CN106203683A - A kind of modeling method of power customer electro-load forecast system - Google Patents
A kind of modeling method of power customer electro-load forecast system Download PDFInfo
- Publication number
- CN106203683A CN106203683A CN201610497287.2A CN201610497287A CN106203683A CN 106203683 A CN106203683 A CN 106203683A CN 201610497287 A CN201610497287 A CN 201610497287A CN 106203683 A CN106203683 A CN 106203683A
- Authority
- CN
- China
- Prior art keywords
- day
- prediction
- load
- error
- power load
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Pending
Links
- 238000000034 method Methods 0.000 title claims abstract description 34
- 239000011159 matrix material Substances 0.000 claims abstract description 22
- 238000012937 correction Methods 0.000 claims abstract description 21
- 238000012546 transfer Methods 0.000 claims abstract description 17
- 230000008878 coupling Effects 0.000 claims abstract description 3
- 238000010168 coupling process Methods 0.000 claims abstract description 3
- 238000005859 coupling reaction Methods 0.000 claims abstract description 3
- 238000000513 principal component analysis Methods 0.000 claims description 19
- 230000003203 everyday effect Effects 0.000 claims description 16
- 230000002068 genetic effect Effects 0.000 claims description 15
- 238000013528 artificial neural network Methods 0.000 claims description 14
- 230000002354 daily effect Effects 0.000 claims description 11
- 238000005457 optimization Methods 0.000 claims description 7
- 230000007935 neutral effect Effects 0.000 description 11
- 238000010276 construction Methods 0.000 description 8
- 230000005611 electricity Effects 0.000 description 7
- 239000000470 constituent Substances 0.000 description 6
- 238000005094 computer simulation Methods 0.000 description 3
- 238000011161 development Methods 0.000 description 3
- 241000196324 Embryophyta Species 0.000 description 2
- 238000005516 engineering process Methods 0.000 description 2
- 238000010353 genetic engineering Methods 0.000 description 2
- 238000012549 training Methods 0.000 description 2
- 235000006508 Nelumbo nucifera Nutrition 0.000 description 1
- 240000002853 Nelumbo nucifera Species 0.000 description 1
- 235000006510 Nelumbo pentapetala Nutrition 0.000 description 1
- 208000024780 Urticaria Diseases 0.000 description 1
- 230000003044 adaptive effect Effects 0.000 description 1
- 238000000540 analysis of variance Methods 0.000 description 1
- 230000015572 biosynthetic process Effects 0.000 description 1
- 239000003245 coal Substances 0.000 description 1
- 230000001186 cumulative effect Effects 0.000 description 1
- 230000007812 deficiency Effects 0.000 description 1
- 238000004519 manufacturing process Methods 0.000 description 1
- 230000035772 mutation Effects 0.000 description 1
- 238000003062 neural network model Methods 0.000 description 1
- 238000004321 preservation Methods 0.000 description 1
- 238000012545 processing Methods 0.000 description 1
- 238000004064 recycling Methods 0.000 description 1
- 238000011160 research Methods 0.000 description 1
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
- G06Q10/04—Forecasting or optimisation specially adapted for administrative or management purposes, e.g. linear programming or "cutting stock problem"
-
- 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)
- Economics (AREA)
- Human Resources & Organizations (AREA)
- Strategic Management (AREA)
- Theoretical Computer Science (AREA)
- Physics & Mathematics (AREA)
- Health & Medical Sciences (AREA)
- Marketing (AREA)
- General Physics & Mathematics (AREA)
- General Business, Economics & Management (AREA)
- Tourism & Hospitality (AREA)
- Quality & Reliability (AREA)
- Game Theory and Decision Science (AREA)
- Operations Research (AREA)
- Development Economics (AREA)
- Entrepreneurship & Innovation (AREA)
- Public Health (AREA)
- Water Supply & Treatment (AREA)
- General Health & Medical Sciences (AREA)
- Primary Health Care (AREA)
- Management, Administration, Business Operations System, And Electronic Commerce (AREA)
- Supply And Distribution Of Alternating Current (AREA)
Abstract
The invention discloses the modeling method of a kind of power customer electro-load forecast, including: the power load of the history day of several years in past is hived off, is divided into set on working day and nonworkdays set；Respectively the power load of set on working day and nonworkdays set is modeled according to grouping result；To prediction day by whether working day mates with grouping result, with the model of coupling correspondence set, prediction daily load is predicted, obtains preliminary forecasting result；Also include utilizing markov to ask for forecast error transfer matrix, obtain predicting day error correction values；It is worth to finally predict the outcome by preliminary forecasting result and prediction day error correction.The present invention realizes the prediction to power load, and the investment planning for electrical network provides foundation, improves network operation efficiency.
