CN109426889A - Short-term load forecasting method based on KPCA in conjunction with improvement neural network - Google Patents
Short-term load forecasting method based on KPCA in conjunction with improvement neural network Download PDFInfo
- Publication number
- CN109426889A CN109426889A CN201710780654.4A CN201710780654A CN109426889A CN 109426889 A CN109426889 A CN 109426889A CN 201710780654 A CN201710780654 A CN 201710780654A CN 109426889 A CN109426889 A CN 109426889A
- Authority
- CN
- China
- Prior art keywords
- neural network
- short
- value
- model
- load forecasting
- 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
- 238000013528 artificial neural network Methods 0.000 title claims abstract description 54
- 230000006872 improvement Effects 0.000 title claims abstract description 14
- 238000013277 forecasting method Methods 0.000 title claims abstract description 13
- 230000009467 reduction Effects 0.000 claims abstract description 20
- 238000003062 neural network model Methods 0.000 claims abstract description 13
- 239000002245 particle Substances 0.000 claims description 26
- 239000011159 matrix material Substances 0.000 claims description 16
- 238000012549 training Methods 0.000 claims description 12
- 230000001537 neural effect Effects 0.000 claims description 5
- 239000000284 extract Substances 0.000 claims description 4
- 239000000203 mixture Substances 0.000 claims description 4
- 230000008859 change Effects 0.000 claims description 3
- 238000000034 method Methods 0.000 abstract description 23
- 238000000513 principal component analysis Methods 0.000 description 6
- 230000000694 effects Effects 0.000 description 4
- 235000006508 Nelumbo nucifera Nutrition 0.000 description 3
- 240000002853 Nelumbo nucifera Species 0.000 description 3
- 235000006510 Nelumbo pentapetala Nutrition 0.000 description 3
- 230000006870 function Effects 0.000 description 3
- 230000007812 deficiency Effects 0.000 description 2
- 238000013507 mapping Methods 0.000 description 2
- 238000005457 optimization Methods 0.000 description 2
- 238000012360 testing method Methods 0.000 description 2
- 230000008901 benefit Effects 0.000 description 1
- 238000011161 development Methods 0.000 description 1
- 238000010586 diagram Methods 0.000 description 1
- 238000011156 evaluation Methods 0.000 description 1
- 238000003780 insertion Methods 0.000 description 1
- 230000037431 insertion Effects 0.000 description 1
- 238000011160 research Methods 0.000 description 1
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06Q—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
- G06Q10/00—Administration; Management
- G06Q10/04—Forecasting or optimisation specially adapted for administrative or management purposes, e.g. linear programming or "cutting stock problem"
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/004—Artificial life, i.e. computing arrangements simulating life
- G06N3/006—Artificial life, i.e. computing arrangements simulating life based on simulated virtual individual or collective life forms, e.g. social simulations or particle swarm optimisation [PSO]
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/02—Neural networks
-
- 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)
- Physics & Mathematics (AREA)
- Health & Medical Sciences (AREA)
- Economics (AREA)
- General Physics & Mathematics (AREA)
- Strategic Management (AREA)
- General Health & Medical Sciences (AREA)
- Human Resources & Organizations (AREA)
- Life Sciences & Earth Sciences (AREA)
- Mathematical Physics (AREA)
- Software Systems (AREA)
- General Business, Economics & Management (AREA)
- Marketing (AREA)
- Tourism & Hospitality (AREA)
- General Engineering & Computer Science (AREA)
- Artificial Intelligence (AREA)
- Biomedical Technology (AREA)
- Biophysics (AREA)
- Computational Linguistics (AREA)
- Data Mining & Analysis (AREA)
- Evolutionary Computation (AREA)
- Molecular Biology (AREA)
- Computing Systems (AREA)
- Public Health (AREA)
- Water Supply & Treatment (AREA)
- Primary Health Care (AREA)
- Development Economics (AREA)
- Game Theory and Decision Science (AREA)
- Entrepreneurship & Innovation (AREA)
- Operations Research (AREA)
- Quality & Reliability (AREA)
- Management, Administration, Business Operations System, And Electronic Commerce (AREA)
Abstract
The invention discloses a kind of short-term load forecasting methods based on KPCA in conjunction with improvement neural network.This method comprises the following steps: (1) it analyzes, choose the principal element and historical data for influencing load, it is preliminary to constitute neural network sample collection；(2) dimensionality reduction decoupling is carried out to input neural network sample collection using core principle component analysis algorithm；(3) the neural network sample collection after decoupling dimensionality reduction as the input quantity for improving neural network model and is trained to obtain prediction model；It (4) will be in forecast sample input trained prediction model；(5) output valve of prediction model is modified, as short-term load forecasting value.The present invention carries out short-term load forecasting to electric system with the model improved in conjunction with neural network by core principle component analysis, simplifies model structure and accelerates convergence efficiency；By being corrected to output valve, the accidental error of model output is reduced, the precision of load prediction is improved.
