CN106529704A - Monthly maximum power load forecasting method and apparatus - Google Patents
Monthly maximum power load forecasting method and apparatus Download PDFInfo
- Publication number
- CN106529704A CN106529704A CN201610930965.XA CN201610930965A CN106529704A CN 106529704 A CN106529704 A CN 106529704A CN 201610930965 A CN201610930965 A CN 201610930965A CN 106529704 A CN106529704 A CN 106529704A
- Authority
- CN
- China
- Prior art keywords
- data
- maximum power
- power load
- moon
- month
- 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
- 238000013277 forecasting method Methods 0.000 title claims abstract description 26
- PEDCQBHIVMGVHV-UHFFFAOYSA-N Glycerine Chemical compound OCC(O)CO PEDCQBHIVMGVHV-UHFFFAOYSA-N 0.000 claims description 144
- 238000000034 method Methods 0.000 claims description 71
- 230000002159 abnormal effect Effects 0.000 claims description 23
- 230000005611 electricity Effects 0.000 claims description 8
- 238000013481 data capture Methods 0.000 claims description 7
- 238000004364 calculation method Methods 0.000 claims description 4
- 230000008901 benefit Effects 0.000 abstract description 4
- 230000002829 reductive effect Effects 0.000 abstract description 4
- 230000008569 process Effects 0.000 description 15
- 238000010586 diagram Methods 0.000 description 9
- 238000011160 research Methods 0.000 description 8
- 238000004458 analytical method Methods 0.000 description 7
- 238000004422 calculation algorithm Methods 0.000 description 7
- 238000004590 computer program Methods 0.000 description 7
- 230000008859 change Effects 0.000 description 6
- 238000009826 distribution Methods 0.000 description 6
- 235000006508 Nelumbo nucifera Nutrition 0.000 description 5
- 240000002853 Nelumbo nucifera Species 0.000 description 5
- 235000006510 Nelumbo pentapetala Nutrition 0.000 description 5
- 238000013459 approach Methods 0.000 description 5
- 238000005516 engineering process Methods 0.000 description 5
- 238000012545 processing Methods 0.000 description 5
- 238000012937 correction Methods 0.000 description 4
- 238000011161 development Methods 0.000 description 4
- 230000000694 effects Effects 0.000 description 4
- 230000006870 function Effects 0.000 description 4
- 241000287196 Asthenes Species 0.000 description 3
- 230000003044 adaptive effect Effects 0.000 description 3
- 238000013528 artificial neural network Methods 0.000 description 3
- 230000001419 dependent effect Effects 0.000 description 3
- 238000011156 evaluation Methods 0.000 description 3
- 238000007689 inspection Methods 0.000 description 3
- 238000004519 manufacturing process Methods 0.000 description 3
- 230000005540 biological transmission Effects 0.000 description 2
- 230000015572 biosynthetic process Effects 0.000 description 2
- 230000007935 neutral effect Effects 0.000 description 2
- 238000002360 preparation method Methods 0.000 description 2
- 238000004321 preservation Methods 0.000 description 2
- 238000007619 statistical method Methods 0.000 description 2
- 238000003860 storage Methods 0.000 description 2
- 238000003786 synthesis reaction Methods 0.000 description 2
- 238000012360 testing method Methods 0.000 description 2
- 235000008733 Citrus aurantifolia Nutrition 0.000 description 1
- 244000131316 Panax pseudoginseng Species 0.000 description 1
- 235000005035 Panax pseudoginseng ssp. pseudoginseng Nutrition 0.000 description 1
- 235000003140 Panax quinquefolius Nutrition 0.000 description 1
- 235000011941 Tilia x europaea Nutrition 0.000 description 1
- 230000001174 ascending effect Effects 0.000 description 1
- 239000012141 concentrate Substances 0.000 description 1
- 238000005520 cutting process Methods 0.000 description 1
- 238000013506 data mapping Methods 0.000 description 1
- 238000007418 data mining Methods 0.000 description 1
- 230000007812 deficiency Effects 0.000 description 1
- 230000008034 disappearance Effects 0.000 description 1
- 230000003631 expected effect Effects 0.000 description 1
- 238000009472 formulation Methods 0.000 description 1
- 239000000446 fuel Substances 0.000 description 1
- 235000008434 ginseng Nutrition 0.000 description 1
- 239000004615 ingredient Substances 0.000 description 1
- 239000004571 lime Substances 0.000 description 1
- 230000000670 limiting effect Effects 0.000 description 1
- 230000007774 longterm Effects 0.000 description 1
- 238000013507 mapping Methods 0.000 description 1
- 238000012067 mathematical method Methods 0.000 description 1
- 238000013178 mathematical model Methods 0.000 description 1
- 239000000203 mixture Substances 0.000 description 1
- 230000003472 neutralizing effect Effects 0.000 description 1
- 230000003287 optical effect Effects 0.000 description 1
- 230000036961 partial effect Effects 0.000 description 1
- 230000009467 reduction Effects 0.000 description 1
- 230000001373 regressive effect Effects 0.000 description 1
- 238000012216 screening Methods 0.000 description 1
- 241000894007 species Species 0.000 description 1
- 238000005309 stochastic process Methods 0.000 description 1
- 230000002123 temporal effect Effects 0.000 description 1
- 230000036962 time dependent Effects 0.000 description 1
- 238000012549 training Methods 0.000 description 1
- 230000009466 transformation Effects 0.000 description 1
- 230000000007 visual effect Effects 0.000 description 1
- 238000012800 visualization 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
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F16/00—Information retrieval; Database structures therefor; File system structures therefor
- G06F16/20—Information retrieval; Database structures therefor; File system structures therefor of structured data, e.g. relational data
- G06F16/24—Querying
- G06F16/245—Query processing
- G06F16/2458—Special types of queries, e.g. statistical queries, fuzzy queries or distributed queries
- G06F16/2462—Approximate or statistical queries
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06Q—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
- G06Q50/00—Information and communication technology [ICT] specially adapted for implementation of business processes of specific business sectors, e.g. utilities or tourism
- G06Q50/06—Energy or water supply
-
- Y—GENERAL TAGGING OF NEW TECHNOLOGICAL DEVELOPMENTS; GENERAL TAGGING OF CROSS-SECTIONAL TECHNOLOGIES SPANNING OVER SEVERAL SECTIONS OF THE IPC; TECHNICAL SUBJECTS COVERED BY FORMER USPC CROSS-REFERENCE ART COLLECTIONS [XRACs] AND DIGESTS
- Y02—TECHNOLOGIES OR APPLICATIONS FOR MITIGATION OR ADAPTATION AGAINST CLIMATE CHANGE
- Y02D—CLIMATE CHANGE MITIGATION TECHNOLOGIES IN INFORMATION AND COMMUNICATION TECHNOLOGIES [ICT], I.E. INFORMATION AND COMMUNICATION TECHNOLOGIES AIMING AT THE REDUCTION OF THEIR OWN ENERGY USE
- Y02D10/00—Energy efficient computing, e.g. low power processors, power management or thermal management
Landscapes
- Engineering & Computer Science (AREA)
- Business, Economics & Management (AREA)
- Physics & Mathematics (AREA)
- Theoretical Computer Science (AREA)
- Economics (AREA)
- Human Resources & Organizations (AREA)
- Strategic Management (AREA)
- General Physics & Mathematics (AREA)
- Probability & Statistics with Applications (AREA)
- Marketing (AREA)
- General Business, Economics & Management (AREA)
- Tourism & Hospitality (AREA)
- Health & Medical Sciences (AREA)
- Entrepreneurship & Innovation (AREA)
- Databases & Information Systems (AREA)
- Operations Research (AREA)
- Game Theory and Decision Science (AREA)
- Fuzzy Systems (AREA)
- Mathematical Physics (AREA)
- Software Systems (AREA)
- Computational Linguistics (AREA)
- Data Mining & Analysis (AREA)
- Quality & Reliability (AREA)
- General Engineering & Computer Science (AREA)
- Development Economics (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)
Abstract
The invention provides a monthly maximum power load forecasting method and apparatus. The monthly maximum power load forecasting method includes the steps: acquiring the real time load data, daily maximum load data, and monthly maximum power load data of each year from a historical database; and according to the monthly maximum power load data, establishing a link relative ratio increment forecasting model or a link relative ratio growth rate forecasting model, and carrying out monthly maximum power load data forecasting. By means of the monthly maximum power load forecasting method and apparatus, the forecasting accuracy can be improved, so that the operation risk for a power plant and a power grid can be reduced. The monthly maximum power load forecasting method and apparatus have the advantages of being less in data bulk, being high in operation speed and solving the problem that the sample size is limited.