Description
Technical field
The present invention relates to electricity demand forecasting field, more particularly, it relates to a kind of power customer electro-load forecast system
Modeling method.
Background technology
Accurate Prediction power consumption is the basis of Operation of Electric Systems and planning, is that electric power enterprise formulates the placing plan, operation
The basis of strategy.As a kind of specialty goods, the production of electric energy, carry, distribute and use have compared with other industrial products bright
Aobvious different feature.Electric power can directly cannot store as the resource such as coal, oil, and it produces, transport and consume can only be simultaneously
Carrying out, glut can cause the huge wasting of resources, supplies deficiency and then can have a strong impact on daily life and bring
Economic loss.In addition, huge investment makes the construction of electrical network have certain time stickiness.Therefore, this will be realistic existing to not
Carry out the Accurate Prediction of power consumption, provide foundation for future electrical energy and electric grid investment planning construction, make power construction meet national warp
Ji development and the needs of people's lives.
Summary of the invention
In view of this, the invention provides the system modeling method of a kind of power customer electro-load forecast, it is achieved to
The prediction of electric load, the investment planning for electrical network provides foundation, improves network operation efficiency.
To achieve these goals, it is proposed that scheme as follows:
The modeling method of a kind of power customer electro-load forecast, it is characterised in that including:
The power load of the history day of several years in past is hived off, is divided into set on working day and nonworkdays set；
Respectively the power load of set on working day and nonworkdays set is modeled according to grouping result；
To prediction day by whether working day mates with grouping result, with the model of coupling correspondence set, prediction day is born
Lotus is predicted, and obtains preliminary forecasting result；
In a preferred embodiment of the invention, also include utilizing markov to ask for forecast error transfer matrix,
To prediction day error correction values；
It is worth to finally predict the outcome by preliminary forecasting result and prediction day error correction.
In a preferred embodiment of the invention, the power load of set on working day and nonworkdays set is entered respectively
Row modeling, with some after principal component analysis of the temperature of every day in each set, neighbouring daily load curve and neighbouring degree/day
New feature is as input, with the load curve on the same day for output, uses genetic algorithm optimization BP neural network to carry out model construction.
In a preferred embodiment of the invention, described employing genetic algorithm optimization BP neural network carries out model construction
During also include modified weight method: use back-propagation algorithm that weights and the deviation of network are carried out training repeatedly and adjusted
Whole, make the vector of output and Mean Vector close to, when the error sum of squares of network output layer is less than the error specified
Shi Xunlian completes, and preserves weights and the deviation of network.
In a preferred embodiment of the invention, to each set with the temperature of every day, neighbouring daily load curve and neighbour
Recently the temperature some new features after principal component analysis are as input, including:
To when under mean daily temperature and same data acquisition system current record close on three days mean temperatures, and same data
Three day 24 hours every day of interior 24 power load of closing on of the lower current record of set carries out principal component analysis, by final some
New feature is as mode input, to reduce variable dimension.
In a preferred embodiment of the invention, described preliminary forecasting result includes:
Judge to predict whether day is working day, if then matching set on working day working day, if nonworkdays is then
It is fitted on nonworkdays set；
In the forecast model of prediction day place corresponding set, input as under prediction mean daily temperature and same data acquisition system
Current record close on three days mean temperatures, and under same data acquisition system current record to close on three day 24 hours every day interior
Some new features that 24 power loads obtain after carrying out principal component analysis, are output as predicting the load curve of day.
