CN103514488A - Electrical power system short-term load forecasting device and method based on combination forecasting model - Google Patents
Electrical power system short-term load forecasting device and method based on combination forecasting model Download PDFInfo
- Publication number
- CN103514488A CN103514488A CN201310461805.1A CN201310461805A CN103514488A CN 103514488 A CN103514488 A CN 103514488A CN 201310461805 A CN201310461805 A CN 201310461805A CN 103514488 A CN103514488 A CN 103514488A
- Authority
- CN
- China
- Prior art keywords
- module
- forecasting
- load
- prediction
- model
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Granted
Links
Images
Landscapes
- Supply And Distribution Of Alternating Current (AREA)
- Management, Administration, Business Operations System, And Electronic Commerce (AREA)
Abstract
The invention provides an electrical power system short-term load forecasting device and method based on a combination forecasting model. The device comprises a data collecting module, an input module, a single model forecasting module, a weight coefficient determining module, a combination forecasting module, a forecasting result assessment module, a forecasting effect simulated analysis module and an output module. When a single forecasting model is selected, an AR model based on a Kalman filtering algorithm and the Kalman filtering algorithm are provided. When a combination model weight coefficient is computed, an improved weight-coefficient-variable is provided. The accuracy and the stability of load forecasting are further improved, the modular structure of the device is beneficial to device upgrading and maintaining, and decision support is provided for economic, safe and reliable operation of an electrical power system.
Description
Technical field
What the present invention relates to is a kind of power-system short-term load forecasting device, the present invention also relates to a kind of power-system short-term load forecasting method.
Background technology
In Operation of Electric Systems, control and planning management, load prediction has determined the reasonable arrangement of generating, transmission and disttrbution, is the important component part of Power System Planning.Wherein, the topmost application of short-term load forecasting is to provide data for generation schedule program, be used for determining the operating scheme that meets safety requirements, operation constraint and physical environment and device-restrictive, the security of operation of power networks, reliability and economy are played an important role.How to improve precision of prediction and be and study at present theoretical center and the emphasis with method of short-term load forecasting, short-term load forecasting has become and has realized one of modern important content of power system management accurately.
For a long time, Chinese scholars has been carried out extensive and deep research to Load Prediction In Power Systems theory, has proposed many effective methods, as regression analysis, and time series method, neural network, wavelet analysis method etc.For a certain forecasting problem, can set up multiple Forecasting Methodology.Different Forecasting Methodologies provides different information of forecastings and different precision of predictions.If simply the larger method of predicated error is given up to fall, tends to lose some useful informations.The method of science is, different Forecasting Methodologies is carried out to suitable combination, and the information being provided to fully utilize the whole bag of tricks, improves precision of prediction as far as possible.In order to make full use of the useful information of the single model of each load prediction, to improve the degree of accuracy of load prediction, several different Forecasting Methodologies are combined and become combination forecasting.Linear Regression Model in One Unknown and Grey System Model are combined in the propositions such as Zhang Yan, the least absolute value method that the absolute value sum minimum of take is objective function is determined weight coefficient (Zhang Yan, Ma Chuansheng, Wei can. the application of definite research-least absolute value method of weight in combined prediction. and traffic and transportation system engineering and information .2006,6 (4): 125-129).Wei Youhua select on the Individual forecast basis of conventional exponential smoothing, trend analysis method, time series method and grey method, set up combination forecasting and adopt least square method determine combination forecasting weight (Wei Youhua. the combination forecasting method of short-term power prediction. Geophysical Ano Geochemical Exploration computing technique .2005,27 (2): 178-180).The persons such as Yogesh Bichpuriya combine with expert system approach, artificial neural network and three kinds of single methods of ARMA time series, and adopt variance and covariance method to determine that weight coefficient forecasts (Yogesh Bichpuriya to power-system short-term load, M.S.S.Rao, S.A.Soman.Combination approaches for short term load forecasting.2010 9
thinternational Power and Energy Conference, IPEC 2010:818-823).But said method all has much room for improvement on precision and stability, the value of forecasting is undesirable.
Summary of the invention
The object of the present invention is to provide the load prediction device that a kind of precision is high, stability is strong.The present invention also aims to provide a kind of power-system short-term load forecasting method based on combination forecasting.
