the user.
Step 1, screening according to the energy consumption characteristic parameter vector of the user to obtain a screened user characteristic data vector set which is:
Figure BDA00028249598200001411
obtaining N screened user characteristic data vectors, and constructing a screened user characteristic data vector set;
the set of the screened user characteristic data vectors is as follows:
Figure BDA00028249598200001412
wherein the content of the first and second substances,
Figure BDA00028249598200001413
for the z-th filtered user feature data vector in the filtered user feature data vector set, i.e. the m-thzEnergy consumption characteristic parameter vector of each user, z belongs to [1, N ∈]N is the number of the screened user characteristic data vectors;
step 2, dividing users with similar load modes into the same cluster by adopting an improved k-means clustering method for the screened user characteristic data vector set, and further calculating the reference load of the temperature control load user; these curves were analyzed using k-means clustering based on the TCL regression characterization parameters, and the results are shown in FIG. 6
The improved k-means clustering method in the step 2 has the clustering objective function of k-means as follows:
Figure BDA0002824959820000151
wherein SSE is the sum of squares of the minimum errors and is an objective function; numkRepresenting the number of sample set data points of the kth cluster;
Figure BDA0002824959820000152
is the n-th in the k-th classkUser feature data vector K1, 2, …, K, nk＝1,2,…,numkK is the number of clusters; u. ofkFor the kth cluster center, the calculation formula is:
Figure BDA0002824959820000153
the symbol "| | |" indicates taking the vector modulo;
step 2, the classification of users with similar load patterns into the same cluster is:
step 2.1, initialize clustering center u ═ u1,u2,…,uK,}；
Wherein u iskFor the kth cluster center, K ═ 1,2, …, K, the initialization process is as follows:
step 2.2, randomly selecting a point from the input data point set D as a first clustering center u1；
Step 2.3, for each point in the dataset
Figure BDA0002824959820000154
Calculating the distance to the nearest cluster center in the selected cluster centers:
Figure BDA0002824959820000155
wherein the content of the first and second substances,
Figure BDA0002824959820000156
are data points
Figure BDA0002824959820000157
Distance to the nearest cluster center; u. ofkIs the k-th cluster center;
step 2.4, selecting a new data point as a new clustering center according to the following selection principle:
Figure BDA0002824959820000158
wherein u isk+1To find the (k + 1) th cluster center,
Figure BDA0002824959820000159
representing the data point corresponding to the maximum cluster;
step 2.5, repeating the step 2.3 and the step 2.4 until K clustering centers are selected;
the initial clustering center can be optimized through the steps;
step 2.6, calculate all sample points
Figure BDA00028249598200001510
To each cluster center u ═ u { [ u ]1,u2,…,uKDividing the distance into K clusters according to the nearest principle;
step 2.7, according to the formula
Figure BDA00028249598200001511
Updating K clustering centers;
wherein u iskIs the k-th cluster center; numkRepresenting the number of sample set data points of the kth cluster;
Figure BDA00028249598200001512
is the n-th in the k-th classkUser feature data vector K1, 2, …, K, nk＝1,2,…,numkK is the number of clusters;
step 2.8, repeating steps 2.6 and 2.7 until the maximum Iteration number Iteration is 500 or the clustering center u is not changed any more; at this time, updated K clusters c are obtained1、c2、…、cKThe cluster centers are respectively u1、u2、…、uKThe number of data points in each cluster is num1、num2、…、numK
Defining the cluster center of the cluster with the maximum data point number as a temperature control load user reference point:
Dbasic＝uKbasic＝[cpbasic、αbasic、βbasic、γbasic、Rbasic]
wherein: kbasic＝argMax(numk) Indicating the cluster number containing the most data points; cp (p)basicIs the temperature turning point of the reference load model; alpha is alphabasicIs the base load value of the reference load model;βbasicis the regression intercept of the reference load model; gamma raybasicIs the regression slope of the reference load model; rbasicIs a reference load model regression coefficient;
step 2, calculating the user reference load of the temperature control load as follows:
user temperature control user reference load may be defined as:
Wtclbasic,i＝[βbasic+γbasic*(Toutbasic,i-cpbasic)-αbasic]*Sbasic,i
wherein:
Figure BDA0002824959820000161
Toutbasic,ithe average temperature of the area where the temperature control load user is located in i hours is represented; n is the number of users with temperature control load;
Figure BDA0002824959820000162
is m atzOutdoor temperature of the ith hour of the individual user.
Sbasic,iThe definition is as follows:
Figure BDA0002824959820000163
wherein Toutbasic,iThe average temperature of the area where the temperature control load user is located in the ith hour is represented; cp (p)basicIs the temperature turning point of the reference load model.
And obtaining the user reference load of the temperature control load according to the parameters and the steps.
And step 3: calculating the refrigerating capacity of each hour of a user under the unit power of a temperature control load, further calculating the indoor temperature of each hour of the user, calculating the real-time power of the temperature control load of each hour of the user, further calculating the real-time energy consumption cost of the temperature control load of each hour of the user, and constructing an evaluation index and a comment subset corresponding to the evaluation index through the indoor temperature of each hour of the user and the real-time energy consumption cost of the temperature control load of each hour of the user;
and 3, calculating the refrigerating capacity of the user per hour under the unit power of the temperature control load as follows:
Figure BDA0002824959820000164
wherein z ∈ [1, N ]]N is the number of users with temperature control load, I belongs to [1, I ∈]And I is the total number of hours,
Figure BDA0002824959820000165
is m atzThe refrigerating capacity of the temperature control load per unit power at the ith hour of each user,
Figure BDA0002824959820000166
for user mzThe energy efficiency ratio of the temperature control load,
Figure BDA0002824959820000167
representative user mzThe temperature controlled load operation state at the ith hour,
Figure BDA0002824959820000168
for user mzThe temperature control load rated power of (1);
step 3, further calculating the indoor temperature of the user per hour as follows:
Figure BDA0002824959820000171
wherein the content of the first and second substances,
Figure BDA0002824959820000172
is m atzThe indoor temperature of the i-th hour of the individual user,
Figure BDA0002824959820000173
is m atzThe outdoor temperature of the user for the ith hour,
Figure BDA0002824959820000174
is m atzThe equivalent impedance of the house heat exchange of an individual user,
Figure BDA0002824959820000175
is m atzThe house equivalent specific heat capacity of each user,
Figure BDA0002824959820000176