Description
Technical field
The present invention relates to technical field of power systems, more particularly to it is a kind of based on KPCA with improve in conjunction with neural network
Short-term load forecasting method.
Background technique
Load prediction be carry out Power System Planning, the premise and basis of scheduling, scientific prediction be correct decisions according to
According to and guarantee.Load prediction is found out existing for history value and predicted value by analyzing, studying known meteorological and historical data
Nonlinear Mapping relationship makes load development and pre-estimates and speculate.The temporally difference of dimension, load prediction can be divided into super
In short term, in short term and Mid-long term load forecasting.
1991, neural network was arrived in the use in load prediction to the scholars such as Park.D.C for the first time, and load prediction is ground
Study carefully and entered intelligent algorithm forecast period, and receives the concern and research of more and more scholars.Neural network algorithm reason
By the upper ability for having and being fitted any Nonlinear Mapping, and have the characteristics that method is easy to operate, precision of prediction is high, at
For one of the common method in load prediction field.
There is following problems for traditional neural network model: (1) initial threshold of neural network and weight are difficult to
It determines, makes network convergence rate slack-off in the way of randomly selecting, while easily falling into local minimum；(2) due to electric power
Load is affected by many factors, and prediction algorithm input dimension is high and intercouples, and causes model structure complicated, learning training is tired
It is difficult；(3) common linear dimensionality reduction mode (such as PCA) to be based on the subspace in insertion high-dimensional data space be it is linear or
This precondition of approximately linear, there is certain deficiencies；(4) there are accidental errors for network output, only once to predict reality
Result is tested as final predicted value and lacks preciseness and science.
Summary of the invention
The purpose of the present invention is to provide it is a kind of based on KPCA with improve neural network in conjunction with short-term load forecasting method,
To improve precision of prediction and efficiency.
The technical solution for realizing the aim of the invention is as follows: a kind of short-term negative in conjunction with improvement neural network based on KPCA
Lotus prediction technique, comprises the following steps:
(1) it analyzes, choose the factor and historical data for influencing load, it is preliminary to constitute neural network sample collection；
(2) dimensionality reduction decoupling is carried out to input neural network sample collection using core principle component analysis algorithm；
(3) the neural network sample collection after decoupling dimensionality reduction as the input quantity for improving neural network model and is instructed
Get prediction model；
It (4) will be in forecast sample input step (3) trained prediction model；
(5) output valve of prediction model is modified, as short-term load forecasting value.
Further, neural network sample collection described in step (1) includes historical data, temperature, week type and date
Type.
Further, core principle component analysis algorithm described in step (2), includes the following steps:
A) n sample, each sample are contained to the sample set X ∈ R of d variable compositiond×nIt is standardized；
B) selection gaussian kernel function carries out inner product calculating, obtains matrix K；
C) centralization nuclear matrixIt is wherein the matrix that A is N × N, each element
For 1/N；
D) to matrixIt is PCA, finds out eigen vector；
E) contribution rate and contribution rate of accumulative total of principal component are found out, the dimension after determining dimensionality reduction extracts principal component amount.
Further, step (3) the improvement neural network model, i.e., using particle swarm algorithm to neural network initial value
Selection optimize, it is specific as follows:
A) weight of neural network and threshold value as the initial position of particle and are initialized into particle group parameters；
B) using neural metwork training error as the fitness of particle swarm algorithm；
C) individual optimal value and global optimum are found；
D) speed of more new particle and position；
E) global optimum of particle is exported in specified iterative steps or error；
F) using optimal value as the initial weight of neural network algorithm and threshold value；
G) training neural network model is until iterative steps terminate or set error less than initial.