In a preferred embodiment of the invention, the described absolute error correction utilizing markov to obtain prediction day
Value, specifically includes following steps:
Obtain absolute error sequence samples；
Utilize the mean variance method absolute error sequence state interval division to load prediction results；
Calculate Markov forecast techniques error state probability transfer matrix；
Calculate the state probability of future time instance forecast error according to state probability transfer matrix, obtain error prediction value.
In a preferred embodiment of the invention, described by preliminary forecasting result and prediction day error correction be worth to
Predict the outcome eventually, predict that the BP neutral net that genetic algorithm is improved by day error correction values is calculated with specific reference to what markov obtained
The initial predicted result that method obtains is modified, and obtains final load prediction curve.
In a preferred embodiment of the invention, the several years are preferably 1 year, i.e. the electricity consumption to 1 year history day of past
Load includes by working day carries out grouping method:
Obtain user's day power load of a year in the past；
Obtain date and time information every day；
Being then to divide the power load set under working day working day into, weekend, country's legal festivals and holidays then divide inoperative into
Power load set under.
Through as shown from the above technical solution, first user's history power load is divided into set on working day with non-by the method
Working day gathers, and utilizes genetic algorithm optimization BP neural network algorithm to set up the use of each set after using principal component analysis dimensionality reduction
Electric load model；Then, by prediction day by whether working day matches corresponding history electricity consumption set, according to history power load collection
The model closed obtains predicting the initial predicted value of day power load；Recycling markov asks for error transfer matrix, obtains pre-
Survey the error correction values of day；It is worth to predict the final electro-load forecast of day finally according to initial predicted value and error correction
Value.Compared with prior art, the present invention combines principal component analysis, the BP neutral net of improvement and markov, it is proposed that one
Plant the system modeling method being applicable to power customer electro-load forecast, it is possible to realize customer electricity load prediction, for not sending a telegram here
Power and electric grid investment planning construction provide foundation, make power construction meet the needs of the national economic development and people's lives.
Accompanying drawing explanation
In order to be illustrated more clearly that the embodiment of the present invention or technical scheme of the prior art, below will be to embodiment or existing
In having technology to describe, the required accompanying drawing used is briefly described, it should be apparent that, the accompanying drawing in describing below is only this
Inventive embodiment, for those of ordinary skill in the art, on the premise of not paying creative work, it is also possible to according to
The accompanying drawing provided obtains other accompanying drawing.
Fig. 1 shows that disclosed in one embodiment of the invention, a kind of system being applied to power customer electro-load forecast is built
The schematic flow sheet of mould method；
Fig. 2 is principal component analysis dimensionality reduction flow chart disclosed in the embodiment of the present invention；
Fig. 3 is BP neural network algorithm modeling structure figure disclosed in the embodiment of the present invention；
Fig. 4 is that genetic algorithm disclosed in the embodiment of the present invention determines BP neural network algorithm weights flow chart；
Fig. 5 is markov error correction flow chart disclosed in the embodiment of the present invention；
Detailed description of the invention
Below in conjunction with the accompanying drawing in the embodiment of the present invention, the technical scheme in the embodiment of the present invention is carried out clear, complete
Describe, it is clear that described embodiment is only a part of embodiment of the present invention rather than whole embodiments wholely.Based on
Embodiment in the present invention, it is every other that those of ordinary skill in the art are obtained under not making creative work premise
Embodiment, broadly falls into the scope of protection of the invention.
A kind of flowchart of the electricity sales amount Forecasting Methodology that Fig. 1 provides for the embodiment of the present invention, including:
Step S11: to the power load of 1 year history day of past by whether working day hives off；
Obtain the date and time information of 1 year every day in the past, determine that every day is working day or nonworkdays, wherein, nonworkdays
Including weekend and festivals or holidays, the history power load whole year is divided into two set, set on working day and nonworkdays set.
Step S12: the power load of each set is modeled according to grouping result；
Power load influence factor is numerous, it will be carried out principal component analysis, reduce data dimension, finally determine that model is defeated
Enter under prediction day identity set first three day every day 24 load point, first three day mean temperature under prediction same day day and identity set
Some new features after principal component analysis, are output as predicting 24 power load values of day.
After determining input and output, to the historical data in two set, use genetic algorithm optimization BP neural network algorithm
It is modeled.