is m atzRefrigerating capacity of the temperature control load at the ith hour of each user under unit power;
and 3, calculating the real-time power of the temperature control load of each hour of the user as follows:
Figure BDA0002824959820000177
wherein the content of the first and second substances,
Figure BDA0002824959820000178
m thzThe temperature control load real-time power of the ith hour of each user,
Figure BDA0002824959820000179
representative user mzThe temperature controlled load operation state at the ith hour,
Figure BDA00028249598200001710
for user mzThe temperature control load rated power of (1);
step 3, further calculating the real-time energy consumption cost of the temperature control load per hour of the user as follows:
Figure BDA00028249598200001711
wherein the content of the first and second substances,
Figure BDA00028249598200001712
is m atzThe real-time energy consumption cost of the temperature control load of the ith hour of each user,
Figure BDA00028249598200001713
m thzTemperature controlled load real time power, ToU, for the ith hour of an individual useriThe time-of-use electricity price of the ith hour;
step 3, selecting the indoor temperature of each hour of the user and the real-time energy consumption cost of the temperature control load of each hour of the user as evaluation indexes, and defining a comment subset and the weight of the evaluation indexes;
through the m < th > mzIndoor temperature of individual user, mthzEstablishing an evaluation index of the electric charge of each user;
the evaluation indexes are as follows:
Figure BDA00028249598200001714
wherein, UmzIs m atzSet of fuzzy variables of individual users, first fuzzy variable
Figure BDA00028249598200001715
Is m atzIndoor temperature of individual user, second fuzzy variable
Figure BDA00028249598200001716
Is m atzThe electricity charge of the individual user;
step 3, the comment subset is:
v1{ "cold", "comfort", "hot" }
v2As { "cheap", "moderate", "expensive" }
Wherein v is1Denotes the m-thzSubset of comments on the indoor temperature of an individual user, v2Denotes the m-thzA comment subset of the individual user's electricity charge;
and 3, the weight vector of the evaluation index is as follows:
A＝{a1,a2}
wherein, a1Denotes the m-thzComment weight of indoor temperature of individual user, a2Represents the m-thzOf individual usersComment weight of the electricity charge;
the human thermal comfort index is calculated as follows:
Figure BDA0002824959820000181
wherein Tin is the ambient temperature, and PMV is the index of human thermal comfort;
the text takes the elasticity range value of the human comfort level index: PMVa＞PMVb＞PMVc＞PMVdAt this time, the corresponding indoor ambient temperature is Ta,Tb,Tc,Td；
In the blurring process, for a temperature at Tb～TcIs absolutely comfortable; and for a temperature range of Ta～Tb,Tc～TdFor relative comfort, i.e. not completely belonging to the "comfort" concept, the model uses a trapezoidal function to quantify the degree of membership:
the fuzzy membership function model is as follows:
Figure BDA0002824959820000182
Figure BDA0002824959820000183
Figure BDA0002824959820000184
wherein the content of the first and second substances,
Figure BDA0002824959820000185
is m atzH fuzzy variable for each user to comment set vhThe membership degree of the g-th comment, h is 1,2, g is 1,2, 3;
Figure BDA0002824959820000186
is m atzH fuzzy variable of each user; a, b, c and d represent the upper and lower limits of the membership function;
when the fuzzy variable is the indoor temperature
Figure BDA0002824959820000187
The upper and lower bounds of the fuzzy membership function respectively correspond to Ta,Tb,Tc,Td(ii) a At this point, the comment subset v can be accessed1{ "Cold", "comfort", "Heat" } and fuzzy variables
Figure BDA0002824959820000188
And establishing a membership relationship.
Similarly, when the fuzzy variable is the electricity charge, the comment subset v can be selected2Membership of { "cheap", "moderate", "expensive" } to the fuzzy variable electricity rate, i.e.
Figure BDA0002824959820000189
Evaluation vector calculation processing
Using fuzzy synthesis operator to make the weight vector A of evaluation index and evaluation matrix RmSynthesizing to obtain an evaluation result vector B:
Figure BDA0002824959820000191
wherein
Figure BDA00028249598200001915
Represents the fuzzy synthesis operator and the fuzzy synthesis operator,
Figure BDA0002824959820000192
in order to take the value of the weight,
Figure BDA0002824959820000193
is the fuzzy degree of membership.
Processing the fuzzy comprehensive evaluation vector by a weighted average principle to obtain a user satisfaction quantification result:
Figure BDA0002824959820000194
wherein the content of the first and second substances,
Figure BDA0002824959820000195
is m atzSatisfaction of the individual user; bm,gIs m atzThe g membership value of each user; num is a natural number, and the purpose is to control the weight distribution caused by larger membership;
and 3.3, optimizing the comprehensive satisfaction degree of the user by taking one day as a time unit, and outputting an optimization result of each day.
The objective function is:
Figure BDA0002824959820000196
wherein: j ═ D × 24+1, D × 24+2, …, D × 24+ 24; d-1, 2, …, Mod [ I/24 ]]Represents the jth time in 1-24 of day D, and the total Data is mod [ I/24 ]]Day;
Figure BDA0002824959820000197
Figure BDA0002824959820000198
for user mzAt time j on day D the room temperature was
Figure BDA0002824959820000199
The satisfaction of (2);
Figure BDA00028249598200001910
for user mzThe indoor temperature possible value at time j on day D,
al＝1,2,…,(Tinm,j-1-Toutm,j) 0.1, i.e. all possible room temperatures are traversed by a division value of 0.1, i.e.:
Figure BDA00028249598200001911
for the user, only the above sum (Tin)m,j-1-Toutm,j) The/0.1 set temperature can be selected, so the optimization problem is to traverse all the selected temperatures, and the aim of simulating the decision of the user is achieved by optimizing the comprehensive satisfaction pair.
Through the satisfaction optimizing model, the user mzThe decision result is:
Figure BDA00028249598200001912
i.e. user mzIndoor temperature was set at time j on day D
Figure BDA00028249598200001913
And the comprehensive satisfaction is optimal.
Bring this temperature value into user mzTemperature controlled load thermoelectric model, available user mzEnergy consumption at time j on day D, recorded as
Figure BDA00028249598200001914
And 4, step 4: by comparing and analyzing the benchmark load data, the response potential of the resident temperature control load demand is evaluated.
The reference load of the user with the temperature control load obtained in the step 1 at the moment i is Wtclbasic,i(ii) a From step 2.3, user m is availablezEnergy consumption at time j on day D of
Figure BDA0002824959820000201
The temperature control load user demand response potential is evaluated herein in terms of time periods of days. User mzThe response potential on day D was:
Figure BDA0002824959820000202
wherein the content of the first and second substances,
Figure BDA0002824959820000203
for user mzCorresponding potential at day D; wtclbasic,(D-1)*24+jThe reference load at the j moment on the D day;
Figure BDA0002824959820000204
) For user mzDemand response optimized energy consumption at day D, j.