Further, the output valve of prediction model is modified described in step (5), comprising the following steps: with prediction
Based on continuous 10 output valves of model, it is averaged；By 10 output valves compared with average value, if output valve and flat
The error of mean value is greater than 10% and rejects the biggish value of the fluctuation；Remaining output valve is averaged again as final mould
Type output valve, i.e. short-term load forecasting value.
Compared with prior art, the present invention its remarkable advantage are as follows: (1) using core principle component analysis to input quantity dimensionality reduction, letter
Change model structure and accelerates convergence efficiency；(2) selection for utilizing particle swarm algorithm optimization neural network initial value, overcomes tradition
Neural network convergence easily falls into the limitation of Local Minimum slowly；(3) it is corrected using to output valve, avoids the accidental of model output
Error, improves the precision of prediction, and prediction numerical value is more representative.
Detailed description of the invention
Fig. 1 is overview flow chart of the invention.
Fig. 2 is the flow chart that population improves neural network algorithm in the present invention；
Fig. 3 is prediction effect schematic diagram of the present invention.
Specific embodiment
The present invention is further described with somewhere short-term load forecasting example with reference to the accompanying drawing.
In order to improve the precision and efficiency of short-term load forecasting, the present invention proposes a kind of based on core principle component analysis and improvement
The short-term load forecasting method that neural network combines.As shown in Figure 1, introducing core principle component analysis to input quantity dimensionality reduction, simplify
Model structure increases convergence efficiency；Using the selection of particle swarm algorithm optimization neural network initial value, traditional neural net is overcome
Network convergence easily falls into the limitation of Local Minimum slowly；Finally output valve is corrected, avoids the accidental error of model output, is improved
The precision of prediction, prediction numerical value are more representative.
The present invention is based on short-term load forecasting method of the KPCA in conjunction with improvement neural network, comprise the following steps:
(1) it analyzes, choose the factor and historical data for influencing load, it is preliminary to constitute neural network sample collection；
The neural network sample collection includes historical data, temperature, week type and date type.
(2) dimensionality reduction decoupling is carried out to input neural network sample collection using core principle component analysis algorithm；
The core principle component analysis algorithm, includes the following steps:
A) n sample, each sample are contained to the sample set X ∈ R of d variable compositiond×nIt is standardized；
B) selection gaussian kernel function carries out inner product calculating, obtains matrix K；
C) centralization nuclear matrixIt is wherein the matrix that A is N × N, each element
For 1/N；
D) to matrixIt is PCA, finds out eigen vector；
E) contribution rate and contribution rate of accumulative total of principal component are found out, the dimension after determining dimensionality reduction extracts principal component amount.
(3) the neural network sample collection after decoupling dimensionality reduction as the input quantity for improving neural network model and is instructed
Get prediction model；
The improvement neural network model optimizes the selection of neural network initial value using particle swarm algorithm,
It is specific as follows:
A) weight of neural network and threshold value as the initial position of particle and are initialized into particle group parameters；
B) using neural metwork training error as the fitness of particle swarm algorithm；
C) individual optimal value and global optimum are found；
D) speed of more new particle and position；
E) global optimum of particle is exported in specified iterative steps or error；
F) using optimal value as the initial weight of neural network algorithm and threshold value；
G) training neural network model is until iterative steps terminate or set error less than initial.
It (4) will be in forecast sample input step (3) trained prediction model；
(5) output valve of prediction model is modified, as short-term load forecasting value.
The output valve to prediction model is modified, comprising the following steps: with continuous 10 output of prediction model
Based on value, it is averaged；By 10 output valves compared with average value, if output valve and the error of average value are greater than 10%
Then reject the biggish value of the fluctuation；Remaining output valve is averaged again as final model output value, i.e., it is short-term negative
Lotus predicted value.
