CN107665385A - A kind of short-term load forecasting method based on SVMs of micro-grid system - Google Patents
A kind of short-term load forecasting method based on SVMs of micro-grid system Download PDFInfo
- Publication number
- CN107665385A CN107665385A CN201711036155.0A CN201711036155A CN107665385A CN 107665385 A CN107665385 A CN 107665385A CN 201711036155 A CN201711036155 A CN 201711036155A CN 107665385 A CN107665385 A CN 107665385A
- Authority
- CN
- China
- Prior art keywords
- load
- mrow
- micro
- sample
- data
- 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
- 238000012706 support-vector machine Methods 0.000 title claims abstract description 21
- 238000013277 forecasting method Methods 0.000 title claims abstract description 16
- 239000013598 vector Substances 0.000 claims abstract description 31
- 238000012549 training Methods 0.000 claims abstract description 28
- 238000000034 method Methods 0.000 claims abstract description 20
- 238000012360 testing method Methods 0.000 claims abstract description 13
- 230000006870 function Effects 0.000 claims description 17
- 230000000694 effects Effects 0.000 claims description 9
- 238000006243 chemical reaction Methods 0.000 claims description 7
- 230000005494 condensation Effects 0.000 claims description 3
- 238000009833 condensation Methods 0.000 claims description 3
- 239000000203 mixture Substances 0.000 claims description 3
- 238000004458 analytical method Methods 0.000 claims description 2
- 230000009466 transformation Effects 0.000 claims description 2
- 238000002790 cross-validation Methods 0.000 abstract description 7
- 238000000513 principal component analysis Methods 0.000 abstract description 6
- 230000008859 change Effects 0.000 description 9
- 230000005611 electricity Effects 0.000 description 4
- 238000004146 energy storage Methods 0.000 description 3
- 238000007726 management method Methods 0.000 description 3
- 230000002950 deficient Effects 0.000 description 2
- 238000011161 development Methods 0.000 description 2
- 238000010438 heat treatment Methods 0.000 description 2
- 238000005457 optimization Methods 0.000 description 2
- PEDCQBHIVMGVHV-UHFFFAOYSA-N Glycerine Chemical compound OCC(O)CO PEDCQBHIVMGVHV-UHFFFAOYSA-N 0.000 description 1
- 235000006508 Nelumbo nucifera Nutrition 0.000 description 1
- 240000002853 Nelumbo nucifera Species 0.000 description 1
- 235000006510 Nelumbo pentapetala Nutrition 0.000 description 1
- 230000006835 compression Effects 0.000 description 1
- 238000007906 compression Methods 0.000 description 1
- 238000010276 construction Methods 0.000 description 1
- 238000001816 cooling Methods 0.000 description 1
- 238000005520 cutting process Methods 0.000 description 1
- 230000007547 defect Effects 0.000 description 1
- 238000010586 diagram Methods 0.000 description 1
- 238000009826 distribution Methods 0.000 description 1
- 238000005516 engineering process Methods 0.000 description 1
- 230000007774 longterm Effects 0.000 description 1
- 238000010801 machine learning Methods 0.000 description 1
- 238000013507 mapping Methods 0.000 description 1
- 238000012986 modification Methods 0.000 description 1
- 230000004048 modification Effects 0.000 description 1
- 238000010606 normalization Methods 0.000 description 1
- 230000035699 permeability Effects 0.000 description 1
- 238000004445 quantitative analysis Methods 0.000 description 1
- 230000009467 reduction Effects 0.000 description 1
- 230000001172 regenerating effect Effects 0.000 description 1
- 238000011160 research Methods 0.000 description 1
- 230000001932 seasonal effect Effects 0.000 description 1
- 238000007619 statistical method Methods 0.000 description 1
- 238000012795 verification Methods 0.000 description 1
- 230000003442 weekly effect Effects 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
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/24—Classification techniques
- G06F18/241—Classification techniques relating to the classification model, e.g. parametric or non-parametric approaches
- G06F18/2411—Classification techniques relating to the classification model, e.g. parametric or non-parametric approaches based on the proximity to a decision surface, e.g. support vector machines
-
- 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
- Engineering & Computer Science (AREA)
- Business, Economics & Management (AREA)
- Theoretical Computer Science (AREA)
- Economics (AREA)
- Physics & Mathematics (AREA)
- Strategic Management (AREA)
- Human Resources & Organizations (AREA)
- General Physics & Mathematics (AREA)
- Marketing (AREA)
- Tourism & Hospitality (AREA)
- General Business, Economics & Management (AREA)
- Health & Medical Sciences (AREA)
- Data Mining & Analysis (AREA)
- Evolutionary Computation (AREA)
- Quality & Reliability (AREA)
- General Engineering & Computer Science (AREA)
- Game Theory and Decision Science (AREA)
- Evolutionary Biology (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Entrepreneurship & Innovation (AREA)