The response potential of the temperature control load user group on day D is:
Figure BDA0002824959820000205
wherein, DRP∑,DResponse potential for the temperature controlled load user population on day D;
Figure BDA0002824959820000206
for user mzCorresponding potential at day D.
The second embodiment of the present invention comprises the steps of:
step 1: based on the data of the intelligent electric meter, acquiring a temperature control load reference curve of a resident user by applying Pearson correlation coefficient, linear regression and k-means clustering;
step 2: under the scene of time-of-use electricity price, based on a temperature control load thermoelectric model, the comprehensive user satisfaction degree considering environment comfort and cost expenditure is quantitatively optimized through a fuzzy comprehensive evaluation method, user decision is simulated, and the load quantity after user response is obtained. By comparing and analyzing the benchmark load data, the response potential of the resident temperature control load demand is evaluated.
2. The demand response load reference curve analysis using smart meter data of claim 1, wherein:
the specific steps of establishing the resident user temperature control load reference curve in the step 1 are as follows:
taking a certain area in east China as an example, 1500 resident users are assumed to exist in the area, and simulation analysis is carried out on TCL demand response conditions under the condition of time-of-use electricity price of the resident users. In subtropical monsoon climates in the region, the temperature change in one year is about 0-35 ℃, high temperature and raininess are realized in summer, the temperature is mild in winter, and typical summer meteorological data are adopted as meteorological data.
Step 1.1: meteorological conditions, temperature, visibility, dew point, air pressure, humidity, wind speed and other factors all influence energy consumption of residents, and the correlation between energy consumption and meteorological factors is analyzed by utilizing the Pearson correlation coefficient. For energy consumption and meteorological time series data X ═ X1,x2,…,xn]And Y ═ Y1,y2,…,yn]The pearson correlation coefficient is:
Figure BDA0002824959820000207
wherein cov (X, Y) is X, the covariance of Y, and σ is the standard deviation. The value range of rho is-1. The maximum correlation between energy consumption is calculated as the temperature, and the Pearson correlation coefficient is 0.61, which shows that the correlation is strong.
Step 1.2: temperature-energy consumption regression analysis
1. In order to represent the specific correlation between the temperature and the load of the resident users, the real-time energy consumption of the temperature control load of a certain user is expressed as follows:
Wtcl＝[β+γ*(T(t)-Cp)-α]*S(t)
wherein, WtclTemperature-controlled load hourly energy consumption, T (t) is a temperature function, alpha is the hourly basic load, alpha is 0.5 kW.h, CpIs the turning point of temperature, CpBeta and gamma are respectively the intercept and slope of the regression model at 17 deg.C, beta is 0.72kWh, and gamma is 0.25 kWh/deg.C.
S (t) is a temperature control load state parameter, and the values are as follows:
Figure BDA0002824959820000211
2. total energy consumption of user for one day
Figure BDA0002824959820000212
Wherein, W is total energy consumption of the user, the temperature control load and the base load are cumulatively summed in one day, and W is 58 kW.h.
The specific process is
1) First, the change point C is determined by traversing the temperature pointspSo that the regression effect is the best, above which point the user energy consumption is sensitive to temperature.
2) The temperature response state of the single residential user is described by regression analysis of the data.
3) Calculating the temperature control load from the temperature response state, and determining whether the user uses the temperature control load and the time period for opening.
Step 1.3: based on correlation analysis and regression analysis, under the 95% significance level, users with rho > 0.5 and regression slope gamma > 0 are judged as temperature control load users, k-means clustering is adopted, and residential users with similar load modes are grouped, so that the temperature control load user reference load is finally obtained.
K-means clustering objective function
Figure BDA0002824959820000213
Wherein SSE is the sum of the squares of errors, wherein CiSample set, u, representing the ith clusteriIs the ith cluster center. x is sample data, k is the number of clusters, and k is 4.
2. Convergence condition of clustering
Figure BDA0002824959820000214
Wherein C isiSample set, u, representing the ith clusteriIs the ith cluster center. This equation is equivalent to the sum of squared errors being the minimum.
3. Hyper-parametric optimization
The sample x corresponding to the (i + 1) th initial cluster center should satisfy:
Figure BDA0002824959820000215
wherein i is more than or equal to 0 and less than n, which indicates that the next clustering center x should satisfy the distance
Figure BDA0002824959820000216
And max.
Figure BDA0002824959820000217
The method specifically comprises the following steps:
Figure BDA0002824959820000218
representing the set of distances from the respective farthest sample point corresponding to the first i cluster centers.
3. The demand response potential analysis using smart meter data of claim 1, wherein:
the specific steps for evaluating the temperature control load response potential of the resident user in the step 2 are as follows:
and 2.1, establishing a time sequence relation among the temperature control load power, the outdoor temperature and the room temperature through the thermoelectric parameters of the temperature control load unit at a certain time.
1. Temperature as a function of energy consumption:
Figure BDA0002824959820000221
wherein, ToutIs the outdoor temperature, at a certain time, Tout＝35℃，TinIs the indoor temperature, Tin＝28℃R is equivalent resistance of house heat exchange, R is 1, C is house equivalent specific heat capacity, C is 5.56, Q is refrigerating capacity per unit time:
Q(t)＝s(t)·P·k
and k is the temperature control load energy efficiency ratio, and k is 3 and represents the refrigerating capacity of the temperature control load under unit power. S represents the operation state of the temperature control load, P is the rated power of the temperature control load, and P is 3 kW.
2. The real-time power of the temperature control load is as follows:
p(t)＝s(t)·P
3. the real-time energy consumption cost of the temperature control load is as follows:
c(t)＝p(t)·pToU(t)
wherein c is the real-time electricity charge, PToUIs a time of use price of electricity, PToU0.8 yuan/(kW. h)
And 2.2, quantitatively analyzing the electricity utilization satisfaction degree of the temperature control load resident users containing two indexes of environment comfort and electricity expense by adopting a fuzzy comprehensive evaluation method.
1. Determining an evaluation factor set U ═ U1,u2,...,umV ═ V }, comment set1,v2,...,vnThe weight vector a ═ a1,a2,...,an}. The values are as follows:
Figure BDA0002824959820000222
2. determining a fuzzy membership function:
the average level of the cold and the heat of most people in a specific environment can be described according to the index of the thermal comfort degree of the human body. Taking the residential houses of residents in summer as an example, neglecting the influences of indoor wind speed, humidity, human body metabolism clothing thermal resistance and the like, the calculation formula can be simplified as follows:
Figure BDA0002824959820000223
where T is ambient temperature. PMV is a human thermal comfort index. T-26 ℃, PMV-0. Thereby determining the fuzzy membership function of the indoor temperature as follows:
Figure BDA0002824959820000224
Figure BDA0002824959820000231
Figure BDA0002824959820000232
wherein a, b, c and d represent the upper and lower limits of the membership function and are determined by PMV values: 22, 24, 26 and 29.
3. Evaluation vector calculation processing
Synthesizing the fuzzy weight vector A and the evaluation matrix R by using a fuzzy synthesis operator to obtain an evaluation result vector B:
Figure BDA0002824959820000233
wherein
Figure BDA0002824959820000239
Representing a fuzzy synthesis operator, taking the operator which is small firstly and then big, taking a as a weight value, and taking a as a1＝a20.5, r is the fuzzy membership degree,
Figure BDA0002824959820000234
obtain B ═ 0.50.50.3.
Processing the fuzzy comprehensive evaluation vector by a weighted average principle to obtain a user satisfaction quantification result:
Figure BDA0002824959820000235
wherein y isuFor user satisfaction, bjFor fuzzy estimation, k is natural number to control the weight distribution caused by larger membership degree, k is 2, and y is obtainedu＝1.729
And 2.3, evaluating fuzzy evaluation factor sets with different electric charges and temperature compositions at a certain time t to obtain the satisfaction of the user in different states, and optimizing to seek the maximum satisfaction of the user.
The objective function is:
Figure BDA0002824959820000236
wherein
Figure BDA0002824959820000237
ciIs a possible indoor temperature value of 24-35 ℃, i is more than or equal to 1 and less than or equal to l, l is a possible number, the division value is 0.1 ℃, l is 110,
Figure BDA0002824959820000238
is the corresponding user satisfaction. The temperature setting satisfies the range: t ismin≤Tin≤Tmax≤Tout，Tmin＝24℃，Tmax＝ToutFor the user, only the limited set temperature can be selected, so the optimization problem is to traverse all the selected temperatures, and the goal of simulating the decision of the user is achieved by optimizing the comprehensive satisfaction pair.
The satisfaction degree of the user at part of the temperature at the moment is as follows:
Figure BDA0002824959820000241
analysis found that the maximum overall user satisfaction was 2.9985 at room temperature of 28.6 ℃.
Fig. 6(a) shows TCL user clustering results, and fig. 6(b) shows corresponding frequency distributions. Analysis shows that users are mainly concentrated on the first type of users, the proportion is 63%, the alpha value is the minimum, the basic energy consumption is relatively lowest, and the beta value of the reaction temperature responsiveness is the maximum, which indicates that the type of users have better load flexibility. It was therefore designated as the TCL reference curve and used for the subsequent response potential analysis.
In the time-of-use electricity price scenario, simulation calculations were performed for a single TCL user, with the results shown in fig. 7. Wherein FIG. 7(a) shows the outdoor temperatures ToutAnd indoor temperature TinIndoor temperature during certain time periodsInside-follow outdoor temperature, such as 10: 00-20: period 00. That is, the resident TCL is not thermostatically controlled at the maximum user satisfaction decision, but is determined by both ambient temperature and electricity cost. Fig. 7(b) is a real-time TCL power curve of a user, and the peak power consumption of the user is mainly shown in 5: 00-10: 00, at the moment, the electricity price is moderate, but the outdoor temperature is increased rapidly, and the user starts the TCL before the high temperature and the high electricity price come, so that the cost is reduced. The required valley value is 1: 00-5: about 00. Analysis shows that the midnight air temperature is relatively low, and a user can choose to reduce TCL energy consumption under the condition of considering comprehensive satisfaction of cost and comfort, and particularly, the temperature is 5: at time 00, TCL is completely off for this period, and both load power and power cost are 0. Fig. 7(c) shows the electricity charge.
And 2.4, simulating the decision of the user and determining the response energy consumption of the user through a satisfaction decision model. And obtaining the demand response potential by comprehensively comparing the user temperature control load reference curves.
The group user related parameter values are as follows:
Figure BDA0002824959820000251
and simulating the demand response behavior of the TCL group of the residential users, analyzing the response behavior of the load group to the time-lapse electricity price, and evaluating the response potential of the load group on the basis of the benchmark load.
The calculation results are shown in fig. 8. Wherein the triangular labeled curve is the total power without requiring response, and the circular labeled curve is the reference load power. The time period for responding to the afterload reduction in fig. 8 is mainly focused on 10: 00-20: 00 and 1: 00-6: 00 midnight time period. During the period from noon to evening, the total load trend of the load groups participating in demand response is more stable than that before response, but the total load trend is obviously peak before response. By combining data analysis of outdoor temperature, time-of-use electricity price and the like, in the midnight time period, the outdoor temperature is lower, the satisfaction of user comfort and cost expenditure is considered comprehensively, the indoor temperature setting of the user basically follows the outdoor temperature, and the TCL energy consumption is obviously reduced. In the time period from noon to afternoon, the electricity price is higher at the moment, the comprehensive satisfaction of the user reflects that the user adjusts the indoor temperature to be higher properly along with the outdoor temperature on the premise of certain comfort level so as to reduce the electricity fee, and the load capacity is relatively stable at the moment.
The calculation results are shown in the table.
Figure BDA0002824959820000252
The calculation result shows that the load and cost of the participation demand response are obviously reduced, the load fluctuation is greatly reduced, and all indexes after the response are better than those before the response. The response potential is a response capacity reduction value and is 14MW & h.
It should be understood that parts of the application not described in detail are prior art.
It should be understood that the above description of the preferred embodiments is given for clearness of understanding and no unnecessary limitations should be understood therefrom, and all changes and modifications may be made by those skilled in the art without departing from the scope of the invention as defined by the appended claims.