Step S13: to prediction day by whether working day mates with grouping result, to mate the model of set to prediction
Daily load is predicted, and obtains preliminary forecasting result.
Judge to predict whether day is working day, and prediction day is referred to corresponding power load type, by putting down of prediction day
All temperature, prediction day place lower first three day power load curve of set and the mean temperature some new features after principal component analysis
As mode input, output corresponding prediction day power load initial predicted value.
Step S14: utilize markov to ask for forecast error transfer matrix, obtains predicting day error correction values.
The forecast error sequence of partial history is divided into n interval, obtains error transfer matrix, and then ask for predicting day
Error correction values.
Step S15: be worth to finally predict the outcome by preliminary forecasting result and prediction day error correction.
The BP neural network algorithm that genetic algorithm is improved by the prediction day error correction values obtained according to markov obtains
Initial predicted result be modified, obtain final load prediction curve.
In above-described embodiment, optionally, power load influence factor's principal component analysis, may include that
Stronger dependency is there is, it is necessary to carry out principal component analysis between each variable factors relevant with power load.
Research to a certain problem relates to k index, total n sample, and the sample matrix x that observation draws is n × k dimension.K in the present invention
Individual index includes predicting the mean temperature of day, prediction day place lower first three day power load curve of set and mean temperature, chooses
The data in past 1 year some days are as sample.It is standardized original matrix x processing, eliminates between index variable due to quantity
The difference of level and the impact that produces.
In above-described embodiment, optionally, power load influence factor principal component analysis flow chart is as in figure 2 it is shown, can wrap
Include:
Step S21: according to normalized matrix x1,x2,x3,,...xk, calculate the correlation matrix R of sample；
Step S22: seek k eigenvalue λ of correlation matrix1,λ2,λ3,...λkWith corresponding characteristic vector e1,e2,
e3,...ek；
Step S23: seek the variance contribution ratio of each main constituent, calculates cumulative proportion in ANOVA, screens main constituent.
The accumulative variance contribution ratio of current m main constituent touches the mark the requirement of message reflection precision, and generally 85%, ask
Obtain m main constituent y1,y2,...ykReplace original variable, using m main constituent as the input of neural network model.
Main constituent expression formula is
In formula: ei=[ei1ei2...eik], eikK dimension corresponding to the ith feature value of the correlation matrix of original variable
Characteristic vector；X is the initial input variable of k dimension, x=[x1x2...xk]T。
In above-described embodiment, optionally, use genetic algorithm optimization BP neural network algorithm to be modeled, may include that
BP neural network algorithm is utilized to set up electro-load forecast model process as it is shown on figure 3, may include that
Step S31: input block.
Before gathering lower first three day every day of 24 load point, prediction same day day with prediction day place and predict that day place set is lower
Three day every day, the mean temperature some new features after principal component analysis were as the BP neutral net input after improving；
Step S32: modified weight unit.
Use back-propagation algorithm that weights and the deviation of network are carried out training repeatedly and adjusted, make vector and the phase of output
Hope vector close to, when network output layer error sum of squares less than specify error time train, preservation network
Weights and deviation；
Step S33: output unit.
With prediction day 24 power load values for output.
In above-described embodiment, optionally, modified weight unit, may include that
First by genetic algorithm, neutral net initial weight is optimized, solution space is oriented one and preferably searches
Rope space.Use genetic algorithm to be trained BP neutral net optimizing weights step as shown in Figure 4, may include that
Step S41: individual UVR exposure and the initialization of population.
Individuality contains all weights and the threshold value of whole BP neutral net.Herein individuality is used the mode of real coding
Encode.Code length is:
S=m × v+v × l+v+l
Wherein, v is node in hidden layer；M is input layer number；L is output layer nodes.
Step S42: individuality is evaluated according to fitness function value.
Fitness function is set as the inverse of neutral net error sum of squares:
Wherein, SE is the error sum of squares between prediction output and the desired output of neutral net.Each individuality is carried out
Decoding obtains one group of BP neural network weight threshold value, calculates the SE of BP neutral net, then calculates according to fitness function
The adaptive value of each individuality.
Step S43: select, intersect, mutation genetic operation.