Description
Technical field
The present invention relates to moon maximum power load prediction technology, especially with regard to a kind of maximum electricity of moon based on month increment
Power load Forecasting Methodology and device.
Background technology
Month maximum power load prediction is mainly used in Power System Planning and formulates generation schedule, can be also used for be
Unite abundance assessment, the formulation of generating contract, Contract generation distribution, Research on electricity price prediction, Real-Time Scheduling etc., so as to improve system operation
Economy and reliability.With the development of China's electric utility, the moon, maximum power load prediction was in Power System Planning and operation
Aspect plays more and more important effect.The load prediction of month maximum power its be substantially the prediction to electric power market demand,
Therefore, the research of moon maximum power load prediction problem, has become one of important topic of modern power systems research, closely
Nian Lai Utilities Electric Co.s while assurance load variations promptly and accurately are made every effort to, by the moon maximum power load prediction importance and compel
Cutting property brings up to unprecedented height, and also the precision of load is put forward higher requirement.
Moon maximum power load prediction is the historical values according to electric load, thus it is speculated that the numerical value in its future, with certain
Technology and model, the development trend for rationally speculating load and the situation being likely to be breached.For a long time, Chinese scholars to the moon most
The theory and method of big load forecast has done substantial amounts of research, it is proposed that various methods, at present using more
The method of short-term moon maximum power load prediction is broadly divided into two classes：Classical Forecasting Methodology and modern Forecasting Methodology.Classical prediction
Method includes：Regression analyses, time series method, gray forecast approach etc., modern Forecasting Methodology mainly include：Expert system approach,
Artificial neural network method etc..
Regression analyses on the basis of dependency relation, are set up between the independent variable and dependent variable in analysis business datum
Regression equation between variable, and using regression equation as forecast model, according to independent variable time span of forecast number change come pre-
Survey dependent variable.When the load of next month is predicted, by factor of influence value (such as PMI, temperature, gross national product, people
Mouthful etc.) and the previous data of power load carry out statistical analysiss, founding mathematical models, and derive future load amount.Its advantage
It is that speed is fast, all kinds of situations preferably can be processed.
The shortcoming of regression analyses is that data are had high demands, especially incomplete in historical data or have the situation of larger error
Under, effect is undesirable, because only that preferably can solve in the case of with linear approach neutralizing non-linear formula, but non-thread
Property model in can only add part uncertain factor (such as temperature or humidity) because uncertain factor is too many, it is impossible to all consider,
Algorithm does not have adaptivity yet, so being difficult to obtain accurate result.
Time series method is the load data according to history, finds its time dependent rule, sets up temporal model, with
The method of prediction future load numerical value, its basic assumption is：Past load variations rule can last till in the future, i.e. future was
The continuity gone.In the algorithm, dependent variable load and independent variable time are stochastic variables, by between actual load and prediction load
Difference process as smooth change procedure.In time series method, widely used model has AR (autoregression) model, MA
(dynamic average) model, ARMA (auto regressive moving average) model etc..The algorithm is relatively low to the data volume dependency in data base, mesh
Front achievement in research is ripe, is widely used in actual production.
But, time series method does not allow problem data to exist, in grid condition because needing past actual load data
Normally, when the change of the factor such as weather is little, prediction effect is good, but changes greatly or exist the situation of bad data in random factor
Under, predict the outcome not ideal.
Gray model abbreviation GM models, it is that the viewpoint and method of fuzzy control are extended in the big system of complexity, will
Automatically control and combine with the mathematical method of operational research, research has the problem of grey majorized model in being widely present in objective world.Ash
The modeling process of color model sets up Differential Equation Model after usually being generated with historical data row, and it is built upon following four
On the basis of：One be the scope of model and time zone be it is given, but the random partial in model, stochastic process are changes；
Two is that the regular ascending sequence with exponential increase is generated after data sequence random superposition；Three can be according to model
The generating mode of mellow lime number, the screening of data, respectively with the amendment of residual error (GM) model adjusting and improve the precision of the model；
Four be the GM Model Groups of system modelling under high-order be first order differential equation system into the gray model with exponential increase, generally made
Grey forecasting model be GM (1,1) model and GM (1, n) model.The advantage of gray model is that the data of requirement are few, need not
There is excessive dependence to data, do not consider the regularity of distribution, do not consider variation tendency, simple operation, be easily verified.
But, gray model is few due to the data for requiring, as data discrete degree increases, precision of prediction can also be deteriorated,
Therefore be not suitable for long-term forecast, and the higher predictive value of significant precision is nearest several data, so while data
The error brought less is larger, is rarely employed in actual production process.
Expert system approach is according to being predicted with the knowledge of expert, method, empirical log.The algorithm is equivalent to by the mankind's
It is that formula or model are applied in actual prediction that experience is abstract, is a kind of unique algorithm.The advantage of the algorithm concentrates multidigit special
The past prediction data of family, by the rule for finding scholarly forecast, is translated into a moon ability for maximum power load prediction, and
Draw, the concordance that can be kept and predict the outcome in the past.
But, expert system approach does not have adaptive ability, to the event bad adaptability being continually changing, therefore is easily known
Know the restriction of species and total amount.
Neural network theory is the learning capacity using neutral net, allows computer learning to be included in historical load data
Mapping relations, recycle this mapping relations to predict future load.Neutral net have stronger adaptive learning ability and
Nonlinear Processing ability, is widely used in moon maximum power load prediction, by adaptive learning training come
Process a large amount of non-linear ingredient and inaccurate rule that nature is present.
Artificial neural network method algorithm has very big application market, but there is also many deficiencies：Such as restrain slowly, be easily absorbed in office
The minimum state in portion；Network structure determines that shortage is effectively instructed, and subjective dependency is strong, excessively complicated, operation time length etc..
The content of the invention
The embodiment of the present invention provides a kind of moon maximum power load forecasting method and device, to improve precision of prediction, reduces
The operations risks in power plant and electrical network, reduction data volume are few, improve arithmetic speed.
To achieve these goals, a kind of moon maximum power load forecasting method is embodiments provided, the moon is most
Big Methods of electric load forecasting includes：
Each annual real-time load data, Daily treatment cost data and moon maximum power load number are obtained from historical data base
According to；
Chain rate incremental forecasting model or sequential growth rate forecast model are set up according to the moon maximum power load data, is entered
The maximum power load data prediction of the row moon.
In one embodiment, each annual real-time load data, Daily treatment cost data and the moon are obtained from historical data base most
Big Power system load data, including：
In obtaining daily from the historical data base, the real-time load data of multiple stoichiometric points, obtains setting annual scope
Interior daily stoichiometric point real-time load data；
Obtain setting the daily Daily treatment cost data in annual scope according to the stoichiometric point real-time load data；
The moon maximum power load in the annual scope of setting monthly is calculated according to the Daily treatment cost data
Data.
In one embodiment, before the moon maximum power load data is obtained, this month maximum power load forecasting method
Also include：
Whether there is missing values in searching the Daily treatment cost data；
The missing values are modified using average enthesis.