Claims (5)

1. A demand response potential analysis method applying smart meter data comprises the following steps:
step 1: constructing energy consumption time sequence data and meteorological time sequence data of each user, calculating a Pearson correlation coefficient between the energy consumption time sequence data and each meteorological factor time sequence data, calculating temperature control load hour energy consumption by a linear regression method, calculating total energy consumption of the user according to the temperature control load hour energy consumption, calculating a temperature turning point of the user, further constructing an energy consumption characteristic parameter vector of the user, and screening according to the energy consumption characteristic parameter vector of the user to obtain a screened user characteristic data vector set;
step 2, dividing users with similar load modes into the same cluster by adopting an improved k-means clustering method for the screened user characteristic data vector set, and further calculating the reference load of the temperature control load user;
and step 3: calculating the refrigerating capacity of the user in each hour under the unit power of the temperature control load, further calculating the indoor temperature of the user in each hour, calculating the real-time power of the temperature control load of the user in each hour, further calculating the real-time energy consumption cost of the temperature control load of the user in each hour, and constructing an evaluation index and a comment subset corresponding to the evaluation index through the indoor temperature of the user in each hour and the real-time energy consumption cost of the temperature control load of the user in each hour;
and 4, step 4: by comparing and analyzing the benchmark load data, the response potential of the resident temperature control load demand is evaluated.
2. The demand response potential analysis method applying the smart meter data according to claim 1, wherein:
step 1, constructing energy consumption time sequence data and meteorological time sequence data of each user, and calculating a Pearson correlation coefficient between the energy consumption time sequence data and each meteorological factor time sequence data as follows:
the user energy consumption time sequence data is as follows:
Xm＝[xm,1,xm,2,…xm,i,…,xm,I]
wherein M is 1,2, …, M is the total number of users, XmThe energy consumption time sequence Data of the mth user comprises user load Data per hour, and the total Data is mod [ I/24 ]]Day, xm,iThe energy consumption data of the mth user in the ith hour is represented, and I is the number of the hours;
the weather time sequence data of the user is as follows:
Ym＝[ym,1,ym,2,…,ym,n]
ym,i＝{vism,i,Toutm,i,dpm,i,prem,i,humm,i,spm,i}
wherein, YmWeather time sequence data of the mth user; y ism,iWeather time series for mth userAccording to the meteorological data set of the ith hour, Toutm,iIs the outdoor temperature, vis, in the meteorological data set of the ith hour in the meteorological time sequence data of the mth userm,iIs visibility dp in meteorological data set of ith hour in meteorological time sequence data of mth userm,iIs the dew point, pre in the meteorological data set of the ith hour in the meteorological time sequence data of the mth userm,iIs the air pressure, hum in the meteorological data set of the ith hour in the meteorological time sequence data of the mth userm,iHumidity, sp, in the meteorological data set of the ith hour in the meteorological time series data of the mth userm,iWind speed in the meteorological data set of the ith hour in the meteorological time series data of the mth user;
the visibility time series data of the mth user is as follows: vism＝{vism,1,vism,2,…,vism,I}；
The Pearson correlation coefficient between the energy consumption time series data and the visibility factor time series data is as follows:
Figure FDA0002824959810000021
wherein, cov (X)m,vism) Is the covariance, vis, between the energy consumption time series data and the visibility time series data of the mth usermIn order for the visibility time series data to be visible,
Figure FDA0002824959810000022
for the standard deviation of the energy consumption time series data,
Figure FDA0002824959810000023
is the standard deviation of the visibility time series data;
Figure FDA0002824959810000024
Figure FDA0002824959810000025
wherein the content of the first and second substances,
Figure FDA0002824959810000026
is the average of the energy consumption of the mth user,
Figure FDA0002824959810000027
is the average value, x, of the visibility time series data of the mth userm,iIs the energy consumption data point, vis, of the m-th user in the ith hourm,iThe visibility in the meteorological data set of the ith hour in the meteorological time series data of the mth user is shown, wherein I is the number of hours;
step 1, calculating the energy consumption of the temperature control load hour by a linear regression method:
the energy consumption when the temperature control load is calculated is as follows:
Wtclm,i＝[βm+γm*(Toutm,i-cpm)-αm]*Sm,i
Figure FDA0002824959810000028
wherein, Wtclm,iThe energy consumption of the temperature control load of the mth user within the ith hour;
Toutm,ii hours outdoor temperature for the mth user;
αmthe base load per hour for the mth user;
cpmthe temperature turning point of the mth user is the variable to be solved;
βmintercept of regression model for mth user;
γmthe slope of the regression model for the mth user;
Sm,ithe temperature control load state parameter is the i hour of the mth user;
step 1, further calculating the total energy consumption of the user according to the temperature control load hour energy consumption as follows:
Figure FDA0002824959810000031
wherein, Wm,Sigma is the total energy consumption of the mth user in the I period;
step 1, calculating the temperature turning point of the user, wherein the specific calculation method comprises the following steps:
according to energy consumption time sequence data Xm＝[xm,1,xm,2,…,xm,n]Outdoor temperature time sequence data Toutm＝[Toutm,1,Toutm,2,…,Toutm,n]Traversing the temperature points in the interval according to the division value of 0.1 ℃, and totaling Lm＝(Max(Toutm)-Min(Toutm) 0.1 temperature values) and as turning point cpm,lThe regression model was introduced, i.e.:
cpm,l＝Min(Toutm)+0.1*l
wherein cpm,lThe temperature turning point of the first regression model of the mth user, wherein L is 0, 1.. and L;
Min(Toutm) The minimum value of the outdoor temperature of the mth user;
solving through a regression model to obtain an energy consumption regression sequence under the model:
fm,l,i＝βm,l+γm,l*(Toutm,i-cpm,l)
wherein f ism,l,iIs the energy consumption regression value, beta, of the mth user at the ith hour in the ith modelm,lIs the intercept, γ, in the first regression model of the mth userm,lSlope in the l regression model for the m user; toutm,iIs the outdoor temperature value, cp, of the mth user at the ith hourm,lIs the turning point in the first regression model of the mth user;
and calculating the regression correlation coefficient under the regression model as follows:
Figure FDA0002824959810000032
wherein R ism,l 2Regression correlation coefficient in the l model for the m-th user, fm,l,iFor the ith user, the ith energy consumption value in the ith model, i.e., energy consumption regression value, xm,iThe energy consumption data of the mth user in the ith hour;
thus forming a set of l model regression correlation coefficients for the mth user as:
Figure FDA0002824959810000033
obtaining the following results by taking the maximum value:
Figure FDA0002824959810000034
i.e. at*The maximum regression coefficient of the mth user is obtained from the model, and the l*The model is an optimal regression model of the mth user, and the load energy consumption temperature turning point of the mth user is as follows:
Figure FDA0002824959810000035
the corresponding intercept, slope, and base load correspond to the ith*Individual model values, i.e. the resulting regression model parameters, were:
Figure FDA0002824959810000041
at the moment, the regression correlation coefficient is maximum, the regression error is minimum, and the regression model effect is optimal;
carrying out regression analysis through the optimal regression model of the single user, and establishing a temperature control load equation of the single user:
Figure FDA0002824959810000042
Figure FDA0002824959810000043
wherein, Wtclm,iIs the energy consumption regression value of the mth user at the ith hour,
Figure FDA0002824959810000044
the energy consumption temperature inflection point of the mth user;
step 1, the energy consumption characteristic parameter vector of the user is as follows:
Figure FDA0002824959810000045
wherein M is ∈ [1, M ∈]And M is the total number of users,
Figure FDA00028249598100000415
is the energy consumption characteristic parameter vector, cp, of the mth user* mThe energy consumption temperature inflection point of the mth user;
Figure FDA0002824959810000047
the base energy consumption for the mth user;
Figure FDA0002824959810000048
the energy consumption regression intercept of the mth user represents the basic energy consumption level of the mth user for starting the temperature control load;
Figure FDA0002824959810000049
the energy consumption regression slope of the mth user represents the temperature sensitivity,
Figure FDA00028249598100000410
the regression correlation coefficient of the mth user;
step 1, screening according to the energy consumption characteristic parameter vector of the user to obtain a screened user characteristic data vector set which is:
Figure FDA00028249598100000411
obtaining N screened user characteristic data vectors, and constructing a screened user characteristic data vector set;
the set of the screened user characteristic data vectors is as follows:
Figure FDA00028249598100000412
wherein the content of the first and second substances,
Figure FDA00028249598100000413
for the z-th filtered user feature data vector in the filtered user feature data vector set, i.e. the m-thzEnergy consumption characteristic parameter vector of each user, z belongs to [1, N ∈]And N is the number of the screened user characteristic data vectors.
3. The demand response potential analysis method applying the smart meter data according to claim 1, wherein:
the improved k-means clustering method in the step 2 has the clustering objective function of k-means as follows:
Figure FDA00028249598100000414
wherein SSE is the sum of squares of the minimum errors and is an objective function; numkRepresenting the number of sample set data points of the kth cluster;
Figure FDA0002824959810000051
is the n-th in the k-th classkUser feature data vector K1, 2, …, K, nk＝1,2,…,numkK is the number of clusters; u. ofkFor the kth cluster center, the calculation formula is:
Figure FDA0002824959810000052
the symbol "| | |" indicates taking the vector modulo;
step 2, the classification of users with similar load patterns into the same cluster is:
step 2.1, initialize clustering center u ═ u1,u2,…,uK,}；
Wherein u iskFor the kth cluster center, K ═ 1,2, …, K, the initialization process is as follows:
step 2.2, randomly selecting a point from the input data point set D as a first clustering center u1；
Step 2.3, for each point in the dataset
Figure FDA0002824959810000053
Calculating the distance to the nearest cluster center in the selected cluster centers:
Figure FDA0002824959810000054
wherein the content of the first and second substances,
Figure FDA0002824959810000055
are data points
Figure FDA0002824959810000056
Distance to the nearest cluster center; u. ofkIs the k-th cluster center;
step 2.4, selecting a new data point as a new clustering center according to the following selection principle:
Figure FDA0002824959810000057
wherein u isk+1To find the (k + 1) th cluster center,
Figure FDA0002824959810000058
representing the data point corresponding to the maximum cluster;
step 2.5, repeating the step 2.3 and the step 2.4 until K clustering centers are selected;
the initial clustering center can be optimized through the steps;
step 2.6, calculate all sample points
Figure FDA0002824959810000059
To each cluster center u ═ u { [ u ]1,u2,…,uKDividing the distance into K clusters according to the nearest principle;
step 2.7, according to the formula
Figure FDA00028249598100000510
Updating K clustering centers;
wherein u iskIs the k-th cluster center; numkRepresenting the number of sample set data points of the kth cluster;
Figure FDA00028249598100000511
is the n-th in the k-th classkUser feature data vector K1, 2, …, K, nk＝1,2,…,numkK is the number of clusters;
step 2.8, repeating steps 2.6 and 2.7 until the maximum Iteration number Iteration is 500 or the clustering center u is not changed any more; at this time, updated K clusters c are obtained1、c2、…、cKThe cluster centers are respectively u1、u2、…、uKThe number of data points in each cluster is num1、num2、…、numK
Defining the cluster center of the cluster with the maximum data point number as a temperature control load user reference point:
Figure FDA0002824959810000061
wherein: kbasic＝argMax(numk) Indicating the cluster number containing the most data points; cp (p)basicIs the temperature turning point of the reference load model; alpha is alphabasicIs the base load value of the reference load model; beta is abasicIs the regression intercept of the reference load model; gamma raybasicIs the regression slope of the reference load model; rbasicIs a reference load model regression coefficient;
step 2, calculating the user reference load of the temperature control load as follows:
user temperature control user reference load may be defined as:
Wtclbasic,i＝[βbasic+γbasic*(Toutbasic,i-cpbasic)-αbasic]*Sbasic,i
wherein:
Figure FDA0002824959810000062
Toutbasic,ithe average temperature of the area where the temperature control load user is located in i hours is represented; n is the number of users with temperature control load;
Figure FDA00028249598100000612
is m atzOutdoor temperature of the ith hour of the individual user;
Sbasic,ithe definition is as follows:
Figure FDA0002824959810000063
wherein Toutbasic,iThe average temperature of the area where the temperature control load user is located in the ith hour is represented; cp (p)basicIs the temperature turning point of the reference load model;
and obtaining the user reference load of the temperature control load according to the parameters and the steps.
4. The demand response potential analysis method applying the smart meter data according to claim 1, wherein:
and 3, calculating the refrigerating capacity of the user per hour under the unit power of the temperature control load as follows:
Figure FDA0002824959810000064
wherein z ∈ [1, N ]]N is the number of users with temperature control load, I belongs to [1, I ∈]And I is the total number of hours,
Figure FDA0002824959810000065
is m atzThe refrigerating capacity of the temperature control load per unit power at the ith hour of each user,
Figure FDA0002824959810000066
for user mzThe energy efficiency ratio of the temperature control load,
Figure FDA0002824959810000067
representative user mzThe temperature controlled load operation state at the ith hour,
Figure FDA0002824959810000068
for user mzThe temperature control load rated power of (1);
step 3, further calculating the indoor temperature of the user per hour as follows:
Figure FDA0002824959810000069
wherein the content of the first and second substances,
Figure FDA00028249598100000610
is m atzThe indoor temperature of the i-th hour of the individual user,
Figure FDA00028249598100000611
is m atzUser' sThe outdoor temperature at the i-th hour,
Figure FDA0002824959810000071