- Bioinformatics & Computational Biology (AREA)
- Operations Research (AREA)
- Development Economics (AREA)
- Bioinformatics & Cheminformatics (AREA)
- Artificial Intelligence (AREA)
- Life Sciences & Earth Sciences (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 a kind of short-term load forecasting method based on SVMs of micro-grid system, the method selected by the use of principal component analysis as input vector, choose ten input vectors, after all training samples and test sample has been obtained, the sample data of input is normalized;And by grid data service and cross-validation method, optimize the width parameter and penalty coefficient in kernel function;Establish micro-capacitance sensor on-line load forcasting model.The short-term load forecasting method based on SVMs of the micro-grid system of the present invention, using information such as real-time weather information, historical load data, festivals or holidays, the online Real-time Load prediction of micro-capacitance sensor can be achieved.
Description
Technical field
The present invention relates to network load field, and in particular to a kind of micro-grid system based on the short-term of SVMs
Load forecasting method.
Background technology
Referring to Fig. 1, microgrid refers to that self-contr ol, protection and management can be realized, homemade micro power network system is (below
Abbreviation micro-grid system), generally comprise power distribution network, grid-connected switch, photo-voltaic power supply, wind-force power supply, energy storage, important load, secondary
Load, energy management module, load prediction module and Weather information module.Micro-grid system both may be incorporated into bulk power grid operation,
And can departs from bulk power grid independent operating.Regenerative resource and clean energy resource system access can be realized demand by micro-grid system
Management and the existing energy maximally utilize, and electronic system, energy-storage system and load be combined by micro-capacitance sensor, pass through related control
The cooperation of device processed, provides a user heat energy and electric energy, can effectively support bulk power grid in good time, play a part of the peak load that disappears.
Short-term load forecasting is the important component of microgrid energy management system, be to distributed power source, as wind-powered electricity generation,
Controllable micro- source such as photovoltaic, micro- gas turbine, diesel engine and energy storage optimizes the basis of scheduling, and prediction result will directly affect micro-
Operation of power networks strategy and power trade.Correlative study shows, compared with high micro-grid load prediction error operating cost will be caused big
Width increase.Microgrid load prediction at present is typically carried out using the load forecasting method of bulk power grid, is all based on historical load data
It is predicted with weather data, it is difficult to which Accurate Prediction part short term change comprehensively, can not support online Real-Time Scheduling system
Unite and the decision support of control and other miscellaneous functions are run to microgrid.
The characteristics of microgrid electric load is often to change, and changes with hour, day, week, year, constantly rises and falls, has larger
Periodicity and fluctuation.Meanwhile load also changes with factors such as season, weather, temperature, the levels of economic development.General micro-capacitance sensor
Supply district it is smaller, microgrid load fluctuation is bigger, regular poor.Meanwhile the weather forecasting of a small range is typically inaccurate, because
This load prediction is accomplished.
The final purpose of Load Prediction In Power Systems problem based on SVMs is to find one to have good popularization energy
Mapping of the factor from influence load of power to load.Because the performance of the load forecasting model using intelligent algorithm construction is from root
The quality and quantity of training data is depended in sheet, the SVMs as machine learning method needs to first pass through training sample
It is trained, then could be predicted with the network trained, and the precision of forecast model and generalization ability easily by
The influence of sample input variable, the select permeability of input variable turn into the key of load prediction data pretreatment.
The content of the invention
The purpose of the present invention is the defects of overcoming prior art, there is provided a kind of micro-grid system based on SVMs
Short-term load forecasting method, imported by real-time online cutting load testing and real time meteorological data, it is pre- that real-time online carries out load
Survey, substantially increase the load prediction precision of micro-grid system.