Power-system short-term load forecasting device based on combination forecasting of the present invention comprises data acquisition module, load module, single model prediction module, determines weight coefficient module, combined prediction module, the evaluation module that predicts the outcome, prediction effect simulation analysis module and output module;
Described data acquisition module is for to electrical network integral point 24 hours every days load, (unit is: MW/h) gather; Described load module can be for input historical load data and data processing; Described single model prediction module is for calculating the predicted value of three kinds of Individual forecast models; Described definite weight coefficient module is for the weight coefficient of calculation combination prediction; Described combined prediction module is used for calculating finally and predicts the outcome; The described evaluation module that predicts the outcome is for checking the various error criterions of precision of prediction; The simulation analysis module of described prediction effect is for carrying out simulation test and analysis to load prediction effect; Described output module is for showing and the predicting the outcome of output load;
Signal flow between each module is: data acquisition module, load module, single model prediction module, the simulation analysis module of determining weight coefficient module, combined prediction module, prediction effect, the evaluation module that predicts the outcome, output module are connected in series successively.Each module must be carried out in order, and the output of a upper module is the input of next module, and the historical load of electrical network, by after data acquisition module, is carried out data processing through load module; Data available after processing enters Individual forecast module, and the Output rusults of three kinds of Individual forecast models is for determining the input of weight coefficient module, if i kind method is at the prediction error e of t period
it>(α is given threshold value to α, this method is selected α=3%), represent that the predicted value sudden change of section at this moment of this kind of method is very large, can give up, getting its weight is zero, and remaining various single forecasting models are from the equal weight of new distribution, utilize simulation analysis module and evaluation of result module to predicting the outcome, to carry out accuracy requirement check, if met the requirements, enter output module, if undesirable, return data acquisition module, re-starts the data acquisition of other periods.
Power-system short-term load forecasting method based on combination forecasting of the present invention is:
When to the setting up of single model, kalman filter method, chaos Kalman filter method have been proposed, to built-up pattern weight coefficient really regularly, take can variable weight method, following methods and embodiment have been proposed:
(1) load data of electrical network gathered and suitably process, forming available Load Time Series;
(2) three kinds of single forecasting model CALCULATING PREDICTION result f that adopt respectively least square method, Kalman filter method, chaos Kalman filter method to set up
1t, f
2t, f
3t;
(3) ask for the weight coefficient w of combining prediction model
1, w
2, w
3;
(4) obtain final forecasting model f
t=w
1f
1t+ w
2f
2t+ w
3f
3t, t=1,2 ..., n;
(5) model that utilizes (4) to set up carries out Load Forecasting.
Power-system short-term load forecasting device based on combination forecasting of the present invention has adopted modular construction, and modular construction is beneficial to device upgrading and safeguards; Can carry out online acquisition data, line modeling, online forecasting, be a kind of real-time online predictor; Compare and increased prediction effect simulation analysis module, the evaluation module that predicts the outcome with device in the past, make application person grasp in real time predicated error, make correct judgement and decision-making.
Power-system short-term load forecasting method based on combination forecasting of the present invention, with respect to traditional load forecasting method and device advantage, be, while choosing Individual forecast model, AR model based on Kalman filtering algorithm and the forecasting model of chaos Kalman filtering algorithm have been proposed, the vector using forecasting model parameter as Kalman filter status, application Kalman filtering algorithm estimated prediction model parameter, sets up forecasting model; When calculation combination Model Weight coefficient, a kind of method of improved variable weight coefficient is proposed, to compare with the method for not variable weight in the past, load prediction precision is higher.Proposition of the present invention has further improved precision and the stability of load prediction, for electric system economy, safety and reliability service provide decision support.
Accompanying drawing explanation
The module composition frame chart of the power-system short-term load forecasting device of Fig. 1 based on combination forecasting;
The power-system short-term load forecasting method process flow diagram of Fig. 2 based on combination forecasting;
Embodiment
Below in conjunction with accompanying drawing, the present invention is described in more detail.
In conjunction with Fig. 1, specific embodiment of the invention scheme is described in detail.