In one embodiment, before the moon maximum power load data is obtained, this month maximum power load forecasting method
Also include：
Whether there is exceptional value in searching the Daily treatment cost data, the exceptional value is modified.
In one embodiment, whether there is exceptional value in searching the Daily treatment cost data, the exceptional value is repaiied
Just, including：
Exceptional value in the Daily treatment cost data is searched using lateral likeness method；
The exceptional value is modified using average enthesis.
In one embodiment, whether there is exceptional value in searching the Daily treatment cost data, the exceptional value is repaiied
Just, including：
From deleting unrelidble data Xi in the n Daily treatment cost data, the flat of (n-1) individual Daily treatment cost data is calculated
AverageWith standard deviation δ, ifThen think that Xi is exceptional value；
Use the meansigma methodssReplace the suspicious data Xi；
Wherein, K is the difference of actual value and standard value.
In one embodiment, chain rate incremental forecasting model is set up according to the moon maximum power load data, including：
Load maximum increment of each moon relative to last month is calculated according to the moon maximum power load data；
Chain rate incremental forecasting model is set up according to the load maximum increment.
In one embodiment, sequential growth rate forecast model is set up according to the moon maximum power load data, including：
Sequential growth rate of each moon relative to last month is calculated according to the moon maximum power load data；
Sequential growth rate forecast model is set up according to the sequential growth rate.
In one embodiment, maximum power load forecasting method also includes within this month：
Calculate the actual natural law being separated by with last month each moon；
The chain rate incremental forecasting model is modified according to the actual natural law.
In one embodiment, maximum power load forecasting method also includes within this month：
Set the range of predicted value of every month；
The predictive value that moon maximum power load data prediction is obtained is compared with the range of predicted value；
If the predictive value is replaced with the maximum or minima of the range of predicted value not in the range of predicted value
Change the predictive value.
In one embodiment, maximum power load forecasting method also includes within this month：
Absolute error or relative error are calculated according to moon maximum power load actual value and moon maximum power predicted load；
The absolute error or relative error are fed back to into chain rate incremental forecasting model or sequential growth rate forecast model enters
Row Modifying model.
To achieve these goals, the embodiment of the present invention additionally provides a kind of moon maximum power load prediction device, the moon
Maximum power load prediction device includes：
Data capture unit, for each annual real-time load data, Daily treatment cost data are obtained from historical data base
And moon maximum power load data；
Predicting unit, for setting up chain rate incremental forecasting model or sequential growth rate according to the moon maximum power load data
Rate forecast model, carries out a moon maximum power load data prediction.
In one embodiment, data capture unit described in this month maximum power load prediction device includes：
Real time data acquisition module, for the Real-time Load of multiple stoichiometric points in obtaining daily from the historical data base
Data, obtain setting stoichiometric point real-time load data daily in annual scope；
Daily load data generation module, for obtaining setting every in annual scope according to the stoichiometric point real-time load data
It the Daily treatment cost data；
Month load data generation module, it is every in the annual scope of setting for being calculated according to the Daily treatment cost data
The moon maximum power load data of the moon.
In one embodiment, moon maximum power load prediction device also includes：
Missing values searching unit, whether there is missing values for searching in the Daily treatment cost data；
Missing values amending unit, for being modified to the missing values using average enthesis.
In one embodiment, moon maximum power load prediction device also includes：
Abnormal amending unit, for searching with the presence or absence of exceptional value in the Daily treatment cost data, to the exceptional value
It is modified.
In one embodiment, the abnormal amending unit includes：
Abnormal searching modul, for the exceptional value in the Daily treatment cost data is searched using lateral likeness method；
Abnormal correcting module, for being modified to the exceptional value using average enthesis.
In one embodiment, the abnormal amending unit includes：
Abnormal determining module, for from deleting unrelidble data Xi in the n Daily treatment cost data, calculating (n-1) individual
The meansigma methodss of Daily treatment cost dataWith standard deviation δ, ifThen think that Xi is exceptional value；
Replace module, for using the meansigma methodssReplace the suspicious data Xi；
Wherein, K is the difference of actual value and standard value.
In one embodiment, the predicting unit includes：
Incremental computations module, for according to the moon maximum power load data calculate each moon relative to last month load most
Big value increment；
Incremental forecasting model building module, for setting up chain rate incremental forecasting model according to the load maximum increment.
In one embodiment, sequential growth rate forecast model is set up according to the moon maximum power load data, including：
Sequential growth rate computing module, for calculating each moon relative to last month according to the moon maximum power load data
Sequential growth rate；
Rate of increase forecast model sets up module, for setting up sequential growth rate forecast model according to the sequential growth rate.
In one embodiment, moon maximum power load prediction device also includes：
Natural law computing unit, for calculating the actual natural law that each moon and last month be separated by；
Amending unit, for being modified to the chain rate incremental forecasting model according to the actual natural law.
In one embodiment, moon maximum power load prediction device also includes：
Range setting module, for setting the range of predicted value of every month；
Comparing unit, for the predictive value that moon maximum power load data prediction is obtained is carried out with the range of predicted value
Relatively；
Predictive value replacement unit, replaces the predictive value for the maximum with the range of predicted value or minima.
In one embodiment, moon maximum power load prediction device also includes：
Error calculation unit, for calculating exhausted according to moon maximum power load actual value and moon maximum power predicted load
To error or relative error；
Modifying model unit, for the absolute error or relative error are fed back to chain rate incremental forecasting model or chain rate
Rate of increase forecast model carries out Modifying model.
Using the present invention, precision of prediction can be improved, so as to reduce the operations risks in power plant and electrical network；Number of the present invention
Few according to amount, fast operation solves the problems, such as that sample size is limited.
Description of the drawings
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
Accompanying drawing to be used needed for having technology description is briefly described, it should be apparent that, drawings in the following description are only this
Some embodiments of invention, for those of ordinary skill in the art, on the premise of not paying creative work, can be with
Other accompanying drawings are obtained according to these accompanying drawings.
Moon maximum power load forecasting method flow charts of the Fig. 1 for the embodiment of the present invention；
Moon maximum power load data generation method flow charts of the Fig. 2 for the embodiment of the present invention；
Missing values querying method flow charts of the Fig. 3 for the embodiment of the present invention；
Abnormal value correction method flow charts of the Fig. 4 for one embodiment of the invention；
Abnormal value correction method flow charts of the Fig. 5 for another embodiment of the present invention；
Each time moon load maximum broken line graphs of Fig. 6 A to Fig. 6 L for the embodiment of the present invention；
Structured flowcharts of the Fig. 7 for the moon maximum power load prediction device of the embodiment of the present invention；
Structured flowcharts of the Fig. 8 for the data capture unit of the embodiment of the present invention；
Structured flowcharts of the Fig. 9 for the moon maximum power load prediction device of one embodiment of the invention；
Structured flowcharts of the Figure 10 for the moon maximum power load prediction device of one embodiment of the invention；
Structured flowcharts of the Figure 11 for the abnormal amending unit of the embodiment of the present invention；
Structured flowcharts of the Figure 12 for the abnormal amending unit of the embodiment of the present invention；
Structured flowcharts of the Figure 13 for the predicting unit of the embodiment of the present invention；
Structured flowcharts of the Figure 14 for the predicting unit of the embodiment of the present invention；
Structured flowcharts of the Figure 15 for the moon maximum power load prediction device of one embodiment of the invention；
Structured flowcharts of the Figure 16 for the moon maximum power load prediction device of one embodiment of the invention；
Structured flowcharts of the Figure 17 for the moon maximum power load prediction device of one embodiment of the invention.
Specific embodiment
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
Site preparation is described, it is clear that described embodiment is only a part of embodiment of the invention, rather than the embodiment of whole.It is based on
Embodiment in the present invention, it is every other that those of ordinary skill in the art are obtained under the premise of creative work is not made
Embodiment, belongs to the scope of protection of the invention.