is m atzThe equivalent impedance of the house heat exchange of an individual user,
Figure FDA0002824959810000072
is m atzThe house equivalent specific heat capacity of each user,
Figure FDA0002824959810000073
is m atzRefrigerating capacity of the temperature control load at the ith hour of each user under unit power;
and 3, calculating the real-time power of the temperature control load of each hour of the user as follows:
Figure FDA0002824959810000074
wherein the content of the first and second substances,
Figure FDA0002824959810000075
m thzThe temperature control load real-time power of the ith hour of each user,
Figure FDA0002824959810000076
representative user mzThe temperature controlled load operation state at the ith hour,
Figure FDA0002824959810000077
for user mzThe temperature control load rated power of (1);
step 3, further calculating the real-time energy consumption cost of the temperature control load per hour of the user as follows:
Figure FDA0002824959810000078
wherein the content of the first and second substances,
Figure FDA0002824959810000079
is m atzThe real-time energy consumption cost of the temperature control load of the ith hour of each user,
Figure FDA00028249598100000710
m thzTemperature controlled load real time power, ToU, for the ith hour of an individual useriThe time-of-use electricity price of the ith hour;
step 3, selecting the indoor temperature of each hour of the user and the real-time energy consumption cost of the temperature control load of each hour of the user as evaluation indexes, and defining a comment subset and the weight of the evaluation indexes;
through the m < th > mzIndoor temperature of individual user, mthzEstablishing an evaluation index of the electric charge of each user;
the evaluation indexes are as follows:
Figure FDA00028249598100000711
wherein, UmzIs m atzSet of fuzzy variables of individual users, first fuzzy variable
Figure FDA00028249598100000712
Is m atzIndoor temperature of individual user, second fuzzy variable
Figure FDA00028249598100000713
Is m atzThe electricity charge of the individual user;
step 3, the comment subset is:
v1{ "cold", "comfort", "hot" }
v2As { "cheap", "moderate", "expensive" }
Wherein v is1Denotes the m-thzSubset of comments on the indoor temperature of an individual user, v2Denotes the m-thzA comment subset of the individual user's electricity charge;
and 3, the weight vector of the evaluation index is as follows:
A＝{a1,a2}
wherein, a1Denotes the m-thzComment weight of indoor temperature of individual user, a2Represents the m-thzThe comment weight of the electric charge of each user;
the human thermal comfort index is calculated as follows:
Figure FDA0002824959810000081
wherein Tin is the ambient temperature, and PMV is the index of human thermal comfort;
the text takes the elasticity range value of the human comfort level index: PMVa＞PMVb＞PMVc＞PMVdAt this time, the corresponding indoor ambient temperature is Ta,Tb,Tc,Td；
In the blurring process, for a temperature at Tb～TcIs absolutely comfortable; and for a temperature range of Ta～Tb,Tc～TdFor relative comfort, i.e. not completely belonging to the "comfort" concept, the model uses a trapezoidal function to quantify the degree of membership:
the fuzzy membership function model is as follows:
Figure FDA0002824959810000082
Figure FDA0002824959810000083
Figure FDA0002824959810000084
wherein the content of the first and second substances,
Figure FDA0002824959810000085
is m atzH fuzzy variable for each user to comment set vhThe membership degree of the g-th comment, h is 1,2, g is 1,2, 3;
Figure FDA0002824959810000086
is m atzH fuzzy variable of each user; a, b, c and d represent the upper and lower limits of the membership function;
when the fuzzy variable is the indoor temperature
Figure FDA0002824959810000087
The upper and lower bounds of the fuzzy membership function respectively correspond to Ta,Tb,Tc,Td(ii) a At this point, the comment subset v can be accessed1{ "Cold", "comfort", "Heat" } and fuzzy variables
Figure FDA0002824959810000088
Establishing a membership relationship;
similarly, when the fuzzy variable is the electricity charge, the comment subset v can be selected2Membership of { "cheap", "moderate", "expensive" } to the fuzzy variable electricity rate, i.e.
Figure FDA0002824959810000089
Evaluation vector calculation processing
Using fuzzy synthesis operator to make the weight vector A of evaluation index and evaluation matrix RmSynthesizing to obtain an evaluation result vector B:
Figure FDA0002824959810000091
wherein
Figure FDA0002824959810000092
Represents the fuzzy synthesis operator and the fuzzy synthesis operator,
Figure FDA0002824959810000093
in order to take the value of the weight,
Figure FDA0002824959810000094
is fuzzy membership;
processing the fuzzy comprehensive evaluation vector by a weighted average principle to obtain a user satisfaction quantification result:
Figure FDA0002824959810000095
wherein the content of the first and second substances,
Figure FDA0002824959810000096
is m atzSatisfaction of the individual user; bm,gIs m atzThe g membership value of each user; num is a natural number, and the purpose is to control the weight distribution caused by larger membership;
and 3.3, optimizing the comprehensive satisfaction degree of the user by taking one day as a time unit, and outputting an optimization objective function of each day as follows:
Figure FDA0002824959810000097
wherein: j ═ D × 24+1, D × 24+2, …, D × 24+ 24; d-1, 2, …, Mod [ I/24 ]]Represents the jth time in 1-24 of day D, and the total Data is mod [ I/24 ]]Day;
Figure FDA0002824959810000098
Figure FDA0002824959810000099
for user mzAt time j on day D the room temperature was
Figure FDA00028249598100000910
The satisfaction of (2);
Figure FDA00028249598100000911
for user mzThe indoor temperature possible value at time j on day D,
al＝1,2,…,(Tinm,j-1-Toutm,j) 0.1, i.e. all possible room temperatures are traversed by a division value of 0.1, i.e.:
Figure FDA00028249598100000912
for the user, only the above sum (Tin)m,j-1-Toutm,j) The/0.1 set temperature can be selected, so the optimization problem is to traverse all the selected temperatures, and the aim of simulating the decision of the user is achieved by optimizing the comprehensive satisfaction;
through the satisfaction optimizing model, the user mzThe decision result is:
Figure FDA00028249598100000913
i.e. user mzIndoor temperature was set at time j on day D
Figure FDA00028249598100000914
The comprehensive satisfaction is optimal;
bring this temperature value into user mzTemperature controlled load thermoelectric model, available user mzEnergy consumption at time j on day D, recorded as
Figure FDA00028249598100000915
5. The demand response potential analysis method applying the smart meter data according to claim 1, wherein:
the step 4 is as follows:
the reference load of the user with the temperature control load obtained in the step 1 at the moment i is Wtclbasic,i(ii) a From step 2.3, user m is availablezEnergy consumption at time j on day D of
Figure FDA0002824959810000101
Evaluating the demand response potential of a temperature control load user by taking days as time length; user mzThe response potential on day D was:
Figure FDA0002824959810000102
wherein the content of the first and second substances,
Figure FDA0002824959810000103
for user mzCorresponding potential at day D; wtclbasic,(D-1)*24+jThe reference load at the j moment on the D day;
Figure FDA0002824959810000106
) For user mzEnergy consumption after demand response optimization at day D, time j;
the response potential of the temperature control load user group on day D is:
Figure FDA0002824959810000104
wherein, DRP∑,DResponse potential for the temperature controlled load user population on day D;
Figure FDA0002824959810000105
for user mzCorresponding potential at day D.
CN202011447244.6A 2020-12-09 2020-12-09 Demand response potential analysis method applying intelligent ammeter data Active CN112465385B (en)