Realizing the technical scheme of above-mentioned purpose is:A kind of micro-grid system based on SVMs (Support
Vector Machine, SVM) short-term load forecasting method, comprise the following steps:
S1, the selecting step of sample characteristics:Set prediction load value as output valve, set influence load prediction factor as
Input vector, the factor for influenceing load include historical load sequence, time factor and temperature information;
The determination step of S2, training sample and test sample:Utilize principal component analysis (Principal Component
Analysis, PCA) after method analyzed historical data, choose ten dimensional input vectors of input sample data, ten dimension
Input vector include first 1 day mutually load in the same time, last week on the same day the load of synchronization, the average load of upper one week, when
Carve, the dew point of which day of one week, dry-bulb temperature, condensation vapor, whether festivals or holidays, Real-time Load and real-time temperature, it is described defeated
Enter the vector that sample data is ten dimensions, smooth and normalized is carried out to input sample data, obtains training sample and test
The sample set of sample;
S3, the normalized step of load data:, be to institute after the sample set of training sample and test sample is obtained
The input sample data for stating sample set are normalized, and every by ten dimensional input vectors of input sample data one-dimensional returns
One changes into [0,1] section, if maximum of the leading dimension on all samples is maxvalue, minimum value minvalue, does
Following linear transformation:
Wherein, x is the input vector value before conversion;For the input vector value after conversion;
S4, Kernel and parameter optimisation step:Select RBF (Radial Basis Function, RBF)
As SVM kernel function, and pass through the width parameter σ in grid data service and the cross-validation method optimization kernel function2With punish
Penalty factor c, establish the on-line load forcasting model of micro-grid system;
S5, prediction effect judge step:Judging basis of the average relative error as prediction effect are chosen,
Wherein, A (i) is actual negative charge values;F (i) is prediction load value;eMAPEFor average relative error.
A kind of 2. short-term load forecasting side based on SVMs of micro-grid system according to claim 1
Method, it is characterised in that in ten dimensional input vector, the moment is in units of hour or half an hour.
A kind of 3. short-term load forecasting side based on SVMs of micro-grid system according to claim 1
Method, it is characterised in that in the Kernel and parameter optimisation step, parameter optimisation procedure is equivalent in width parameter σ2With
Optimal value is determined on the grid of the Two Dimensional Uniform division of penalty coefficient c compositions.
The short-term load forecasting method based on SVMs of the micro-grid system of the present invention, utilizes principal component analysis
As the method for input vector selection, ten input vectors are chosen, it is right after all training samples and test sample has been obtained
The sample data of input is normalized;And by grid data service and cross-validation method, optimize the width in kernel function
Parameter σ2With penalty coefficient c;Establish micro-capacitance sensor on-line load forcasting model.The present invention passes through real-time load data and historical load
Data are combined, and train forecast model using online weather data, the Accurate Prediction of micro-grid load are realized, for micro-
Network optimizationization is run.
Brief description of the drawings
Fig. 1 is the structural representation of micro-grid system;
Fig. 2 is the flow chart of the short-term load forecasting method based on SVMs of the micro-grid system of the present invention;
The comparison diagram of the actual load that Fig. 3 is 14 days and prediction load;
Fig. 4 is the prediction load of 14 days and the average relative error of actual load.
Embodiment
In order that those skilled in the art are better understood when technical scheme, it is right below in conjunction with the accompanying drawings
Its embodiment is described in detail:
Referring to Fig. 2, highly preferred embodiment of the present invention, a kind of the bearing in short term based on SVMs of micro-grid system
Lotus Forecasting Methodology, comprises the following steps:
S1, the selecting step of sample characteristics:
Power-system short-term load forecasting is a multivariable prediction problem, and it can be seen as function regression problem.Prediction
Load value y is the output valve of function, and influences factor such as historical load sequence, time factor and the Temperature Factor of load accordingly
Deng function input vector x is used as, each composition of training data is SVM (Support Vector Machine, supporting vector
Machine) characteristic quantity input vector (each element in x is referred to as a feature).Different data sequences influences the side of model
Case, which information should be included in training data actually, and it is the load prediction based on SVM to select which factor as characteristic quantity
Committed step.It is now discussed with influenceing each factor of prediction, by studying the correlation of various factors and load, so as to combine
The size and data of correlation determine the model of input sample.