Main terms according to the present invention are as follows：
Electric load：In power system, the electrical power needed for electrical equipment is referred to as electric load or electric power.
Load forecast：Load forecast is according to the operation characteristic of system, increase-volume decision-making, natural conditions and society
The factors such as impact, under conditions of certain required precision is met, it is determined that the load data of following certain particular moment.Power load
Lotus prediction is one of groundworks of administration section such as power system electricity consumption, plan, scheduling.
Moon maximum power load prediction：Electric load is divided by size, can be divided into peak load, average load and minimum
Load.Peak load is also known as maximum load or peakload, and it was associated with certain statistic record time, has day maximum negative
Lotus, moon maximum power load and annual peak load point.Month maximum power load refers to the maximum day power load of certain month year Nei.
From planning and the angle predicted, the moon maximum power load be an extremely important parameter, it is establishment power system day operation side
The important evidence of formula and yearly operation mode.Moon maximum power load prediction is set for determination power system generating equipment and power transmission and transformation
Standby capacity is very important, is conducive to selecting appropriate machine set type and rational power supply architecture, and determines fuel gage
Draw etc., can also be that the normal operation of the peak value, the capacity of hydroenergy storage station and a transmission facility of research power system is carried
Support for data.
Moon maximum power load prediction is the historical values according to moon maximum power load, thus it is speculated that the numerical value in its future, fortune
With certain technology and model, the development trend for rationally speculating load and the situation being likely to be breached.By to the moon over the years maximum electricity
The visual presentation of power load value, due to the moon load increment diversity of the annual same period it is little, that is to say, that with certain phase
Like property, based on this rule can by the increment of the same period of previous year as this year same period the foundation that changes of increment.Month increment
The basic thought of peak load prediction is exactly to be predicted according to the data of the previous year same period and former phases, and by maximum negative
Lotus difference occurs day and history maximum and history minima carries out Modifying model respectively.The present invention only considers close month load
Measuring and trend for change, has evaded impact of various uncertain factors to load variations to a great extent.
Moon maximum power load forecasting method flow charts of the Fig. 1 for the embodiment of the present invention, as shown in figure 1, this month maximum electricity
Power load Forecasting Methodology includes：
S101：Each annual real-time load data, Daily treatment cost data and moon maximum power are obtained from historical data base
Load data；
S102：Chain rate incremental forecasting model or sequential growth rate prediction mould are set up according to the moon maximum power load data
Type, carries out a moon maximum power load data prediction.
Flow diagram as shown in Figure 1, the application set up forecast model by the moon maximum power load data for obtaining,
A moon maximum power load data prediction is carried out, precision of prediction can be improved, the operations risks of power plant and electrical network are reduced.
The purpose of S101 is to obtain a moon maximum power load data, in an embodiment, as shown in Fig. 2 can pass through as follows
Step is realized：
S201：The real-time load data of multiple stoichiometric points in obtaining daily from the historical data base, obtains setting year
Stoichiometric point real-time load data daily in the range of degree.Multiple stoichiometric points can be arranged according to time interval within one day, such as per 20
Minute a stoichiometric point is set, then daily in have 96 stoichiometric points.
S202：Obtain setting the daily Daily treatment cost in annual scope according to the stoichiometric point real-time load data
Data.
S203：The moon maximum power in the annual scope of setting monthly is calculated according to the Daily treatment cost data
Load data.
The daily stoichiometric point real-time load data obtained in S201 may be because of the shortage of data of some stoichiometric points
Cause data imperfect, will so cause moon maximum power load data inaccurate, will finally affect moon maximum power load number
According to prediction.In order to obtain the predictive value of accurate moon maximum power load data, needing to inquire about in Daily treatment cost data is
It is no to there are missing values, in an embodiment, as shown in figure 3, missing values inquiry can be carried out by the following method：
S301：Whether there is missing values in searching the Daily treatment cost data.
S302：The missing values are modified using average enthesis.Due to the load of interior adjacent stoichiometric point on the same day
Essentially identical, in adjacent several days, type identical daily load is essentially identical.If there is missing values, can be filled up by average
Method is modified to missing values, for example the shortage of data of the 3rd stoichiometric point, can calculate the 2nd stoichiometric point and the 4th stoichiometric point
Meansigma methodss, as the data of the 3rd stoichiometric point.
In addition, the daily stoichiometric point real-time load data obtained in S201 is it is possible that certain stoichiometric point or some meters
The data appearance for measuring point is abnormal, or the daily load data of one day exception occur and so will cause moon maximum power load data not
Accurately, by the final prediction for affecting moon maximum power load data.So the present invention obtain moon maximum power load data it
Before, whether there is exceptional value during the Daily treatment cost data can also be searched, exceptional value is modified.
In one embodiment, as shown in figure 4, whether there is exceptional value in searching the Daily treatment cost data, to described different
Constant value is modified, and can realize by the following method：
S401：Exceptional value in the Daily treatment cost data is searched using lateral likeness method.
S402：The exceptional value is modified using average enthesis.Due to the load of interior adjacent stoichiometric point on the same day
Essentially identical, in adjacent several days, type identical daily load is essentially identical.If there is exceptional value, can be filled up by average
Method is modified to exceptional value, and for example the data of the 3rd stoichiometric point are exceptional value, can calculate the 2nd stoichiometric point and the 4th meter
The meansigma methodss of amount point, as the data of the 3rd stoichiometric point.
In one embodiment, as shown in figure 5, whether there is exceptional value in searching the Daily treatment cost data, to described different
Constant value is modified, and can realize by the following method：
S501：Using T test criterions, from deleting unrelidble data Xi in the n Daily treatment cost data, calculate (n-1)
The meansigma methodss of individual Daily treatment cost dataWith standard deviation δ, ifThen think that Xi is exceptional value.Wherein, K is actual value
With the difference of standard value, can be checked in from α tables according to the significance selected by sample size n, be entered using Hash table in system
The preservation of the capable value.
S502：If Xi is exceptional value, meansigma methodss are usedReplace the suspicious data Xi.
The present invention can carry out Data Mining before modeling.Specifically, preanalysis are carried out to load data, according to day,
The moon, season, year at times, region-by-region carry out statistical computation, analysis can be represented with forms such as pie chart, broken line graph, block diagram, forms
As a result.By statistical analysiss, load curve trend and Changing Pattern can be apparent from, be that model selects to establish statistics base
Plinth.
When the present invention carries out moon maximum power load data and predicts, chain rate incremental forecasting model and/or chain rate can be set up
Rate of increase forecast model, illustrates that chain rate incremental forecasting model and sequential growth rate forecast model set up process separately below.
1st, set up chain rate incremental forecasting model
Chain rate is a statistics term, is that current period statistical data is compared with last issue, such as in July, 2016 and 2016 6
Month compares, and is chain rate.Chain rate increment refers to that this month statistical data compares statistical data increment last month.
Load maximum increment of each moon relative to last month is calculated according to the moon maximum power load data：Note a years 1
The load maximum in the moon-December isEach moon the load maximum increment (chain rate increment) with respect to last month
ForThen have：
Chain rate incremental forecasting model is set up according to the load maximum increment：If need prediction is a+1 years i-th
The load maximum of individual month, is designated asThen its predictive value：
2nd, set up sequential growth rate forecast model
Sequential growth rate of each moon relative to last month is calculated according to the moon maximum power load data：Sequential growth rate=
(this month maximum power load value-last month maximum power load value)/last month maximum power load value × 100%.
Sequential growth rate forecast model is set up according to the sequential growth rate：If need prediction is i-th of a+1 years
The load maximum of the moon, is designated asThen its predictive value：
The load maximum in 1st month years of a+1：
Above-mentioned chain rate incremental forecasting model and sequential growth rate forecast model are at intervals of one based on this month with last month
What calendar month was obtained, as the interval time that the load maximum of former and later two months is produced differs and is set to one month, so can be
There is day number of days different based on peak load on the basis of chain rate incremental forecasting model to be modified, amendment principle is by upper one year
The value of same period increment be multiplied by the ratio of a number of days different.Illustrate natural law makeover process：
Table 1 represents 2015,2,016 two years 3, the exact date at April load maximum place.