Step S44: reach maximum evolutionary generation, or error is less than setting value, jumps out algorithm.
After genetic manipulation completes, it is taken in whole genetic manipulation the optimum individual obtained and weighs as the initial of neutral net
Value, then use BP neutral net to be trained, calculate its error, and constantly revise its weight threshold, until meeting required precision.
In above-described embodiment, optionally, Markov model correction flow chart is as it is shown in figure 5, may include that
The most basic feature of Markov chain is under the conditions of the state of system " now " is known, the state of its " in the future "
Unrelated with the state in " past ".If the time series of certain things or certain phenomenon with various state can be considered that Ma Er can
Husband's chain, then the state according to the n moment can dope the state in n+1 moment, here it is application Markov-chain model solves each
Plant the basic thought of forecasting problem.Being predicted error correction with markov in the present invention, concrete steps are as shown in Figure 5.
Step S51: obtain absolute error sequence samples.
After choosing sample, determine the electro-load forecast result absolute error sequence of every dayWherein Pn
For actual power load sequence,For prediction power load sequence, it it is all the column vector of 24 dimensions.
Step S52: utilize the mean variance method absolute error sequence state interval division to load prediction results；
For absolute error sequence δ predicted the outcomenIts average isMean square deviation is s, utilizes average-all
After variance staging, typically sequence can be divided into 5 grades: Wherein a1,a4Value exists
Value in [1.0,1.5], a2,a3Value in [0.3,0.6].
Step S53: calculate Markov forecast techniques error state probability transfer matrix；
According to Markov Theory, it is considered to the absolute error data of sample, have multiple error state every day it may happen that,
If error is at EiIn the range of, then event is in state Ei；Error was from the E of first dayiBecome the E of second dayjProbability, the most just
It it is state EiE is become through 1 stepjProbability be:
In formula:For forecast error in sample from EiOne step transfers to EjTransfer number；NiFor state EiOccur is total
Number of times (if), then 1 step state probability transfer matrix is
State probability transfer matrix is a n rank square formation, has two features:
i.Matrix each element nonnegativity；
ii.I.e. matrix often row sum is 1.
And k walks state probability transfer matrix P(k)=(P(1))k
Step S54: calculate the state probability of future time instance forecast error according to state probability transfer matrix, obtain error
Predictive value.
Choose electro-load forecast absolute error δ when the day before yesterday (n)nThe state at place, as original state, is asked for treating pre-
Survey the state vector of day, initial state vector Vn。
The state vector of the electro-load forecast absolute error of (n+1) day to be predicted
Vn+1=VnP(1)
The state vector of the electro-load forecast absolute error of (n+2) day to be predicted
Vn+2=VnP(2)
Find V respectivelyn+1Or Vn+2The value of middle maximum probability:
A) position finding the value place of a maximum probability is i.e. the interval residing for error to be predicted.Take the equal of interval
Value is i.e. the predictive value δ of errorn+1And δn+2, E1TakeHalf, E5In like manner.
If b) value of maximum probability is more than 1, then the predictive value taking error is the equal of these maximum correspondences intervals
Value.
Such as, the absolute error initial value that predicts the outcome on January 1 is δ1, state vector is V1, then the absolute of January 2 misses
Difference state vector V2=V1P, vector V2The position at maximum value place is i.e. the interval residing for absolute error on January 2, January 3
The absolute error of day is vector V3=V1P(2), vector V3The position at maximum value place is i.e. residing for the absolute error on January 3
Interval, the average taking interval is January 2 and the predictive value δ of absolute error on January 32,δ3。
As seen from the above embodiment: the invention discloses a kind of system modelling being applied to power customer electro-load forecast
Method.First user's history power load is divided into two classes by the method, is class and nonworkdays class on working day respectively, then uses
What genetic algorithm improved BP algorithm determined each class uses electric model.And then, will prediction day according to whether working day
It is fitted on corresponding history electricity consumption type.The power load of day is predicted in model prediction further according to history electricity consumption type.The present invention adopts
Determine mode input with principal component analysis, utilize the BP neural network algorithm improved to determine forecast model, and by markov
Model prediction result is carried out error correction, it is possible to realize power user consumption load prediction, advise for future electrical energy and electric grid investment
Draw to build and foundation is provided, make power construction meet the needs of the national economic development and people's lives.