(1) historical load sequence
Historical load sequence, i.e. former load data.Generally, according to the continuity principle of load prediction and analogizing original
Then, after the change procedure for first developing things being extrapolated to the things that develops up, so as to be made prediction to the latter.Typically
Ground, the feature considered in x include the previous day or information more in early time.Precisely, if y=L- { d } represents to be predicted
The load of d days a few days ago, because it is very big with the load L correlations of day to be predicted, therefore it is elected to be the essential spy of input sample data
One of sign.
(2) time factor
Time factor has a major impact to power system load, and its performance mainly has at 3 points:The seasonal periodicity of load, the moon
Periodically, weekly pattern, festivals or holidays attribute.As short-term load forecasting, the cycle attribute of a consideration load and festivals or holidays
Attribute.Load has the feature by cycle change, and the load during great festivals or holidays is even more to have obvious fluctuation.
(3) Temperature Factor
Temperature Factor is the principal element for influenceing short-term load forecasting.Under some weather conditions, power load and electricity
Can drastically it rise, this causes the quantitative analysis of electric load and meteorological relation to turn into the research emphasis of researcher.Brief analysis
The influence that meteorological element changes to short term, it is found that the influence of temperature is the most notable, when weather acutely turns cold (heating)
When, there will be a large amount of heating (cooling) load to put into operation;And when mean temperature continues too high or too low, the phase with the former time
Same date is compared, and daily load will have a greater change.
(4) other influence factors
The economic factor such as population of power system institute overlay area, industrial level, agricultural etc. all reflect certain electric power
Workload demand, the change of load must be restricted by socio-economic development.However, influence of the economic factor to load is very
Slowly, only paid attention in long term load forecasting.Meanwhile power system is made up of many independent users, different user choosing
The time and mode for selecting electricity consumption vary, due to the randomness of its load behavior, such as the switching of user power utilization equipment and
Some accidents all will cause system total load that certain randomness, such as the influence that earthquake disaster is brought is presented.These with
Machine is often unpredictable, accordingly, it is difficult to consider in load prediction.
The determination step of S2, training sample and test sample:
For feature selection issues in the power-system short-term load forecasting algorithm based on SVM, conventional researchers do
Extensive work simultaneously determines the characteristic quantity of sample with a variety of methods.Principal component analysis (Principal Component
Analysis, PCA) as currently used solve the problems, such as the method for Input variable selection, it is upper all relatively simple in theoretical and application
Easily.On the one hand, the efficiency of prediction with compression samples space, can not only be improved;On the other hand, can also eliminate due to being deposited between variable
The reduction of forecast model generalization ability caused by correlation, so as to be effectively improved the precision of prediction of model.
The work that the present invention is done according to conventional researcher, after historical data is analyzed, it is defined below inputting sample
The input vector of notebook data:
(1) first 1 day load mutually in the same time;
(2) load of the last week of synchronization on the same day;
(3) average load of one week on;
(4) moment (hour or half an hour be unit);
Which day of (5) one weeks;
(6) dry-bulb temperature;
(7) dew point of condensation vapor;
(8) whether festivals or holidays;
(9) Real-time Load;
(10) real-time temperature.
Input sample data are the vector of ten dimensions, and smooth and normalized, composing training sample are carried out to historical data
With the sample set of test sample.
S3, the normalized step of load data
After the sample set of all training samples and test sample is obtained, the input sample data of sample set are carried out
Normalized, main cause are:
(l) data for avoiding changing in a big way flood the data of smaller range change;
(2) avoid occurring numerical value difficulty in calculating, it is such as linear because core value needs to calculate the inner product of characteristic vector in calculating
Core and polynomial kernel, big characteristic value may cause numerical computations difficult.
Normalization is carried out according to dimension, i.e., one-dimensional the every of ten dimensional input vectors is normalized into required section
It is interior.Assuming that maximum of the leading dimension on all samples is maxvalue, minimum value minvalue, then can do such as lower linear
Conversion:
Wherein, x is the input vector value before conversion;For the input vector value after conversion;Thus x is mapped as [0,1]
Section.
S4, Kernel and parameter optimisation step:
The selection of kernel function influences very big, kernel functions of the present invention selection RBF as SVM on the precision of load prediction.It is logical
Cross it is a large amount of experimental studies have found that, the width parameter σ in kernel function2SVM performance is played with penalty coefficient c extremely important
Effect.