There is day in 1 load maximum of table
So 2015 3, April load maximum date of occurrence at intervals of 2015/4/12-2015/3/14=29 days, relatively
In 31 days of in March, 2015, this time interval was 29/31=0.9355 month；And 2016 3, April load maximum
It is divided into during birthday 2016/4/19-2016/3/12=38 days, relative to 31 days of in March, 2016, this time interval was
38/31=1.2258 individual month.
And in sequential growth rate forecast model, the Forecasting Methodology of employing is defaulted as load maximum between two months and day occurs
Interval be always 1 month, it is possible to consideration this factor is added in Forecasting Methodology.
Firstly the need of the actual natural law that calculating each moon was separated by with last month, then according to the actual natural law to the ring
It is modified than incremental forecasting model, specifically：Remember that the annual i-th month natural law of a isThis month load is most
Big value date of occurrence with the natural law that last month, load maximum date of occurrence was separated by isSo this time
At intervals of(note, the value in January within individual monthIndividual month).
If need prediction is the load maximum in i-th month years of a+1, it is designated asThen its predictive value
Month maximum power predicted load sometimes can be substantially too big or substantially too little, this extreme in order to prevent
Situation, can be realized by moon load maximum experience storehouse, specifically, can set the range of predicted value of every month, by the moon most
The predictive value that big Power system load data prediction is obtained is compared with the range of predicted value, if the predictive value is not described
In range of predicted value, the predictive value is replaced with the maximum or minima of the range of predicted value.
This range of predicted value can be determined with the history maximum of month load maximum, minima, such
Scope development over time can produce some changes, but may eventually form an almost unchanged experience storehouse, the experience storehouse pair
The load maximum range of every month is made that restriction, predictive value can be caused to be unlikely to deviation too big.
Remember that the annual i-th month load maximum predictive values of a+1 areThe range of predicted value of this month isThen revised predictive value
In order to check the quality for predicting the outcome, it is necessary to work out a set of evaluation index and forecast model is tested and is evaluated
(model evaluation), i.e., contrasted predictive value with actual value after having had actual value, checks the precision of prediction.According to prediction
Effect assessment principle and convention, present invention employs the standard that relative error and mean absolute error predict the outcome as inspection：
Absolute error and relative error：
If Y represents actual value,Predictive value is represented, is then claimedFor absolute error, claimFor relative error, sometimes
Relative error also uses percentTo represent.
Mean absolute error MAE：
Just have negative as forecast error has, offset in order to avoid positive and negative, therefore the absolute value for taking error carries out synthesis and counts
Its average is calculated, this is one of aggregative indicator of error analyses.
After being verified, absolute error or relative error are fed back to into forecast model, by adjusting modeling parameters, improve pre-
Survey model accuracy.
It is last in the present invention, can preserve and predict the outcome：When load prediction is carried out, it then follows " first segment, predict again "
Principle, i.e., the data of each dimension of each year are first segmented out from mass data, then is respectively put into model carries out load prediction, finally will
Predict the outcome and be compared with real data, evaluate to predicting the outcome, and feed back to forecast model, by adjusting modeling ginseng
Number, improves precision of forecasting model, finally draws satisfactory precision, by this preservation that predicts the outcome, and be applied to practical business
In, prediction work terminates.
The moon maximum power load method of the present invention is illustrated with reference to specific example.
This section by somewhere 2012 to the maximum power load prediction of 2016 days, by the preparation of historical data,
The structure of model, prediction and outcome evaluation tell about the implementation process of Forecasting Methodology.
First, history data store is created, builds data platform, 2012 to 2016 all power consumptions and electricity consumption are born
Lotus data are stored in data warehouse, and data field sample table is as shown in table 2：
The power load stored in 2 data warehouse of table and power consumption sample table
Modeling desired data is extracted from data warehouse：As 1 day to 2016 January in 2012 is only stored in data base
The data of July 31, data volume are little, and data are exported to CSV forms, have an X-rayed table function with Excel, extract in January, 2012 extremely
The moon maximum power load and date that every month, maximum power load occurred in 55 month in July, 2016, such as table 3 and table
Shown in 4：
Table in January, 3 2012 maximum power load month in and month out to 2016 7
The date that table in January, 4 2012, maximum power load occurred month in and month out to 2016 7
This month maximum power load is calculated based on table 3 and 4 data of table and has occurred with maximum power load generation last month day
The natural law that day differs, is designated as Δi, it is assumed that the date that current period maximum occurs is xi, the date that last maximum occurs is xi-1, root
According to formula：
According to above-mentioned formula, a moon maximum power load number of days different such as table 5 can be drawn：
Table May maximum power load number of days different
(history maximum is called respectively minimum with history with reference to the maximum and minima in this time period monthly
Value), obtain per the maximum of maximum power load month in and month out and minima is as shown in table 6 respectively：
Table in January, 6 2012 maximum power demand history maxima and minima month in and month out to 2016 7
For data obtained above, need to calculate missing values and exceptional value：
Missing values, are 5.34% by statistical data miss rate, and missing data is inquired about and supplemented from scheduling daily paper Intranet, is mended
Only a few data disappearances after finishing are charged, according to the rule of data distribution, is filled up with the meansigma methodss in adjacent two days.
Exceptional value, by T test rules, is found out the exceptional value in data, and is modified with the average in the adjacent date.
In order to it is directly perceived, be well understood to data distribution situation, all being described property of data can be counted, check data
The regularity of distribution, by each moon in time maximum broken line graph, such as Fig. 6 A to Fig. 6 L.As can be seen that the maximum in corresponding month in each year
Distribution trend reaches unanimity, and can be predicted using incremental model.
When setting up chain rate incremental forecasting model, using the changing value of the same period of upper one year and early stage as this year current period and last issue
Difference Δ i, i.e., by the same period upper one year the moon maximum power load, the early stage of upper one year moon maximum power load and this year
Early stage peak load predicting the moon maximum power load value of this year current period.
The computing formula of chain rate incremental forecasting model is：
The predictive value in the April, 2014 obtained based on this model is as shown in table 7 below：
Table 7
Natural law amendment can be occurred the rate of increase correction model of day by maximum, and computing formula is as follows：
The predictive value for obtaining is as shown in table 8：
Table 8
Rule of thumb predictive value does not create the highest again or new low point, so on the basis of it there is day amendment according to the history same period most
Big value and minimaModifying model is carried out again, and amendment principle is：If according to the revised prediction of generation day
Value is more than same period history maximum, then the moon value of maximum power load be changed into the history maximum of the same period, if predictive value is little
In same period history minima, then history minima of the value of the peak load of this month for the same period, otherwise predictive value are constant, finally
Revised model is optimum model.I.e.：
It is as shown in table 9 according to the predictive value obtained after extreme value amendment：
Table 9
By that analogy, the predictive value in the whole year in 2014 can be obtained, it is as shown in table 10 below：
Table 10
Present invention employs the standard that relative error and mean absolute error predict the outcome as inspection：
1st, absolute error and relative error
If Y represents actual value,Predictive value is represented, is then claimedFor absolute error, claimFor relative error.Have
When relative error also use percentTo represent.This is a kind of intuitively expression of error, in power system
It is commonly used as a kind of performance assessment criteria.