Claims (9)
1. the modeling method of a power customer electro-load forecast, it is characterised in that including:
The power load of the history day of several years in past is hived off, is divided into set on working day and nonworkdays set；
Respectively the power load of set on working day and nonworkdays set is modeled according to grouping result；
To prediction day by whether working day mates with grouping result, with the model of coupling correspondence set, prediction daily load is entered
Row prediction, obtains preliminary forecasting result.
Method the most according to claim 1, it is characterised in that also include utilizing markov to ask for forecast error transfer square
Battle array, obtains predicting day error correction values；
It is worth to finally predict the outcome by preliminary forecasting result and prediction day error correction.
Method the most according to claim 1, it is characterised in that to set on working day and the power load of nonworkdays set
It is modeled respectively, with the temperature of every day, neighbouring daily load curve and neighbouring degree/day in each set after principal component analysis
Some new features as input, with the load curve on the same day for output, use genetic algorithm optimization BP neural network to carry out mould
Type builds.
Method the most according to claim 3, it is characterised in that described employing genetic algorithm optimization BP neural network carries out mould
Type building process also includes modified weight method: use back-propagation algorithm that weights and the deviation of network are carried out instruction repeatedly
Practice and adjust, make the vector of output and Mean Vector close to, when the error sum of squares of network output layer is less than specifying
Train during error, preserved weights and the deviation of network.
5. according to the method described in claim 3 or 4, it is characterised in that to each set with the temperature of every day, neighbouring daily load
Curve and the neighbouring degree/day some new features after principal component analysis as input, including:
To when under mean daily temperature and same data acquisition system current record close on three days mean temperatures, and same data acquisition system
Three day 24 hours every day of interior 24 power load of closing on of lower current record carries out principal component analysis, by final some new spy
Levy as mode input, to reduce variable dimension.
Method the most according to claim 1 and 2, it is characterised in that described preliminary forecasting result includes:
Judge to predict whether day is working day, if then matching set on working day working day, if nonworkdays then matches
Nonworkdays set；
Prediction day place corresponding set forecast model in, input into prediction mean daily temperature and same data acquisition system under when
Front record close on three days mean temperatures, and under same data acquisition system current record close on three day 24 hours interior 24 every day
Some new features that individual power load obtains after carrying out principal component analysis, are output as predicting the load curve of day.
Method the most according to claim 2, it is characterised in that the described absolute error utilizing markov to obtain prediction day
Correction value, specifically includes following steps:
Obtain absolute error sequence samples；
Utilize the mean variance method absolute error sequence state interval division to load prediction results；
Calculate Markov forecast techniques error state probability transfer matrix；
Calculate the state probability of future time instance forecast error according to state probability transfer matrix, obtain error prediction value.
8. according to the method described in claim 2 or 7, it is characterised in that described by preliminary forecasting result and prediction day error repair
On the occasion of finally being predicted the outcome, predict, with specific reference to what markov obtained, the BP that genetic algorithm is improved by day error correction values
The initial predicted result that neural network algorithm obtains is modified, and obtains final load prediction curve.