Priority Applications (1)

Application Number Priority Date Filing Date Title
CN202011447244.6A CN112465385B (en) 2020-12-09 2020-12-09 Demand response potential analysis method applying intelligent ammeter data

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
CN202011447244.6A CN112465385B (en) 2020-12-09 2020-12-09 Demand response potential analysis method applying intelligent ammeter data

Publications (2)

Publication Number Publication Date
CN112465385A true CN112465385A (en) 2021-03-09
CN112465385B CN112465385B (en) 2023-12-01

Family

ID=74801537

Family Applications (1)

Application Number Title Priority Date Filing Date
CN202011447244.6A Active CN112465385B (en) 2020-12-09 2020-12-09 Demand response potential analysis method applying intelligent ammeter data

Country Status (1)

Country Link
CN (1) CN112465385B (en)

Cited By (5)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN113627741A (en) * 2021-07-20 2021-11-09 国网湖南省电力有限公司 Comprehensive evaluation method and device for operation state of charging pile electric energy metering system
CN113822482A (en) * 2021-09-24 2021-12-21 内蒙古电力(集团)有限责任公司内蒙古电力科学研究院分公司 Method and device for establishing load prediction model of comprehensive energy system
CN115330280A (en) * 2022-10-14 2022-11-11 国网山东省电力公司营销服务中心(计量中心) Method and system for evaluating adjustable potential of air conditioner load demand response in aggregated load
CN116089847A (en) * 2023-04-06 2023-05-09 国网湖北省电力有限公司营销服务中心(计量中心) Distributed adjustable resource clustering method based on covariance agent
CN117762106A (en) * 2023-12-23 2024-03-26 济宁市铠铠食品有限公司 Method for monitoring processing process of poultry blood product based on Internet of things

Citations (5)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN108593990A (en) * 2018-06-04 2018-09-28 国网天津市电力公司 A kind of stealing detection method and application based on electric power users electricity consumption behavior pattern
CN110580585A (en) * 2019-09-11 2019-12-17 东南大学 Power consumer clustering power consumption behavior characteristic analysis method based on load decomposition
CN111461761A (en) * 2020-02-29 2020-07-28 国网江苏省电力有限公司苏州供电分公司 Resident user portrait method based on multi-dimensional fine-grained behavior data
CN111754029A (en) * 2020-06-08 2020-10-09 深圳供电局有限公司 Community load prediction system
CN111967723A (en) * 2020-07-24 2020-11-20 南昌大学 User peak regulation potential analysis method based on data mining

Patent Citations (5)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN108593990A (en) * 2018-06-04 2018-09-28 国网天津市电力公司 A kind of stealing detection method and application based on electric power users electricity consumption behavior pattern
CN110580585A (en) * 2019-09-11 2019-12-17 东南大学 Power consumer clustering power consumption behavior characteristic analysis method based on load decomposition
CN111461761A (en) * 2020-02-29 2020-07-28 国网江苏省电力有限公司苏州供电分公司 Resident user portrait method based on multi-dimensional fine-grained behavior data
CN111754029A (en) * 2020-06-08 2020-10-09 深圳供电局有限公司 Community load prediction system
CN111967723A (en) * 2020-07-24 2020-11-20 南昌大学 User peak regulation potential analysis method based on data mining

Non-Patent Citations (2)

* Cited by examiner, † Cited by third party
Title
罗金满;温兆聪;董文杰;李兆伟;刘卓贤;吴浩天;: "基于居民用户画像的综合能源需求响应潜力量化模型", 可再生能源, no. 10 *
谢敦见: ""温控负荷的需求响应潜力评估及其协同优化管理研究"", 《中国优秀博硕士学位论文全文数据库(硕士) 工程科技Ⅱ辑》 *

Cited By (8)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN113627741A (en) * 2021-07-20 2021-11-09 国网湖南省电力有限公司 Comprehensive evaluation method and device for operation state of charging pile electric energy metering system
CN113627741B (en) * 2021-07-20 2023-12-12 国网湖南省电力有限公司 Comprehensive evaluation method and device for operation state of charging pile electric energy metering system
CN113822482A (en) * 2021-09-24 2021-12-21 内蒙古电力(集团)有限责任公司内蒙古电力科学研究院分公司 Method and device for establishing load prediction model of comprehensive energy system
CN115330280A (en) * 2022-10-14 2022-11-11 国网山东省电力公司营销服务中心(计量中心) Method and system for evaluating adjustable potential of air conditioner load demand response in aggregated load
CN115330280B (en) * 2022-10-14 2023-02-21 国网山东省电力公司营销服务中心(计量中心) Method and system for evaluating adjustable potential of air conditioner load demand response in aggregated load
CN116089847A (en) * 2023-04-06 2023-05-09 国网湖北省电力有限公司营销服务中心(计量中心) Distributed adjustable resource clustering method based on covariance agent
CN116089847B (en) * 2023-04-06 2023-06-23 国网湖北省电力有限公司营销服务中心(计量中心) Distributed adjustable resource clustering method based on covariance agent
CN117762106A (en) * 2023-12-23 2024-03-26 济宁市铠铠食品有限公司 Method for monitoring processing process of poultry blood product based on Internet of things

Also Published As

Publication number Publication date
CN112465385B (en) 2023-12-01

Similar Documents

Publication Publication Date Title
CN112465385A (en) Demand response potential analysis method applying intelligent electric meter data
Lindberg et al. Modelling electric and heat load profiles of non-residential buildings for use in long-term aggregate load forecasts
Yildiz et al. A review and analysis of regression and machine learning models on commercial building electricity load forecasting
Zhang et al. An optimal and learning-based demand response and home energy management system
Platt et al. Adaptive HVAC zone modeling for sustainable buildings
Faruqui et al. Dynamic pricing of electricity in the mid-Atlantic region: econometric results from the Baltimore gas and electric company experiment
Yang et al. A new method to develop typical weather years in different climates for building energy use studies
Jawad et al. Machine learning based cost effective electricity load forecasting model using correlated meteorological parameters
Zhou et al. Benchmarking air-conditioning energy performance of residential rooms based on regression and clustering techniques
CN110298487A (en) It is a kind of for meeting the room temperature prediction technique of users &#39; individualized requirement
Pedersen Use of different methodologies for thermal load and energy estimations in buildings including meteorological and sociological input parameters
Obringer et al. Evaluating the climate sensitivity of coupled electricity-natural gas demand using a multivariate framework
da Fonseca et al. Thermal comfort maintenance in demand response programs: A critical review
Swan Residential sector energy and GHG emissions model for the assessment of new technologies
Yan et al. A study on temperature-setting behavior for room air conditioners based on big data
Zhou et al. Comparison of different machine learning algorithms for predicting air-conditioning operating behavior in open-plan offices
CN112669173A (en) Short-term load prediction method based on multi-granularity features and XGboost model
Albert et al. Building dynamic thermal profiles of energy consumption for individuals and neighborhoods
CN106570729A (en) Air conditioner reliability influence factor-based regional clustering method
Zhang et al. Effect of physical, environmental, and social factors on prediction of building energy consumption for public buildings based on real-world big data
Li et al. Exploring potential of energy flexibility in buildings for energy system services
Lutzenhiser et al. The “Average American” unmasked: Social structure and differences in household energy use and carbon emissions
Shi et al. Predicting household energy consumption in an aging society
CN117132334A (en) Intelligent pricing method, system, electronic device and storage medium for electric power retail
Middelhauve On the role of districts as renewable energy hubs

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