2nd, mean absolute error MAE
Just have negative as forecast error has, offset in order to avoid positive and negative, therefore the absolute value for taking error carries out synthesis and counts
Its average is calculated, this is one of aggregative indicator of error analyses.Such as table 11, data instance of the present invention to predict 2014, lead to
Crossing the error rate that predicts the outcome that obtains of inspection is：
Table 11
By predicting the outcome for table 11, it can be seen that by the result of basic forecast model prediction, global error rate is
2.79%, after once correcting, error rate is reduced to 2.34%, and after second-order correction, error rate is reduced to 1.87%, integrally reduces
0.92 percentage point, forecast error rate is maintained at 2 percentage points or so, drastically increases the precision of load prediction, reaches
Expected effect.
Finally, corresponding visualization tool is derived and be saved into revised predicting the outcome, and write report, achieve.
Moon maximum power load prediction apparatus structure block diagrams of the Fig. 7 for the embodiment of the present invention, as shown in fig. 7, this month maximum
Load forecast device includes：
Data capture unit 701, for each annual real-time load data, Daily treatment cost number are obtained from historical data base
According to and moon maximum power load data；
Predicting unit 702, for setting up chain rate incremental forecasting model or chain rate according to the moon maximum power load data
Rate of increase forecast model, carries out a moon maximum power load data prediction.
In one embodiment, as shown in figure 8, the data capture unit 701 includes：
Real time data acquisition module 801, in obtaining daily from the historical data base multiple stoichiometric points it is real-time
Load data, obtains setting stoichiometric point real-time load data daily in annual scope；
Daily load data generation module 802, for obtaining setting annual scope according to the stoichiometric point real-time load data
The interior daily Daily treatment cost data；
Month load data generation module 803, for being calculated the annual scope of setting according to the Daily treatment cost data
The interior moon maximum power load data monthly.
In one embodiment, as shown in figure 9, this month maximum power load prediction device also includes：
Missing values searching unit 901, whether there is missing values for searching in the Daily treatment cost data；
Missing values amending unit 902, for being modified to the missing values using average enthesis.
In one embodiment, as shown in Figure 10, this month maximum power load prediction device also includes：
Abnormal amending unit 1001, for searching with the presence or absence of exceptional value in the Daily treatment cost data, to described different
Constant value is modified.
In one embodiment, as shown in figure 11, abnormal amending unit 1001 includes：
Abnormal searching modul 1101, for the exception in the Daily treatment cost data is searched using lateral likeness method
Value；
Abnormal correcting module 1102, for being modified to the exceptional value using average enthesis.
In one embodiment, as shown in figure 12, abnormal amending unit 1001 includes：
Abnormal determining module 1201, for from deleting unrelidble data Xi in the n Daily treatment cost data, calculating (n-
1) meansigma methodss of individual Daily treatment cost dataWith standard deviation δ, ifThen think that Xi is exceptional value；
Replace module 1202, for using the meansigma methodssReplace the suspicious data Xi；
Wherein, K is the difference of actual value and standard value.
In one embodiment, as shown in figure 13, the predicting unit 702 includes：
Incremental computations module 1301, for calculating bearing relative to last month of each moon according to the moon maximum power load data
Lotus maximum increment；
Incremental forecasting model building module 1302, for setting up chain rate incremental forecasting mould according to the load maximum increment
Type.
In one embodiment, as shown in figure 14, the predicting unit 702 includes：
Sequential growth rate computing module 1401, for according to the moon maximum power load data calculate each moon relative to
The sequential growth rate of the moon；
Rate of increase forecast model sets up module 1402, for setting up sequential growth rate prediction mould according to the sequential growth rate
Type.
In one embodiment, as shown in figure 15, this month maximum power load prediction device also includes：
Natural law computing unit 1501, for calculating the actual natural law that each moon and last month be separated by；
Amending unit 1502, for being modified to the chain rate incremental forecasting model according to the actual natural law.
In one embodiment, as shown in figure 16, this month maximum power load prediction device also includes：
Range setting module 1601, for setting the range of predicted value of every month；
Comparing unit 1602, for moon maximum power load data is predicted the predictive value and the range of predicted value that obtain
It is compared；
Predictive value replacement unit 1603, replaces the prediction for the maximum with the range of predicted value or minima
Value.
In one embodiment, as shown in figure 17, this month maximum power load prediction device also includes：
Error calculation unit 1701, based on according to moon maximum power load actual value and moon maximum power predicted load
Calculate absolute error or relative error；
Modifying model unit 1702, for by the absolute error or relative error feed back to chain rate incremental forecasting model or
Sequential growth rate forecast model carries out Modifying model.
Those skilled in the art are it should be appreciated that embodiments of the invention can be provided as method, system or computer program
Product.Therefore, the present invention can adopt complete hardware embodiment, complete software embodiment or with reference to the reality in terms of software and hardware
Apply the form of example.And, the present invention can be using the computer for wherein including computer usable program code at one or more
The computer program implemented in usable storage medium (including but not limited to disk memory, CD-ROM, optical memory etc.) is produced
The form of product.
The present invention be with reference to method according to embodiments of the present invention, equipment (system), and computer program flow process
Figure and/or block diagram are describing.It should be understood that can be by computer program instructions flowchart and/or each stream in block diagram
The combination of journey and/or square frame and flow chart and/or flow process and/or square frame in block diagram.These computer programs can be provided
The processor of general purpose computer, special-purpose computer, Embedded Processor or other programmable data processing devices is instructed to produce
A raw machine so that produced for reality by the instruction of computer or the computing device of other programmable data processing devices
The device of the function of specifying in present one flow process of flow chart or one square frame of multiple flow processs and/or block diagram or multiple square frames.
These computer program instructions may be alternatively stored in and can guide computer or other programmable data processing devices with spy
Determine in the computer-readable memory that mode works so that the instruction being stored in the computer-readable memory is produced to be included referring to
Make the manufacture of device, the command device realize in one flow process of flow chart or one square frame of multiple flow processs and/or block diagram or
The function of specifying in multiple square frames.
These computer program instructions can be also loaded in computer or other programmable data processing devices so that in meter
Series of operation steps is performed on calculation machine or other programmable devices to produce computer implemented process, so as in computer or
The instruction performed on other programmable devices is provided for realizing in one flow process of flow chart or multiple flow processs and/or block diagram one
The step of function of specifying in individual square frame or multiple square frames.
Apply specific embodiment to be set forth the principle and embodiment of the present invention in the present invention, above example
Explanation be only intended to help and understand the method for the present invention and its core concept；Simultaneously for one of ordinary skill in the art,
According to the thought of the present invention, will change in specific embodiments and applications, in sum, in this specification
Appearance should not be construed as limiting the invention.
Claims (22)
1. a kind of month maximum power load forecasting method, it is characterised in that include：
Each annual real-time load data, Daily treatment cost data and moon maximum power load data are obtained from historical data base；
Chain rate incremental forecasting model or sequential growth rate forecast model are set up according to the moon maximum power load data, the moon is carried out
Maximum power load data is predicted.
2. according to claim 1 month maximum power load forecasting method, it is characterised in that obtain from historical data base
Each annual real-time load data, Daily treatment cost data and moon maximum power load data, including：
The real-time load data of multiple stoichiometric points in obtaining daily from the historical data base, obtains setting every in annual scope
It stoichiometric point real-time load data；
Obtain setting the daily Daily treatment cost data in annual scope according to the stoichiometric point real-time load data；
The moon maximum power load data in the annual scope of setting monthly is calculated according to the Daily treatment cost data.
3. according to claim 1 month maximum power load forecasting method, it is characterised in that obtaining the moon maximum electricity
Before power load data, also include：
Whether there is missing values in searching the Daily treatment cost data；
The missing values are modified using average enthesis.
4. according to claim 1 month maximum power load forecasting method, it is characterised in that obtaining the moon maximum electricity
Before power load data, also include：
Whether there is exceptional value in searching the Daily treatment cost data, the exceptional value is modified.
5. according to claim 4 month maximum power load forecasting method, it is characterised in that search the Daily treatment cost
Whether there is exceptional value in data, the exceptional value is modified, including：
Exceptional value in the Daily treatment cost data is searched using lateral likeness method；
The exceptional value is modified using average enthesis.