Method the most according to claim 1, it is characterised in that the several years are preferably 1 year, i.e. to history day past 1 year
Power load include by working day carries out grouping method:
Obtain user's day power load of a year in the past；
Obtain date and time information every day；
Being then to divide the power load set under working day working day into, weekend, country's legal festivals and holidays then divide under nonworkdays
Power load set.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201610497287.2A CN106203683A (en) | 2016-06-29 | 2016-06-29 | A kind of modeling method of power customer electro-load forecast system |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201610497287.2A CN106203683A (en) | 2016-06-29 | 2016-06-29 | A kind of modeling method of power customer electro-load forecast system |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN106203683A true CN106203683A (en) | 2016-12-07 |
Family
ID=57463426
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201610497287.2A Pending CN106203683A (en) | 2016-06-29 | 2016-06-29 | A kind of modeling method of power customer electro-load forecast system |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN106203683A (en) |
Cited By (11)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN108805388A (en) * | 2018-04-09 | 2018-11-13 | 中国电力科学研究院有限公司 | A kind of determination method and apparatus of non-coming year Load Time Series scene |
| CN109118027A (en) * | 2018-09-30 | 2019-01-01 | 广东小天才科技有限公司 | Electric quantity prediction model training method, electric quantity prediction method, intelligent equipment and system |
| CN109409570A (en) * | 2018-09-20 | 2019-03-01 | 平安科技（深圳）有限公司 | Electronic device, the building day power load prediction technique and storage medium based on clustering algorithm |
| WO2019134254A1 (en) * | 2018-01-02 | 2019-07-11 | 上海交通大学 | Real-time economic dispatch calculation method using distributed neural network |
| TWI705395B (en) * | 2019-06-28 | 2020-09-21 | 國立政治大學 | Customer value management system |
| CN111861587A (en) * | 2020-08-04 | 2020-10-30 | 上海积成能源科技有限公司 | System and method for analyzing residential electricity consumption behavior based on hidden Markov model and forward algorithm |
| CN111915082A (en) * | 2020-08-03 | 2020-11-10 | 深圳集智数字科技有限公司 | Prediction method, prediction device, storage medium and equipment |
| CN112070268A (en) * | 2020-07-31 | 2020-12-11 | 浙江大有实业有限公司杭州科技发展分公司 | Power load prediction method and device based on hotel demand side response |
| CN112766797A (en) * | 2021-01-30 | 2021-05-07 | 广东新华建工程有限公司 | Photoelectric integrated building energy supply method and system |
| CN115619170A (en) * | 2022-10-28 | 2023-01-17 | 北京国电通网络技术有限公司 | Method, device, equipment, computer medium and program product for adjusting electric quantity load |
| CN116402483A (en) * | 2023-06-09 | 2023-07-07 | 国网山东省电力公司兰陵县供电公司 | Online monitoring method and system for carbon emission of park |
-
2016
- 2016-06-29 CN CN201610497287.2A patent/CN106203683A/en active Pending
Cited By (16)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| WO2019134254A1 (en) * | 2018-01-02 | 2019-07-11 | 上海交通大学 | Real-time economic dispatch calculation method using distributed neural network |
| US11216759B2 (en) | 2018-01-02 | 2022-01-04 | Shanghai Jiao Tong University | Real-time economic dispatch method of power system |
| CN108805388A (en) * | 2018-04-09 | 2018-11-13 | 中国电力科学研究院有限公司 | A kind of determination method and apparatus of non-coming year Load Time Series scene |
| CN109409570A (en) * | 2018-09-20 | 2019-03-01 | 平安科技（深圳）有限公司 | Electronic device, the building day power load prediction technique and storage medium based on clustering algorithm |
| CN109118027A (en) * | 2018-09-30 | 2019-01-01 | 广东小天才科技有限公司 | Electric quantity prediction model training method, electric quantity prediction method, intelligent equipment and system |
| TWI705395B (en) * | 2019-06-28 | 2020-09-21 | 國立政治大學 | Customer value management system |
| CN112070268B (en) * | 2020-07-31 | 2024-01-12 | 浙江大有实业有限公司杭州科技发展分公司 | Power load prediction method and device based on hotel demand side response |
| CN112070268A (en) * | 2020-07-31 | 2020-12-11 | 浙江大有实业有限公司杭州科技发展分公司 | Power load prediction method and device based on hotel demand side response |
| CN111915082A (en) * | 2020-08-03 | 2020-11-10 | 深圳集智数字科技有限公司 | Prediction method, prediction device, storage medium and equipment |
| CN111915082B (en) * | 2020-08-03 | 2024-03-29 | 深圳集智数字科技有限公司 | Prediction method, prediction device, storage medium and prediction equipment |
| CN111861587A (en) * | 2020-08-04 | 2020-10-30 | 上海积成能源科技有限公司 | System and method for analyzing residential electricity consumption behavior based on hidden Markov model and forward algorithm |