Power-system short-term load forecasting device based on combination forecasting of the present invention comprises data acquisition module, load module, single model prediction module, determines weight coefficient module, combined prediction module, the evaluation module that predicts the outcome, prediction effect simulation analysis module and output module;
Described data acquisition module is for to electrical network integral point 24 hours every days load, (unit is: MW/h) gather; Described load module can be for input historical load data and data processing; Described single model prediction module, for calculating the predicted value of three kinds of Individual forecast models, comprises least square model module, Kalman filter forecasting module and chaos Kalman filter forecasting module; Described definite weight coefficient module is for the weight coefficient of calculation combination prediction; Described combined prediction module is used for calculating finally and predicts the outcome; The described evaluation module that predicts the outcome is for checking the various error criterions of precision of prediction; The simulation analysis module of described prediction effect is for carrying out simulation test and analysis to load prediction effect; Described output module is for showing and the predicting the outcome of output load;
Signal flow between each module is: data acquisition module, load module, single model prediction module, the simulation analysis module of determining weight coefficient module, combined prediction module, prediction effect, the evaluation module that predicts the outcome, output module are connected in series successively.Each module must be carried out in order, and the output of a upper module is the input of next module, and the historical load of electrical network, by after data acquisition module, is carried out data processing through load module; Data available after processing enters Individual forecast module, and the Output rusults of three kinds of Individual forecast models is for determining the input of weight coefficient module, if i kind method is at the prediction error e of t period
it>(α is given threshold value to α, this method is selected α=3%), represent that the predicted value sudden change of section at this moment of this kind of method is very large, can give up, getting its weight is zero, and remaining various single forecasting models are from the equal weight of new distribution, utilize simulation analysis module and evaluation of result module to predicting the outcome, to carry out accuracy requirement check, if met the requirements, enter output module, if undesirable, return data acquisition module, re-starts the data acquisition of other periods.
In conjunction with Fig. 2, the specific embodiment of the present invention is described in further detail.
Power-system short-term load forecasting method based on combination forecasting of the present invention is:
New, more effective Forecasting Methodology have been proposed.When to the setting up of single model, kalman filter method has been proposed, chaos Kalman filter method, to built-up pattern weight coefficient really regularly, take can variable weight method, following methods and embodiment have been proposed:
(1) load data of electrical network gathered and suitably process, forming available Load Time Series;
(2) three kinds of single forecasting model CALCULATING PREDICTION result f that adopt respectively least square method, Kalman filter method, chaos Kalman filter method to set up
1t, f
2t, f
3t;
Least square method has list of references introduction, repeat no more herein, below in detail introduce and utilize Kalman filter method to carry out short-term Load Prediction In Power Systems, the method utilizes Kalman filtering recursion formula to set up self-adaptation AR model, the vector using AR model auto-regressive parameter as Kalman filter status, the derivation of Kalman filtering method repeats no more herein, and the variable of the variable in Kalman filtering equations and AR model equation is contrasted, as shown in table 1:
Table 1 is the variable contrast of the corresponding equation of situation steadily
Available (0.001～0.01) * (variance of y (k)) is as J
min, empirical tests can be seen J
minvery little on arithmetic result impact, utilize contrast relationship, can derive the Kalman filtering algorithm of load forecasting model in steady situation:
X (1) wherein, X (2) ..., X (k) is input vector, y (1), and y (2) ..., y (k), for wishing output, works as k=1,2,3
K(k)＝P(k-1)X(k)·[X
T(k)P(k-1)X(k)+J
min]
-1
P(k)=P(k-1)-K(k)X
T(k)P(k-1)
Suppose J
min=(0.001～0.01) * (variance of y (k)), starting condition
p (0)=cI (c>0);
(3) ask for the weight coefficient w of combining prediction model
1, w
2, w
3
According to selected k kind Individual forecast model, the predicted value f to every kind of model in the t period
it(i=1,2,3; T=1,2 ..., n), give equal weight
if i kind method is at the prediction error e of t period
it>α (α is given threshold value, selects α=3% herein), representing that the predicted value sudden change of section at this moment of this kind of method is very large, can give up, getting its weight is zero, remaining various single forecasting models are from the equal weight of new distribution
(4) obtain final forecasting model f
t=w
1f
1t+ w
2f
2t+ w
3f
3t, t=1,2 ..., n;
(5) model that utilizes (4) to set up carries out Load Forecasting.
Claims (2)
1. the power-system short-term load forecasting device based on combination forecasting, comprise data acquisition module, load module, single model prediction module, determine weight coefficient module, combined prediction module, the evaluation module that predicts the outcome, prediction effect simulation analysis module and output module, it is characterized in that:
Described data acquisition module gathers electrical network integral point 24 hours every days load; Described load module input historical load data and data processing; Described single model prediction module is calculated the predicted value of three kinds of Individual forecast models; The weight coefficient of described definite weight coefficient module calculation combination prediction; Described combined prediction module is calculated and is finally predicted the outcome; The various error criterions of the described evaluation module check precision of prediction that predicts the outcome; The simulation analysis module of described prediction effect is carried out simulation test and analysis to load prediction effect; Predicting the outcome of described output module demonstration and output load;
Signal flow between each module is: data acquisition module, load module, single model prediction module, the simulation analysis module of determining weight coefficient module, combined prediction module, prediction effect, the evaluation module that predicts the outcome, output module are connected in series successively; The output of a upper module is the input of next module, and the historical load of electrical network, by after data acquisition module, is carried out data processing through load module; Data available after processing enters Individual forecast module, and the Output rusults of three kinds of Individual forecast models is for determining the input of weight coefficient module, if i kind method is at the prediction error e of t period
it>α, α is given threshold value, α=3%, represents that the predicted value sudden change of section at this moment of this kind of method is very large, gives up, getting its weight is zero, remaining various single forecasting model is redistributed equal weight, utilizes simulation analysis module and evaluation of result module to carry out accuracy requirement check to predicting the outcome, if met the requirements, enters output module, if undesirable, return data acquisition module, re-starts the data acquisition of other periods.