Width parameter σ2Value it is too small when, can cause SVM to training set cross study phenomenon.Conversely, work as width parameter σ2's
When value is too big, deficient study phenomenons of the SVM to training set can be caused.So width parameter σ2Extensive study to SVM plays key
Effect.When respectively the width parameter σ in kernel function2With insensitive loss parameter ε:When being fixed on 10 and 0.001, training set
Originally standard deviation reduces as penalty coefficient c increase occurs, then saturation is constant, situation about then increasing;It is another
Aspect, with penalty coefficient c increase, the training time can also increase therewith.This illustrates that small penalty coefficient c values can be to training number
According to deficient study phenomenon is caused, penalty coefficient c values are too big, and study phenomenon was easily caused to training data and causes Generalization Capability to drop
It is low.When the width parameter expansion in penalty coefficient and kernel function is all immobilized respectively, the standard of training set and test set is equal
Variance is relatively stable, is influenceed by insensitive loss parameter ε value changes little.However, the number of supporting vector is but with unwise
Feel loss parameter ε increase and reduce, while also there is a certain degree of fluctuation the training time.This explanation, insensitive loss parameter ε
Be worth influences less on SVM precision of prediction and generalization ability.
In the present invention, because the quantity of training sample is relatively large, extremely important effect is played to SVM performance
Two parameters carry out optimizing, i.e., the width parameter σ in kernel function2With penalty coefficient c.Efficiency and practicality in view of calculating
Property, grid search (Grid-search) and cross-validation method (CrossValidation, CV) are directly selected in kernel function
Width parameter σ2Selection optimization is carried out with penalty coefficient c.Parameter optimisation procedure is equivalent in width parameter σ2With penalty coefficient c groups
Into Two Dimensional Uniform division grid on determine optimal value.
Cross-validation method is for verifying a kind of statistical analysis technique of the performance of grader, and basic thought is in certain meaning
Initial data (dataset) is grouped under justice, a part collects as training set (trainset), another part as checking
(validationset), grader is trained with training set first, the model that training obtains is being tested using checking collection
(model) performance indications as classification of assessment device, are carried out with this.Conventional CV methods are k times of cross validation (k-
FoldCrossValidation), it is designated as k-CV.
S5, prediction effect judge step:
For the validity of verification algorithm, average relative error is taken as prediction effect judging basis, i.e.,
Wherein, A (i) is actual negative charge values;F (i) is prediction load value;eMAPEFor average relative error.
Fig. 3 and Fig. 4 are referred to, is carried out 14 days using the short-term load forecasting method based on SVMs of the present invention
Load prediction, it predicts load value compared with actual negative charge values, average relative error eMAPEIt is of the invention based on branch for 2.27%
The Accurate Prediction of micro-grid load can be realized by holding the short-term load forecasting method of vector machine.
In summary, the short-term load forecasting method based on SVMs of micro-grid system of the invention, passes through
Real-time load data is combined with historical load data, and using online weather data training forecast model, realizes micro- electricity
The Accurate Prediction of net load, optimize for microgrid and run.
Those of ordinary skill in the art it should be appreciated that the embodiment of the above be intended merely to explanation the present invention,
And be not used as limitation of the invention, as long as in the spirit of the present invention, the change to embodiment described above
Change, modification will all fall in the range of claims of the present invention.