6. according to claim 4 month maximum power load forecasting method, it is characterised in that search the Daily treatment cost
Whether there is exceptional value in data, the exceptional value is modified, including：
From deleting unrelidble data Xi in the n Daily treatment cost data, the meansigma methodss of (n-1) individual Daily treatment cost data are calculatedWith standard deviation δ, ifThen think that Xi is exceptional value；
Use the meansigma methodssReplace the suspicious data Xi；
Wherein, K is the difference of actual value and standard value.
7. according to claim 1 month maximum power load forecasting method, it is characterised in that according to the moon maximum power
Load data sets up chain rate incremental forecasting model, including：
Load maximum increment of each moon relative to last month is calculated according to the moon maximum power load data；
Chain rate incremental forecasting model is set up according to the load maximum increment.
8. according to claim 1 month maximum power load forecasting method, it is characterised in that according to the moon maximum power
Load data sets up sequential growth rate forecast model, including：
Sequential growth rate of each moon relative to last month is calculated according to the moon maximum power load data；
Sequential growth rate forecast model is set up according to the sequential growth rate.
9. according to claim 8 month maximum power load forecasting method, it is characterised in that also include：
Calculate the actual natural law being separated by with last month each moon；
The chain rate incremental forecasting model is modified according to the actual natural law.
10. according to claim 1 month maximum power load forecasting method, it is characterised in that also include：
Set the range of predicted value of every month；
The predictive value that moon maximum power load data prediction is obtained is compared with the range of predicted value；
If the predictive value is not in the range of predicted value, institute is replaced with the maximum or minima of the range of predicted value
State predictive value.
11. according to claim 1 months maximum power load forecasting methods, it is characterised in that also include：
Absolute error or relative error are calculated according to moon maximum power load actual value and moon maximum power predicted load；
The absolute error or relative error are fed back to chain rate incremental forecasting model or sequential growth rate forecast model carries out mould
Type amendment.
12. a kind of month maximum power load prediction devices, it is characterised in that include：
Data capture unit, for each annual real-time load data, Daily treatment cost data and the moon are obtained from historical data base
Maximum power load data；
Predicting unit, for setting up chain rate incremental forecasting model according to the moon maximum power load data or sequential growth rate is pre-
Model is surveyed, a moon maximum power load data prediction is carried out.
13. according to claim 12 months maximum power load prediction devices, it is characterised in that the data capture unit
Including：
Real time data acquisition module, for the Real-time Load number of multiple stoichiometric points in obtaining daily from the historical data base
According to obtaining setting stoichiometric point real-time load data daily in annual scope；
Daily load data generation module, for obtaining setting daily in annual scope according to the stoichiometric point real-time load data
The Daily treatment cost data；
Month load data generation module, for being calculated in the annual scope of setting monthly according to the Daily treatment cost data
The moon maximum power load data.
14. according to claim 12 months maximum power load prediction devices, it is characterised in that also include：
Missing values searching unit, whether there is missing values for searching in the Daily treatment cost data；
Missing values amending unit, for being modified to the missing values using average enthesis.
15. according to claim 12 months maximum power load prediction devices, it is characterised in that also include：
Abnormal amending unit, for searching in the Daily treatment cost data with the presence or absence of exceptional value, is carried out to the exceptional value
Amendment.
16. according to claim 15 months maximum power load prediction devices, it is characterised in that the abnormal amending unit
Including：
Abnormal searching modul, for the exceptional value in the Daily treatment cost data is searched using lateral likeness method；
Abnormal correcting module, for being modified to the exceptional value using average enthesis.
17. according to claim 15 months maximum power load prediction devices, it is characterised in that the abnormal amending unit
Including：
Abnormal determining module, for from deleting unrelidble data Xi in the n Daily treatment cost data, calculating (n-1) individual day most
The meansigma methodss of big load dataWith standard deviation δ, ifThen think that Xi is exceptional value；
Replace module, for using the meansigma methodssReplace the suspicious data Xi；
Wherein, K is the difference of actual value and standard value.
18. according to claim 12 months maximum power load prediction devices, it is characterised in that the predicting unit bag
Include：
Incremental computations module, for calculating load maximum of each moon relative to last month according to the moon maximum power load data
Increment；
Incremental forecasting model building module, for setting up chain rate incremental forecasting model according to the load maximum increment.
19. according to claim 12 months maximum power load prediction devices, it is characterised in that the predicting unit bag
Include：
Sequential growth rate computing module, for calculating chain rate of each moon relative to last month according to the moon maximum power load data
Rate of increase；
Rate of increase forecast model sets up module, for setting up sequential growth rate forecast model according to the sequential growth rate.
20. according to claim 19 months maximum power load prediction devices, it is characterised in that also include：
Natural law computing unit, for calculating the actual natural law that each moon and last month be separated by；
Amending unit, for being modified to the chain rate incremental forecasting model according to the actual natural law.
21. according to claim 12 months maximum power load prediction devices, it is characterised in that also include：
Range setting module, for setting the range of predicted value of every month；
Comparing unit, for the predictive value that moon maximum power load data prediction is obtained is compared with the range of predicted value
Compared with；
Predictive value replacement unit, replaces the predictive value for the maximum with the range of predicted value or minima.
22. according to claim 12 months maximum power load prediction devices, it is characterised in that also include：
Error calculation unit, for being calculated definitely by mistake according to moon maximum power load actual value and moon maximum power predicted load
Difference or relative error；
Modifying model unit, for the absolute error or relative error are fed back to chain rate incremental forecasting model or sequential growth rate
Rate forecast model carries out Modifying model.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201610930965.XA CN106529704A (en) | 2016-10-31 | 2016-10-31 | Monthly maximum power load forecasting method and apparatus |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201610930965.XA CN106529704A (en) | 2016-10-31 | 2016-10-31 | Monthly maximum power load forecasting method and apparatus |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN106529704A true CN106529704A (en) | 2017-03-22 |
Family
ID=58293017
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201610930965.XA Pending CN106529704A (en) | 2016-10-31 | 2016-10-31 | Monthly maximum power load forecasting method and apparatus |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN106529704A (en) |
Cited By (14)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN107292429A (en) * | 2017-06-07 | 2017-10-24 | 上海欧睿供应链管理有限公司 | A kind of Demand Forecast Model system of selection analyzed based on demand characteristics |
| CN107688525A (en) * | 2017-07-26 | 2018-02-13 | 深圳市小牛在线互联网信息咨询有限公司 | The safe TPS of system computational methods, device, storage medium and computer equipment |
| CN108681542A (en) * | 2018-02-12 | 2018-10-19 | 阿里巴巴集团控股有限公司 | A kind of method and device of abnormality detection |
| CN109474077A (en) * | 2018-12-29 | 2019-03-15 | 国网辽宁省电力有限公司电力科学研究院 | Intelligent power distribution switchgear control system and electric load prediction technique |