2. the Forecasting Methodology based on the power-system short-term load forecasting device based on combination forecasting claimed in claim 1, is characterized in that:
(1) load data of electrical network gathered and suitably process, forming available Load Time Series;
(2) calculate respectively the forecast result f of least square method, Kalman filter method, three kinds of single models of chaos Kalman filter method
1t, f
2t, f
3t;
(3) ask for the weight coefficient w of combining prediction model
1, w
2, w
3;
(4) obtain final forecasting model f
t=w
1f
1t+ w
2f
2t+ w
3f
3t, t=1,2 ..., n;
(5) model that utilizes (4) to set up carries out Load Forecasting.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201310461805.1A CN103514488B (en) | 2013-10-08 | 2013-10-08 | Power-system short-term load forecasting device and method based on combination forecasting |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201310461805.1A CN103514488B (en) | 2013-10-08 | 2013-10-08 | Power-system short-term load forecasting device and method based on combination forecasting |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN103514488A true CN103514488A (en) | 2014-01-15 |
| CN103514488B CN103514488B (en) | 2016-06-29 |
Family
ID=49897177
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201310461805.1A Expired - Fee Related CN103514488B (en) | 2013-10-08 | 2013-10-08 | Power-system short-term load forecasting device and method based on combination forecasting |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN103514488B (en) |
Cited By (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN103870887A (en) * | 2014-03-07 | 2014-06-18 | 国家电网公司 | Improved adaptive filtering method-based urban power grid short-term load forecasting method |
| CN104281921A (en) * | 2014-10-10 | 2015-01-14 | 讯腾数码科技（北京）有限公司 | Method for obtaining dynamic risk evaluation data of city underground pipe network |
| CN105205560A (en) * | 2015-09-25 | 2015-12-30 | 国网甘肃省电力公司电力科学研究院 | Photovoltaic power supply power prediction method based on positive and negative error variable weights |
| CN108596407A (en) * | 2018-05-15 | 2018-09-28 | 浙江工业大学 | A kind of daily synthetic load Forecasting Approach for Short-term of micro-capacitance sensor based on manhatton distance and Unscented kalman filtering |
| CN110929941A (en) * | 2019-11-26 | 2020-03-27 | 广东电网有限责任公司 | Short-term power load prediction method and system based on multi-load mode |
| CN116315189A (en) * | 2023-05-25 | 2023-06-23 | 澄瑞电力科技(上海)股份公司 | Data fusion-based battery Bao Re out-of-control prediction method and system |
Citations (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN101609517A (en) * | 2008-06-20 | 2009-12-23 | 上海申瑞电力科技股份有限公司 | Power-system short-term load forecasting method based on intelligent strategy management |
| CN102102626A (en) * | 2011-01-30 | 2011-06-22 | 华北电力大学 | Method for forecasting short-term power in wind power station |
| CN102750542A (en) * | 2012-05-29 | 2012-10-24 | 北京交通大学 | Support vector regression machine wind speed combination forecast method with interpolation being smoothed and optimized |
| CN103218675A (en) * | 2013-05-06 | 2013-07-24 | 国家电网公司 | Short-term load prediction method based on clustering and sliding window |
-
2013
- 2013-10-08 CN CN201310461805.1A patent/CN103514488B/en not_active Expired - Fee Related
Patent Citations (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN101609517A (en) * | 2008-06-20 | 2009-12-23 | 上海申瑞电力科技股份有限公司 | Power-system short-term load forecasting method based on intelligent strategy management |
| CN102102626A (en) * | 2011-01-30 | 2011-06-22 | 华北电力大学 | Method for forecasting short-term power in wind power station |
| CN102750542A (en) * | 2012-05-29 | 2012-10-24 | 北京交通大学 | Support vector regression machine wind speed combination forecast method with interpolation being smoothed and optimized |
| CN103218675A (en) * | 2013-05-06 | 2013-07-24 | 国家电网公司 | Short-term load prediction method based on clustering and sliding window |
Cited By (8)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN103870887A (en) * | 2014-03-07 | 2014-06-18 | 国家电网公司 | Improved adaptive filtering method-based urban power grid short-term load forecasting method |
| CN104281921A (en) * | 2014-10-10 | 2015-01-14 | 讯腾数码科技（北京）有限公司 | Method for obtaining dynamic risk evaluation data of city underground pipe network |