Claims (3)
1. the short-term load forecasting method based on SVMs of a kind of micro-grid system, it is characterised in that including following
Step:
S1, the selecting step of sample characteristics:Prediction load value is set as output valve, sets the factor for influenceing load prediction as input
Vector, the factor for influenceing load include historical load sequence, time factor and temperature information;
The determination step of S2, training sample and test sample:After being analyzed using principal component analytical method historical data, choosing
Take ten dimensional input vectors of input sample data, it is same that ten dimensional input vector includes first 1 day load mutually in the same time, last week
The load of its synchronization, the average load of upper one week, which day at moment, one week, dry-bulb temperature, condensation vapor dew point,
Whether festivals or holidays, Real-time Load and real-time temperature, the input sample data be ten dimension vectors, to input sample data carry out
Smooth and normalized, obtains training sample and the sample set of test sample;
S3, the normalized step of load data:, be to the sample after the sample set of training sample and test sample is obtained
The input sample data of this collection are normalized, and every by ten dimensional input vectors of input sample data one-dimensional normalizes
Into [0,1] section, if maximum of the leading dimension on all samples is maxvalue, minimum value minvalue, do as follows
Linear transformation:
<mrow>
<mover>
<mi>x</mi>
<mo>&OverBar;</mo>
</mover>
<mo>=</mo>
<mfrac>
<mrow>
<mi>x</mi>
<mo>-</mo>
<mi>min</mi>
<mi>v</mi>
<mi>a</mi>
<mi>l</mi>
<mi>u</mi>
<mi>e</mi>
</mrow>
<mrow>
<mi>max</mi>
<mi>v</mi>
<mi>a</mi>
<mi>l</mi>
<mi>u</mi>
<mi>e</mi>
<mo>-</mo>
<mi>min</mi>
<mi>v</mi>
<mi>a</mi>
<mi>l</mi>
<mi>u</mi>
<mi>e</mi>
</mrow>
</mfrac>
</mrow>
Wherein, x is the input vector value before conversion;For the input vector value after conversion;
S4, Kernel and parameter optimisation step:Kernel functions of the RBF as SVM is selected, and passes through grid data service and intersection
Proof method optimizes the width parameter σ in the kernel function2With penalty coefficient c, the on-line load forcasting mould of micro-grid system is established
Type;
S5, prediction effect judge step:Judging basis of the average relative error as prediction effect are chosen,
<mrow>
<msub>
<mi>e</mi>
<mrow>
<mi>M</mi>
<mi>A</mi>
<mi>P</mi>
<mi>E</mi>
</mrow>
</msub>
<mo>=</mo>
<mfrac>
<mn>1</mn>
<mi>n</mi>
</mfrac>
<munderover>
<mo>&Sigma;</mo>
<mrow>
<mi>i</mi>
<mo>=</mo>
<mn>1</mn>
</mrow>
<mi>n</mi>
</munderover>
<mfrac>
<mrow>
<mo>|</mo>
<mi>A</mi>
<mrow>
<mo>(</mo>
<mi>i</mi>
<mo>)</mo>
</mrow>
<mo>-</mo>
<mi>F</mi>
<mrow>
<mo>(</mo>
<mi>i</mi>
<mo>)</mo>
</mrow>
<mo>|</mo>
</mrow>
<mrow>
<mi>A</mi>
<mrow>
<mo>(</mo>
<mi>i</mi>
<mo>)</mo>
</mrow>
</mrow>
</mfrac>
<mo>&times;</mo>
<mn>100</mn>
<mi>%</mi>
</mrow>
Wherein, A (i) is actual negative charge values;F (i) is prediction load value;eMAPEFor average relative error.
2. a kind of short-term load forecasting method based on SVMs of micro-grid system according to claim 1,
Characterized in that, in ten dimensional input vector, the moment is in units of hour or half an hour.
3. a kind of short-term load forecasting method based on SVMs of micro-grid system according to claim 1,
Characterized in that, in the Kernel and parameter optimisation step, parameter optimisation procedure is equivalent in width parameter σ2With punish
Optimal value is determined on the grid of the Two Dimensional Uniform division of penalty factor c compositions.