Embodiment 1
It is of the present invention to include with the short-term load forecasting method improved in conjunction with neural network based on core principle component analysis
Following steps:
Step (1): the principal element and historical data for influencing load are chosen in analysis, preliminary to constitute neural network sample collection.
Electric load is mainly by historical load, and temperature, the factors such as date type influence, therefore choose and predict two days a few days ago loads in the same time
Value, prediction two degree/day a few days ago, date type, week type totally 106 parameters, output layer are to predict that day every 30min's is negative
Lotus data amount to 48 nodes.Week and date type are quantified, are set as 1 to 7 week, except the Spring Festival, New Year's Day are set as 2 in the date,
Remaining festivals or holidays is set as 1, and common day is set as 0.Using somewhere 50 days a few days ago data to be predicted as training set.
Step (2): input data dimensionality reduction is decoupled using core principle component analysis.Core principle component analysis (KPCA) is to utilize core
Expansion of the technology to linear algorithm principal component analysis (PCA).Basic thought be first be mapped in the higher linear space of dimension, then
Space dimensionality reduction is realized using linear algorithm.Core principle component analysis mainly includes the following steps:
N sample, each sample are contained to the sample set X ∈ R of d variable composition firstd×nIt is standardized；
It selects gaussian kernel function to carry out inner product calculating, obtains matrix K；Centralization nuclear matrixWherein
The matrix for being N × N for A, each element are 1/N；To matrixIt is PCA, finds out eigen vector；It is close by applying
Special orthogonalization method, orthogonalization and unitization feature vector；The contribution rate and contribution rate of accumulative total for finding out principal component, determine dimensionality reduction
Dimension afterwards extracts principal component amount.
It is as shown in table 1 the contribution rate and contribution rate of accumulative total of KPCA analysis.
1 eigenvalue contribution rate of table and contribution rate of accumulative total
| Number | Characteristic value | Contribution rate (%) | Contribution rate of accumulative total (%) |
| 1 | 3.1827 | 56.6436 | 56.6436 |
| 2 | 1.1525 | 20.5108 | 77.1544 |
| 3 | 0.3696 | 6.5779 | 83.7323 |
| 4 | 0.1736 | 3.0896 | 86.8219 |
| 5 | 0.1179 | 2.0978 | 88.9197 |
| 6 | 0.0842 | 1.4992 | 90.4188 |
| 7 | 0.0582 | 1.0353 | 91.4541 |
| 8 | 0.0479 | 0.8531 | 92.3072 |
| 9 | 0.0384 | 0.6829 | 92.9901 |
| 10 | 0.0358 | 0.6379 | 93.6281 |
| 11 | 0.0305 | 0.5426 | 94.1707 |
| 12 | 0.0262 | 0.4662 | 94.6369 |
| 13 | 0.0229 | 0.407 | 95.044 |
| 14 | 0.0212 | 0.377 | 95.421 |
| ... | ... | ... | ... |
| 106 | 0 | 0 | 100 |
As shown in Table 1, when input variable is 13, contribution rate of accumulative total has reached 95%, can replace 106 original dimensions
Input variable, network architecture are simplified.
Step (3): the data after dimensionality reduction as the input for improving neural network model and are trained.As shown in Fig. 2,
The selection of neural network initial value is optimized using particle swarm algorithm, the specific steps are as follows:
(3.1) weight of neural network and threshold value as the initial position of particle and are initialized into particle group parameters, initially
Weightc1=c2=2, population M=50, greatest iteration step number 1000；
(3.2) using neural metwork training error as the fitness of particle swarm algorithm, individual optimal value and the overall situation are found most
The figure of merit；
(3.3) speed of more new particle and position:
xid k+1=xid k+vid k
(3.4) in the global optimum of specified iterative steps output particle；
(3.5) using optimal value as the initial weight of neural network algorithm and threshold value；
(3.6) training neural network model is until iterative steps terminate or set error less than initial.
Neural network uses three-decker, and input layer number is 13, and output layer number of nodes is 48, and node in hidden layer is
10, using the training method of LM, training output error is set to 0.01, and network has good generalization ability at this time.