| CN105205560A (en) * | 2015-09-25 | 2015-12-30 | 国网甘肃省电力公司电力科学研究院 | Photovoltaic power supply power prediction method based on positive and negative error variable weights |
| CN108596407A (en) * | 2018-05-15 | 2018-09-28 | 浙江工业大学 | A kind of daily synthetic load Forecasting Approach for Short-term of micro-capacitance sensor based on manhatton distance and Unscented kalman filtering |
| CN108596407B (en) * | 2018-05-15 | 2021-08-03 | 浙江工业大学 | Microgrid daily comprehensive load short-time prediction method based on Manhattan distance and unscented Kalman filtering |
| CN110929941A (en) * | 2019-11-26 | 2020-03-27 | 广东电网有限责任公司 | Short-term power load prediction method and system based on multi-load mode |
| CN116315189A (en) * | 2023-05-25 | 2023-06-23 | 澄瑞电力科技(上海)股份公司 | Data fusion-based battery Bao Re out-of-control prediction method and system |
| CN116315189B (en) * | 2023-05-25 | 2023-08-11 | 澄瑞电力科技(上海)股份公司 | Data fusion-based battery Bao Re out-of-control prediction method and system |
Also Published As
| Publication number | Publication date |
|---|---|
| CN103514488B (en) | 2016-06-29 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN110311376B (en) | Dynamic safety assessment comprehensive model and space-time visualization method for power system | |
| US20200327411A1 (en) | Systems and Method on Deriving Real-time Coordinated Voltage Control Strategies Using Deep Reinforcement Learning | |
| Wan et al. | Optimal prediction intervals of wind power generation | |
| CN103514488A (en) | Electrical power system short-term load forecasting device and method based on combination forecasting model | |
| CN103268519B (en) | Based on the power-system short-term load forecasting method and the device that improve Lyapunov index | |
| CN108584592A (en) | A kind of shock of elevator car abnormity early warning method based on time series predicting model | |
| Fioriti et al. | A novel stochastic method to dispatch microgrids using Monte Carlo scenarios | |
| CN105300692B (en) | A kind of bearing failure diagnosis and Forecasting Methodology based on expanded Kalman filtration algorithm | |
| CN105787606A (en) | Power dispatching online trend early warning system based on ultra short term load prediction | |
| CN107124394A (en) | A kind of powerline network security postures Forecasting Methodology and system | |
| CN105447332B (en) | A kind of reliability index distribution method of Large-Scale Equipment | |
| CN104601604A (en) | Network security situation analyzing method | |
| CN105488335A (en) | Lyapunov exponent based power system load prediction method and apparatus | |
| Lv et al. | Novel deterministic and probabilistic combined system based on deep learning and self-improved optimization algorithm for wind speed forecasting | |
| CN107527121A (en) | A kind of method of the information system running status diagnosis prediction of power network | |
| CN110472851A (en) | A kind of power distribution network risk hidden danger dynamic evaluation model building method neural network based | |
| CN111785093A (en) | Air traffic flow short-term prediction method based on fractal interpolation | |
| CN109961160B (en) | Power grid future operation trend estimation method and system based on tide parameters | |
| CN106772354A (en) | Method for tracking target and device based on parallel fuzzy gaussian sum particle filter | |
| CN112418662A (en) | Power distribution network operation reliability analysis method using artificial neural network | |
| Rivero et al. | Short-term rainfall time series prediction with incomplete data | |
| CN104298859A (en) | Multi-interval-prediction-result weighting integration method | |
| Larsen et al. | Evaluation of scenario reduction methods for stochastic inflow in hydro scheduling models | |
| CN103488891A (en) | Large grid safe stable prevention and control performance penalty index calculation method based on risks | |
| CN111310964B (en) | Load prediction method and device |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| C06 | Publication | ||
| PB01 | Publication | ||
| C10 | Entry into substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| C14 | Grant of patent or utility model | ||
| GR01 | Patent grant | ||
| CF01 | Termination of patent right due to non-payment of annual fee |
Granted publication date: 20160629 Termination date: 20211008 |
|
| CF01 | Termination of patent right due to non-payment of annual fee |