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201711036155.0A CN107665385A (en) | 2017-10-30 | 2017-10-30 | A kind of short-term load forecasting method based on SVMs of micro-grid system |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201711036155.0A CN107665385A (en) | 2017-10-30 | 2017-10-30 | A kind of short-term load forecasting method based on SVMs of micro-grid system |
Publications (1)
Publication Number | Publication Date |
---|---|
CN107665385A true CN107665385A (en) | 2018-02-06 |
Family
ID=61144614
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN201711036155.0A Pending CN107665385A (en) | 2017-10-30 | 2017-10-30 | A kind of short-term load forecasting method based on SVMs of micro-grid system |
Country Status (1)
Country | Link |
---|---|
CN (1) | CN107665385A (en) |
Cited By (9)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN108304976A (en) * | 2018-03-06 | 2018-07-20 | 西安交通大学 | A kind of electric system fining load prediction and analysis method |
CN109345027A (en) * | 2018-10-25 | 2019-02-15 | 国网江苏省电力有限公司盐城供电分公司 | Micro-capacitance sensor short-term load forecasting method based on independent component analysis and support vector machines |
CN109344990A (en) * | 2018-08-02 | 2019-02-15 | 中国电力科学研究院有限公司 | A kind of short-term load forecasting method and system based on DFS and SVM feature selecting |
CN109473954A (en) * | 2018-10-15 | 2019-03-15 | 上海电力学院 | A kind of guard method of novel microgrid failure |
CN110866633A (en) * | 2019-10-25 | 2020-03-06 | 上海电气集团股份有限公司 | Micro-grid ultra-short term load prediction method based on SVR support vector regression |
CN111709548A (en) * | 2020-04-27 | 2020-09-25 | 深圳华工能源技术有限公司 | Power consumer load prediction method, device, equipment and storage medium based on support vector machine |
CN112712203A (en) * | 2020-12-29 | 2021-04-27 | 湖南大学 | Method and system for predicting daily maximum load of power distribution network |
CN113657459A (en) * | 2021-07-28 | 2021-11-16 | 合肥国轩高科动力能源有限公司 | Battery SOC prediction method and medium based on combination of principal component and support vector machine |
CN116050667A (en) * | 2023-03-22 | 2023-05-02 | 国网山东省电力公司烟台供电公司 | Distributed photovoltaic-oriented intelligent power grid load prediction management system and management method |
Citations (7)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN101930486A (en) * | 2010-07-12 | 2010-12-29 | 沈阳工业大学 | Device and method for predicating fan load index of wind powder plant |
CN102214262A (en) * | 2010-04-02 | 2011-10-12 | 上海海洋大学 | Tide predicting method |
CN103488869A (en) * | 2013-08-23 | 2014-01-01 | 上海交通大学 | Wind power generation short-term load forecast method of least squares support vector machine |
CN104239689A (en) * | 2014-08-18 | 2014-12-24 | 华北电力大学 | Optimal relevance vector machine based short-term wind speed prediction method |
CN105303262A (en) * | 2015-11-12 | 2016-02-03 | 河海大学 | Short period load prediction method based on kernel principle component analysis and random forest |
CN106372763A (en) * | 2016-10-25 | 2017-02-01 | 国家电网公司 | Improved-support-vector-machine-based power load prediction correction method |
CN106651020A (en) * | 2016-12-16 | 2017-05-10 | 燕山大学 | Short-term power load prediction method based on big data reduction |
-
2017
- 2017-10-30 CN CN201711036155.0A patent/CN107665385A/en active Pending
Patent Citations (7)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN102214262A (en) * | 2010-04-02 | 2011-10-12 | 上海海洋大学 | Tide predicting method |
CN101930486A (en) * | 2010-07-12 | 2010-12-29 | 沈阳工业大学 | Device and method for predicating fan load index of wind powder plant |
CN103488869A (en) * | 2013-08-23 | 2014-01-01 | 上海交通大学 | Wind power generation short-term load forecast method of least squares support vector machine |
CN104239689A (en) * | 2014-08-18 | 2014-12-24 | 华北电力大学 | Optimal relevance vector machine based short-term wind speed prediction method |
CN105303262A (en) * | 2015-11-12 | 2016-02-03 | 河海大学 | Short period load prediction method based on kernel principle component analysis and random forest |
CN106372763A (en) * | 2016-10-25 | 2017-02-01 | 国家电网公司 | Improved-support-vector-machine-based power load prediction correction method |
CN106651020A (en) * | 2016-12-16 | 2017-05-10 | 燕山大学 | Short-term power load prediction method based on big data reduction |
Cited By (12)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN108304976A (en) * | 2018-03-06 | 2018-07-20 | 西安交通大学 | A kind of electric system fining load prediction and analysis method |
CN109344990A (en) * | 2018-08-02 | 2019-02-15 | 中国电力科学研究院有限公司 | A kind of short-term load forecasting method and system based on DFS and SVM feature selecting |
CN109473954A (en) * | 2018-10-15 | 2019-03-15 | 上海电力学院 | A kind of guard method of novel microgrid failure |
CN109345027A (en) * | 2018-10-25 | 2019-02-15 | 国网江苏省电力有限公司盐城供电分公司 | Micro-capacitance sensor short-term load forecasting method based on independent component analysis and support vector machines |