| CN109492048A (en) * | 2019-01-21 | 2019-03-19 | 国网河北省电力有限公司经济技术研究院 | A kind of extracting method, system and the terminal device of power consumer electrical characteristics |
| CN109636035A (en) * | 2018-12-12 | 2019-04-16 | 北京天诚同创电气有限公司 | Load forecasting model creation method and device, Methods of electric load forecasting and device |
| CN109726197A (en) * | 2018-12-05 | 2019-05-07 | 新奥数能科技有限公司 | A kind of generating set rate of load condensate missing value complement trick method and device |
| CN109918415A (en) * | 2019-02-21 | 2019-06-21 | 智恒科技股份有限公司 | A kind of method and system of the water utilities data prediction of data warehouse technology |
| CN109947812A (en) * | 2018-07-09 | 2019-06-28 | 平安科技（深圳）有限公司 | Consecutive miss value fill method, data analysis set-up, terminal and storage medium |
| CN110096497A (en) * | 2019-03-28 | 2019-08-06 | 中国农业科学院农业信息研究所 | A kind of agricultural output data intelligence cleaning method and system |
| CN110826750A (en) * | 2018-08-08 | 2020-02-21 | 阿里巴巴集团控股有限公司 | Power load prediction method, device, equipment and system |
| CN112365070A (en) * | 2020-11-18 | 2021-02-12 | 深圳供电局有限公司 | Power load prediction method, device, equipment and readable storage medium |
| CN112465266A (en) * | 2020-12-14 | 2021-03-09 | 中国南方电网有限责任公司 | Bus load prediction accuracy analysis method and device and computer equipment |
| CN113590989A (en) * | 2020-04-30 | 2021-11-02 | 北京金山云网络技术有限公司 | Data processing method and device for real-time computing abnormality and electronic equipment |
Citations (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN102999791A (en) * | 2012-11-23 | 2013-03-27 | 广东电网公司电力科学研究院 | Power load forecasting method based on customer segmentation in power industry |
| CN103870886A (en) * | 2014-03-03 | 2014-06-18 | 上海申瑞继保电气有限公司 | Prediction method for monthly maximal power consumption of enterprise |
| CN105320999A (en) * | 2015-11-03 | 2016-02-10 | 国家电网公司 | Combination method for demand response based time-of-use power price strategy suitable for smart grid development |
-
2016
- 2016-10-31 CN CN201610930965.XA patent/CN106529704A/en active Pending
Patent Citations (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN102999791A (en) * | 2012-11-23 | 2013-03-27 | 广东电网公司电力科学研究院 | Power load forecasting method based on customer segmentation in power industry |
| CN103870886A (en) * | 2014-03-03 | 2014-06-18 | 上海申瑞继保电气有限公司 | Prediction method for monthly maximal power consumption of enterprise |
| CN105320999A (en) * | 2015-11-03 | 2016-02-10 | 国家电网公司 | Combination method for demand response based time-of-use power price strategy suitable for smart grid development |
Cited By (18)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN107292429A (en) * | 2017-06-07 | 2017-10-24 | 上海欧睿供应链管理有限公司 | A kind of Demand Forecast Model system of selection analyzed based on demand characteristics |
| CN107688525A (en) * | 2017-07-26 | 2018-02-13 | 深圳市小牛在线互联网信息咨询有限公司 | The safe TPS of system computational methods, device, storage medium and computer equipment |
| CN108681542A (en) * | 2018-02-12 | 2018-10-19 | 阿里巴巴集团控股有限公司 | A kind of method and device of abnormality detection |
| CN109947812A (en) * | 2018-07-09 | 2019-06-28 | 平安科技（深圳）有限公司 | Consecutive miss value fill method, data analysis set-up, terminal and storage medium |
| CN109947812B (en) * | 2018-07-09 | 2023-11-10 | 平安科技（深圳）有限公司 | Continuous missing value filling method, data analysis device, terminal and storage medium |
| CN110826750B (en) * | 2018-08-08 | 2023-09-26 | 阿里巴巴集团控股有限公司 | Power load prediction method, device, equipment and system |
| CN110826750A (en) * | 2018-08-08 | 2020-02-21 | 阿里巴巴集团控股有限公司 | Power load prediction method, device, equipment and system |
| CN109726197A (en) * | 2018-12-05 | 2019-05-07 | 新奥数能科技有限公司 | A kind of generating set rate of load condensate missing value complement trick method and device |
| CN109636035A (en) * | 2018-12-12 | 2019-04-16 | 北京天诚同创电气有限公司 | Load forecasting model creation method and device, Methods of electric load forecasting and device |
| CN109474077A (en) * | 2018-12-29 | 2019-03-15 | 国网辽宁省电力有限公司电力科学研究院 | Intelligent power distribution switchgear control system and electric load prediction technique |
| CN109474077B (en) * | 2018-12-29 | 2024-03-15 | 国网辽宁省电力有限公司电力科学研究院 | Intelligent power distribution switch cabinet control system and electric load prediction method |
| CN109492048A (en) * | 2019-01-21 | 2019-03-19 | 国网河北省电力有限公司经济技术研究院 | A kind of extracting method, system and the terminal device of power consumer electrical characteristics |
| CN109918415A (en) * | 2019-02-21 | 2019-06-21 | 智恒科技股份有限公司 | A kind of method and system of the water utilities data prediction of data warehouse technology |
| CN110096497A (en) * | 2019-03-28 | 2019-08-06 | 中国农业科学院农业信息研究所 | A kind of agricultural output data intelligence cleaning method and system |
| CN113590989A (en) * | 2020-04-30 | 2021-11-02 | 北京金山云网络技术有限公司 | Data processing method and device for real-time computing abnormality and electronic equipment |
| CN112365070A (en) * | 2020-11-18 | 2021-02-12 | 深圳供电局有限公司 | Power load prediction method, device, equipment and readable storage medium |
| CN112365070B (en) * | 2020-11-18 | 2024-05-31 | 深圳供电局有限公司 | Power load prediction method, device, equipment and readable storage medium |
| CN112465266A (en) * | 2020-12-14 | 2021-03-09 | 中国南方电网有限责任公司 | Bus load prediction accuracy analysis method and device and computer equipment |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN106529704A (en) | Monthly maximum power load forecasting method and apparatus | |
| CN107123982B (en) | Power distribution network reliability economic benefit analysis method based on equipment transaction | |
| CN109858758A (en) | A kind of the combination weighting appraisal procedure and system of distribution network electric energy quality | |
| CN102509173B (en) | A kind of based on markovian power system load Accurate Prediction method | |
| CN102426674B (en) | Power system load prediction method based on Markov chain | |
| CN103310286A (en) | Product order prediction method and device with time series characteristics | |
| CN112052992A (en) | Building engineering project progress prediction system and method based on deep learning | |
| CN111639783A (en) | Line loss prediction method and system based on LSTM neural network | |
| CN110009419A (en) | Improvement time series electricity sales amount prediction technique and system based on Economic Climate method | |
| CN107832881B (en) | Wind power prediction error evaluation method considering load level and wind speed segmentation | |
| CN111525615B (en) | Method and system for evaluating output characteristic of mountain photovoltaic power station based on guarantee rate | |
| CN109523077B (en) | Wind power prediction method | |
| CN110705771A (en) | Method and device for predicting and optimizing power generation power of new energy of regional power grid | |
| CN109118012A (en) | A kind of industrial dynamics various dimensions energy consumption cost prediction technique, system, storage medium and terminal | |
| CN111080481A (en) | Electric energy substitution potential gray analysis method based on Markov chain correction | |
| CN106600029A (en) | Macro-economy predictive quantization correction method based on electric power data | |
| CN108256693A (en) | A kind of photovoltaic power generation power prediction method, apparatus and system | |
| CN113610296B (en) | Method for predicting peak value of electrical load in region and method for planning investment of power grid | |
| CN112700050B (en) | Method and system for predicting ultra-short-term 1 st point power of photovoltaic power station | |
| CN107025514A (en) | The evaluation method and power transmission and transforming equipment of a kind of dynamic evaluation transformer equipment state | |
| CN106875057A (en) | A kind of electric power meter short term need Forecasting Methodology based on conditional probability adjustment | |
| CN112241804A (en) | Macroscopic economy leading index construction method and system for energy power data | |
| CN106682840A (en) | Short-term power sale quantity predicting method and system based on daily cumulative distributed electric quantity | |
| CN113780655A (en) | Steel multi-variety demand prediction method based on intelligent supply chain | |
| CN113919564A (en) | Foreign trade enterprise development evaluation system based on electric power prosperity index mining and working method thereof |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| C06 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| RJ01 | Rejection of invention patent application after publication | ||
| RJ01 | Rejection of invention patent application after publication |
Application publication date: 20170322 |