| CN112766797A (en) * | 2021-01-30 | 2021-05-07 | 广东新华建工程有限公司 | Photoelectric integrated building energy supply method and system |
| CN112766797B (en) * | 2021-01-30 | 2023-12-22 | 新华建集团(广东)建设有限公司 | Photoelectric integrated building energy supply method and system |
| CN115619170A (en) * | 2022-10-28 | 2023-01-17 | 北京国电通网络技术有限公司 | Method, device, equipment, computer medium and program product for adjusting electric quantity load |
| CN116402483B (en) * | 2023-06-09 | 2023-08-18 | 国网山东省电力公司兰陵县供电公司 | Online monitoring method and system for carbon emission of park |
| CN116402483A (en) * | 2023-06-09 | 2023-07-07 | 国网山东省电力公司兰陵县供电公司 | Online monitoring method and system for carbon emission of park |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN106203683A (en) | A kind of modeling method of power customer electro-load forecast system | |
| Gao et al. | Day-ahead power forecasting in a large-scale photovoltaic plant based on weather classification using LSTM | |
| Zhao et al. | Combining forecasts of electricity consumption in China with time-varying weights updated by a high-order Markov chain model | |
| CN102629106B (en) | Water supply control method and water supply control system | |
| Glotić et al. | Short-term combined economic and emission hydrothermal optimization by surrogate differential evolution | |
| Al-Ghandoor et al. | Electricity consumption in the industrial sector of Jordan: application of multivariate linear regression and adaptive neuro-fuzzy techniques | |
| CN107622329A (en) | The Methods of electric load forecasting of Memory Neural Networks in short-term is grown based on Multiple Time Scales | |
| CN107909220A (en) | Electric heating load prediction method | |
| CN104978611A (en) | Neural network photovoltaic power generation output prediction method based on grey correlation analysis | |
| CN103793887B (en) | Short-term electric load on-line prediction method based on self-adaptive enhancement algorithm | |
| CN101888087A (en) | Method for realizing distributed super-short-term area load forecasting in distribution network terminal | |
| CN104636822A (en) | Residential load prediction method of elman-based neural network | |
| Olagoke et al. | Short term electric load forecasting using neural network and genetic algorithm | |
| CN105447509A (en) | Short-term power prediction method for photovoltaic power generation system | |
| Wang et al. | A novel decomposition-ensemble forecasting system for dynamic dispatching of smart grid with sub-model selection and intelligent optimization | |
| Petrichenko et al. | District heating demand short-term forecasting | |
| CN105243449A (en) | Method and device for correcting prediction result of electricity selling amount | |
| CN105224997A (en) | A kind of planning load forecasting method containing multiple uncertain factor impact | |
| Kofinas et al. | Daily multivariate forecasting of water demand in a touristic island with the use of artificial neural network and adaptive neuro-fuzzy inference system | |
| Wang et al. | Multi-criteria comprehensive study on predictive algorithm of heating energy consumption of district heating station based on timeseries processing | |
| CN103699947A (en) | Meta learning-based combined prediction method for time-varying nonlinear load of electrical power system | |
| CN104134103A (en) | Method for predicting energy consumption of hot oil pipeline by using modified BP neural network model | |
| CN111724064A (en) | Energy-storage-containing power distribution network planning method based on improved immune algorithm | |
| Huang et al. | Optimal Scheduling of a Residential Energy Prosumer Incorporating Renewable Energy Sources and Energy Storage Systems in a Day-ahead Energy Market | |
| Sjelvgren et al. | Large-scale non-linear programming applied to operations planning |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| C06 | Publication | ||
| PB01 | Publication | ||
| C10 | Entry into substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| CB02 | Change of applicant information | ||
| CB02 | Change of applicant information |
Address after: 710065 Shaanxi hi tech Zone, Xi'an science and technology zone, No. 68 Xi'an Software Park, Han Yun Ge C501 Applicant after: Merrill Lynch data technology Limited by Share Ltd Address before: 710065 Shaanxi hi tech Zone, Xi'an science and technology zone, No. 68 Xi'an Software Park, Han Yun Ge C501 Applicant before: XI'AN MERIT DATA TECHNOLOGY CO., LTD. |
|
| RJ01 | Rejection of invention patent application after publication | ||
| RJ01 | Rejection of invention patent application after publication |
Application publication date: 20161207 |