CN109345027B (en) * | 2018-10-25 | 2021-11-23 | 国网江苏省电力有限公司盐城供电分公司 | Micro-grid short-term load prediction method based on independent component analysis and support vector machine |
CN110866633A (en) * | 2019-10-25 | 2020-03-06 | 上海电气集团股份有限公司 | Micro-grid ultra-short term load prediction method based on SVR support vector regression |
CN110866633B (en) * | 2019-10-25 | 2023-11-24 | 上海电气集团股份有限公司 | Micro-grid ultra-short-term load prediction method based on SVR support vector regression |
CN111709548A (en) * | 2020-04-27 | 2020-09-25 | 深圳华工能源技术有限公司 | Power consumer load prediction method, device, equipment and storage medium based on support vector machine |
CN112712203A (en) * | 2020-12-29 | 2021-04-27 | 湖南大学 | Method and system for predicting daily maximum load of power distribution network |
CN112712203B (en) * | 2020-12-29 | 2023-09-15 | 湖南大学 | Day-highest load prediction method and system for power distribution network |
CN113657459A (en) * | 2021-07-28 | 2021-11-16 | 合肥国轩高科动力能源有限公司 | Battery SOC prediction method and medium based on combination of principal component and support vector machine |
CN116050667A (en) * | 2023-03-22 | 2023-05-02 | 国网山东省电力公司烟台供电公司 | Distributed photovoltaic-oriented intelligent power grid load prediction management system and management method |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
CN107665385A (en) | A kind of short-term load forecasting method based on SVMs of micro-grid system | |
Wu et al. | Site selection decision framework using fuzzy ANP-VIKOR for large commercial rooftop PV system based on sustainability perspective | |
Mavromatidis et al. | Design of distributed energy systems under uncertainty: A two-stage stochastic programming approach | |
Zhou et al. | Application of a novel discrete grey model for forecasting natural gas consumption: A case study of Jiangsu Province in China | |
Czaplicka-Kolarz et al. | Technology foresight for a vision of energy sector development in Poland till 2030. Delphi survey as an element of technology foresighting | |
CN110222882A (en) | A kind of prediction technique and device of electric system Mid-long Term Load | |
Xiao et al. | Multistage decision support framework for sites selection of solar power plants with probabilistic linguistic information | |
CN112950098B (en) | Comprehensive energy system-based energy planning method and device and terminal equipment | |
CN111754037B (en) | Long-term load hybrid prediction method for regional terminal integrated energy supply system | |
CN106251008A (en) | A kind of photovoltaic power Forecasting Methodology chosen based on combining weights similar day | |
Hamzaçebi | Primary energy sources planning based on demand forecasting: The case of Turkey | |
CN111222700A (en) | Day-ahead electricity price probability prediction method based on dynamic network quantile model | |
Badi et al. | Using Grey-TOPSIS approach for solar farm location selection in Libya | |
CN112633762A (en) | Building energy efficiency obtaining method and equipment | |
Pakere et al. | Solar power or solar heat: What will upraise the efficiency of district heating? Multi-criteria analyses approach | |
Watson et al. | Sustainable energy transition: Intermittency policy based on digital mirror actions | |
Omar et al. | Seasonal clustering forecasting technique for intelligent hourly solar irradiance systems | |
Li et al. | Optimal design and operation method of integrated energy system based on stochastic scenario iteration and energy storage attribute evaluation | |
Titz et al. | Identifying drivers and mitigators for congestion and redispatch in the German electric power system with explainable AI | |
CN113887809A (en) | Power distribution network supply and demand balance method, system, medium and computing equipment under double-carbon target | |
Haouraji et al. | A correlative approach, combining energy consumption, urbanization and GDP, for modeling and forecasting Morocco's residential energy consumption | |
Zheng et al. | Short-term energy consumption prediction of electric vehicle charging station using attentional feature engineering and multi-sequence stacked Gated Recurrent Unit | |
Refaai et al. | Energy Management Prediction in Hybrid PV‐Battery Systems Using Deep Learning Architecture | |
Liu et al. | An inexact stochastic–fuzzy jointed chance-constrained programming for regional energy system management under uncertainty | |
Makri et al. | Short term net imbalance volume forecasting through machine and deep learning: a UK case study |
Legal Events
Date | Code | Title | Description |
---|---|---|---|
PB01 | Publication | ||
PB01 | Publication | ||
SE01 | Entry into force of request for substantive examination | ||
SE01 | Entry into force of request for substantive examination | ||
WD01 | Invention patent application deemed withdrawn after publication |
Application publication date: 20180206 |
|
WD01 | Invention patent application deemed withdrawn after publication |