Step (4): forecast sample is inputted in trained prediction model, and the load for exporting day to be predicted every 30min is pre-
Measured value.Since output valve has certain fluctuation, accidental error can be generated, needs to carry out output valve the amendment of following steps: with
Based on continuous 10 tests output valve, it is averaged；By output valve compared with average value, if the error of output valve and average value
The biggish value of the fluctuation is rejected greater than 10%；Remaining output valve is averaged again as final model output value,
The predicted load of day every 30min i.e. to be predicted.
Prediction error assessment index mainly has: average absolute percentage error (MAPE) and worst error value (ME).
Average absolute percentage error (MAPE) is defined as follows:
Worst error (ME) is defined as follows:
In formula Li andRespectively predict that the true value and predicted value of day every 30min load, n are prediction time number, n=
48.Prediction result is as shown in table 1.
To analyze prediction effect of the invention, by the method for the present invention (KPCA-PSOBP) and BP neural network method (BP),
PSO Neural Network method (PSOBP), the prediction knot for being based on the improved PSO Neural Network method (PCA-PSOBP) of PCA
Fruit is compared, while comparing the accuracy of output valve amendment front and back, is carried out using MAPE index and ME index to prediction error
Evaluation, short-term load forecasting effect as shown in figure 3, prediction result and index analysis result as shown in table 2, table 3.
2 prediction result of table and error analysis
3 four kinds of method prediction effects of table compare
In conjunction with table 2, the result of table 3 and Fig. 3 are it is known that the result of the method for the present invention prediction can reflect reality substantially
The case where load, illustrates the feasibility of this method；Compared with BP method, PSO-BP method, PCA-PSOBP method, the present invention
MAPE and the ME error criterion of method are smaller, and runing time is shorter, embody the validity of method.PSO is excellent compared with BP method
Change BP and improves Searching efficiency and accuracy；There is input quantity dimension-reduction treatment compared with no input quantity dimension-reduction treatment, dimension-reduction treatment is gone
In addition to the amount of redundancy in input, simplified model while, improves the efficiency of prediction；Compared with the PCA-PSOBP of linear dimensionality reduction,
KPCA-PSOBP overcomes the deficiency of linear dimensionality reduction, and forecasting efficiency and precision have obtained further raising.It can be obtained by table 3
Know, the MAPE of prediction result obtained by output valve error correcting method is approximately less than the MAPE error of most single tests, table
Validity of the error correcting method in terms of overcoming accidental error in the present invention is illustrated.In conclusion method of the invention can be with
The preferable short-term forecast for realizing electric load, engineering practical value with higher.
Claims (5)
1. a kind of short-term load forecasting method based on KPCA in conjunction with improvement neural network, which is characterized in that including following several
A step:
(1) it analyzes, choose the factor and historical data for influencing load, it is preliminary to constitute neural network sample collection；
(2) dimensionality reduction decoupling is carried out to input neural network sample collection using core principle component analysis algorithm；
(3) the neural network sample collection after decoupling dimensionality reduction as the input quantity for improving neural network model and is trained
To prediction model；
It (4) will be in forecast sample input step (3) trained prediction model；
(5) output valve of prediction model is modified, as short-term load forecasting value.
2. the short-term load forecasting method according to claim 1 based on KPCA in conjunction with improvement neural network, feature
Be: neural network sample collection described in step (1) includes historical data, temperature, week type and date type.
3. the short-term load forecasting method according to claim 1 based on KPCA in conjunction with improvement neural network, feature
Be: core principle component analysis algorithm described in step (2) includes the following steps:
A) n sample, each sample are contained to the sample set X ∈ R of d variable compositiond×nIt is standardized；
B) selection gaussian kernel function carries out inner product calculating, obtains matrix K；
C) centralization nuclear matrixIt is wherein the matrix that A is N × N, each element is 1/
N；
D) to matrixIt is PCA, finds out eigen vector；
E) contribution rate and contribution rate of accumulative total of principal component are found out, the dimension after determining dimensionality reduction extracts principal component amount.
4. the short-term load forecasting method according to claim 1 based on KPCA in conjunction with improvement neural network, feature
Be: step (3) the improvement neural network model carries out the selection of neural network initial value using particle swarm algorithm excellent
Change, specific as follows:
A) weight of neural network and threshold value as the initial position of particle and are initialized into particle group parameters；
B) using neural metwork training error as the fitness of particle swarm algorithm；
C) individual optimal value and global optimum are found；
D) speed of more new particle and position；
E) global optimum of particle is exported in specified iterative steps or error；
F) using optimal value as the initial weight of neural network algorithm and threshold value；
G) training neural network model is until iterative steps terminate or set error less than initial.
5. the short-term load forecasting method according to claim 1 based on KPCA in conjunction with improvement neural network, feature
Be: step is modified the output valve of prediction model described in (5), comprising the following steps: with prediction model continuous 10 times
Output valve based on, be averaged；By 10 output valves compared with average value, if output valve and the error of average value are big
The biggish value of the fluctuation is rejected in 10%；As final model output value, i.e., remaining output valve is averaged again
Short-term load forecasting value.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201710780654.4A CN109426889A (en) | 2017-09-01 | 2017-09-01 | Short-term load forecasting method based on KPCA in conjunction with improvement neural network |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201710780654.4A CN109426889A (en) | 2017-09-01 | 2017-09-01 | Short-term load forecasting method based on KPCA in conjunction with improvement neural network |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN109426889A true CN109426889A (en) | 2019-03-05 |
Family
ID=65513003
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201710780654.4A Pending CN109426889A (en) | 2017-09-01 | 2017-09-01 | Short-term load forecasting method based on KPCA in conjunction with improvement neural network |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN109426889A (en) |
Cited By (17)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN110009140A (en) * | 2019-03-20 | 2019-07-12 | 华中科技大学 | A kind of day Methods of electric load forecasting and prediction meanss |
| CN110110930A (en) * | 2019-05-08 | 2019-08-09 | 西南交通大学 | A kind of Recognition with Recurrent Neural Network Short-Term Load Forecasting Method improving whale algorithm |
| CN110275441A (en) * | 2019-07-02 | 2019-09-24 | 武汉科技大学 | A kind of quick self-adapted decoupling control method of PSORBFD |
| CN110348630A (en) * | 2019-07-09 | 2019-10-18 | 武汉四创自动控制技术有限责任公司 | A kind of isolated island region Methods of electric load forecasting and system |
| CN110414788A (en) * | 2019-06-25 | 2019-11-05 | 国网上海市电力公司 | A kind of power quality prediction technique based on similar day and improvement LSTM |
| CN110443414A (en) * | 2019-07-24 | 2019-11-12 | 东北电力大学 | A kind of ultra-short term Methods of electric load forecasting |
| CN110533484A (en) * | 2019-09-05 | 2019-12-03 | 四川长虹电器股份有限公司 | A kind of product Method for Sales Forecast method based on PCA and improved BP |
| CN110533087A (en) * | 2019-08-16 | 2019-12-03 | 成都电科慧安科技有限公司 | A kind of classification method based on velocity vector crowd cluster |
| CN110889564A (en) * | 2019-12-16 | 2020-03-17 | 吉林大学 | Short-term power load prediction method based on MKPCA-RBFNN |
| CN111461463A (en) * | 2020-04-30 | 2020-07-28 | 南京工程学院 | Short-term load prediction method, system and equipment based on TCN-BP |
| CN112348236A (en) * | 2020-10-23 | 2021-02-09 | 浙江八达电子仪表有限公司 | Abnormal daily load demand prediction system and method for intelligent power consumption monitoring terminal |
| CN112504682A (en) * | 2020-12-21 | 2021-03-16 | 中国人民解放军63963部队 | Chassis engine fault diagnosis method and system based on particle swarm optimization algorithm |
| CN112669169A (en) * | 2020-12-15 | 2021-04-16 | 国网辽宁省电力有限公司阜新供电公司 | Short-term photovoltaic power prediction device and method |
| CN113065715A (en) * | 2021-04-21 | 2021-07-02 | 东南大学 | Multi-load ultra-short-term prediction method for comprehensive energy system |
| CN114372691A (en) * | 2021-12-29 | 2022-04-19 | 国网天津市电力公司 | Electric energy substitution potential estimation method based on holographic perception |
| US11539217B2 (en) | 2019-10-28 | 2022-12-27 | Enphase Energy, Inc. | Method and apparatus for tertiary control of microgrids with integrated over-current protection |
| CN115967077A (en) * | 2022-08-30 | 2023-04-14 | 湖南农业大学 | Short-term prediction power load prediction method and system |
Citations (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN102930347A (en) * | 2012-10-15 | 2013-02-13 | 河海大学 | Method for forecasting short term load under demand response |
| CN105913175A (en) * | 2016-04-07 | 2016-08-31 | 哈尔滨理工大学 | Intelligent power grid short period load prediction method based on improved nerve network algorithm |
-
2017
- 2017-09-01 CN CN201710780654.4A patent/CN109426889A/en active Pending
Patent Citations (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN102930347A (en) * | 2012-10-15 | 2013-02-13 | 河海大学 | Method for forecasting short term load under demand response |
| CN105913175A (en) * | 2016-04-07 | 2016-08-31 | 哈尔滨理工大学 | Intelligent power grid short period load prediction method based on improved nerve network algorithm |
Non-Patent Citations (2)
| Title |
|---|
| 刘畅 等: ""基于KPCA和BP神经网络的短期负荷预测"", 《电测与仪表》 * |
| 周涛 等: ""基于改进神经网络的电力系统中长期负荷预测研究"", 《电气应用》 * |
Cited By (26)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN110009140A (en) * | 2019-03-20 | 2019-07-12 | 华中科技大学 | A kind of day Methods of electric load forecasting and prediction meanss |
| CN110009140B (en) * | 2019-03-20 | 2021-10-08 | 华中科技大学 | Daily power load prediction method and prediction device |
| CN110110930A (en) * | 2019-05-08 | 2019-08-09 | 西南交通大学 | A kind of Recognition with Recurrent Neural Network Short-Term Load Forecasting Method improving whale algorithm |
| CN110110930B (en) * | 2019-05-08 | 2022-03-25 | 西南交通大学 | Recurrent neural network short-term power load prediction method for improving whale algorithm |
| CN110414788A (en) * | 2019-06-25 | 2019-11-05 | 国网上海市电力公司 | A kind of power quality prediction technique based on similar day and improvement LSTM |
| CN110414788B (en) * | 2019-06-25 | 2023-12-08 | 国网上海市电力公司 | Electric energy quality prediction method based on similar days and improved LSTM |
| CN110275441A (en) * | 2019-07-02 | 2019-09-24 | 武汉科技大学 | A kind of quick self-adapted decoupling control method of PSORBFD |
| CN110275441B (en) * | 2019-07-02 | 2022-04-12 | 武汉科技大学 | PSORBFD (particle swarm optimization-based adaptive feedback) rapid self-adaptive decoupling control method |
| CN110348630A (en) * | 2019-07-09 | 2019-10-18 | 武汉四创自动控制技术有限责任公司 | A kind of isolated island region Methods of electric load forecasting and system |
| CN110443414B (en) * | 2019-07-24 | 2022-03-18 | 东北电力大学 | Ultra-short-term power load prediction method |
| CN110443414A (en) * | 2019-07-24 | 2019-11-12 | 东北电力大学 | A kind of ultra-short term Methods of electric load forecasting |
| CN110533087A (en) * | 2019-08-16 | 2019-12-03 | 成都电科慧安科技有限公司 | A kind of classification method based on velocity vector crowd cluster |
| CN110533087B (en) * | 2019-08-16 | 2023-04-07 | 成都电科慧安科技有限公司 | Classification method based on speed vector crowd clustering |
| CN110533484A (en) * | 2019-09-05 | 2019-12-03 | 四川长虹电器股份有限公司 | A kind of product Method for Sales Forecast method based on PCA and improved BP |
| US11539217B2 (en) | 2019-10-28 | 2022-12-27 | Enphase Energy, Inc. | Method and apparatus for tertiary control of microgrids with integrated over-current protection |
| CN110889564A (en) * | 2019-12-16 | 2020-03-17 | 吉林大学 | Short-term power load prediction method based on MKPCA-RBFNN |
| CN111461463A (en) * | 2020-04-30 | 2020-07-28 | 南京工程学院 | Short-term load prediction method, system and equipment based on TCN-BP |
| CN111461463B (en) * | 2020-04-30 | 2023-11-24 | 南京工程学院 | Short-term load prediction method, system and equipment based on TCN-BP |
| CN112348236A (en) * | 2020-10-23 | 2021-02-09 | 浙江八达电子仪表有限公司 | Abnormal daily load demand prediction system and method for intelligent power consumption monitoring terminal |
| CN112348236B (en) * | 2020-10-23 | 2023-12-26 | 浙江八达电子仪表有限公司 | Abnormal daily load demand prediction system and method for intelligent electricity consumption monitoring terminal |
| CN112669169A (en) * | 2020-12-15 | 2021-04-16 | 国网辽宁省电力有限公司阜新供电公司 | Short-term photovoltaic power prediction device and method |
| CN112669169B (en) * | 2020-12-15 | 2024-04-30 | 国网辽宁省电力有限公司阜新供电公司 | Short-term photovoltaic power prediction device and method |
| CN112504682A (en) * | 2020-12-21 | 2021-03-16 | 中国人民解放军63963部队 | Chassis engine fault diagnosis method and system based on particle swarm optimization algorithm |
| CN113065715A (en) * | 2021-04-21 | 2021-07-02 | 东南大学 | Multi-load ultra-short-term prediction method for comprehensive energy system |
| CN114372691A (en) * | 2021-12-29 | 2022-04-19 | 国网天津市电力公司 | Electric energy substitution potential estimation method based on holographic perception |
| CN115967077A (en) * | 2022-08-30 | 2023-04-14 | 湖南农业大学 | Short-term prediction power load prediction method and system |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN109426889A (en) | Short-term load forecasting method based on KPCA in conjunction with improvement neural network | |
| Wang et al. | A seasonal GM (1, 1) model for forecasting the electricity consumption of the primary economic sectors | |
| Li et al. | Prediction for tourism flow based on LSTM neural network | |
| Zhang et al. | Short-term rainfall forecasting using multi-layer perceptron | |
| CN109063911B (en) | Load aggregation grouping prediction method based on gated cycle unit network | |
| CN109359786A (en) | A kind of power station area short-term load forecasting method | |
| CN109978284B (en) | Photovoltaic power generation power time-sharing prediction method based on hybrid neural network model | |
| CN108876001B (en) | Short-term power load prediction method based on twin support vector machine | |
| CN104636801A (en) | Transmission line audible noise prediction method based on BP neural network optimization | |
| CN105447509A (en) | Short-term power prediction method for photovoltaic power generation system | |
| CN113762387B (en) | Multi-element load prediction method for data center station based on hybrid model prediction | |
| CN104820877A (en) | Photovoltaic system generation power prediction method based on cloud adaptive PSO-SNN | |
| Rizwan et al. | Artificial intelligence based approach for short term load forecasting for selected feeders at madina saudi arabia | |
| CN112149890A (en) | Comprehensive energy load prediction method and system based on user energy label | |
| CN116526473A (en) | Particle swarm optimization LSTM-based electrothermal load prediction method | |
| Ghofrani et al. | Hybrid clustering-time series-bayesian neural network short-term load forecasting method | |
| CN115099461A (en) | Solar radiation prediction method and system based on double-branch feature extraction | |
| Li et al. | An Integrated Artificial Neural Network-based Precipitation Revision Model. | |
| CN112507613B (en) | Second-level ultra-short-term photovoltaic power prediction method | |
| CN109919374A (en) | Prediction of Stock Price method based on APSO-BP neural network | |
| CN113239503A (en) | New energy output scene analysis method and system based on improved k-means clustering algorithm | |
| CN117132132A (en) | Photovoltaic power generation power prediction method based on meteorological data | |
| CN115481788A (en) | Load prediction method and system for phase change energy storage system | |
| de Castro et al. | Seasonal rainfall forecast using a neo-fuzzy neuron model | |
| CN111798275B (en) | Domestic flight price prediction method |
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 | ||
| RJ01 | Rejection of invention patent application after publication |
Application publication date: 20190305 |
|
| RJ01 | Rejection of invention patent application after publication |