CN110619429A - Short-term load prediction method based on BP neural network - Google Patents
Short-term load prediction method based on BP neural network Download PDFInfo
- Publication number
- CN110619429A CN110619429A CN201910819380.4A CN201910819380A CN110619429A CN 110619429 A CN110619429 A CN 110619429A CN 201910819380 A CN201910819380 A CN 201910819380A CN 110619429 A CN110619429 A CN 110619429A
- Authority
- CN
- China
- Prior art keywords
- neural network
- load
- day
- layer
- output
- 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 48
- 238000000034 method Methods 0.000 title claims abstract description 16
- 230000006870 function Effects 0.000 claims description 21
- 210000002569 neuron Anatomy 0.000 claims description 9
- 238000003062 neural network model Methods 0.000 claims description 5
- 238000005457 optimization Methods 0.000 description 4
- 230000009286 beneficial effect Effects 0.000 description 2
- 238000012986 modification Methods 0.000 description 2
- 230000004048 modification Effects 0.000 description 2
- 238000010586 diagram Methods 0.000 description 1
- 238000011478 gradient descent method Methods 0.000 description 1
- 238000010248 power generation Methods 0.000 description 1
Classifications
-
- 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
- G06N3/08—Learning methods
- G06N3/084—Backpropagation, e.g. using gradient descent
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06Q—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
- G06Q10/00—Administration; Management
- G06Q10/04—Forecasting or optimisation specially adapted for administrative or management purposes, e.g. linear programming or "cutting stock problem"
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06Q—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
- G06Q50/00—Information and communication technology [ICT] specially adapted for implementation of business processes of specific business sectors, e.g. utilities or tourism
- G06Q50/06—Energy or water supply
Landscapes
- Engineering & Computer Science (AREA)
- Business, Economics & Management (AREA)
- Economics (AREA)
- Theoretical Computer Science (AREA)
- Physics & Mathematics (AREA)
- Health & Medical Sciences (AREA)
- Strategic Management (AREA)
- Human Resources & Organizations (AREA)
- General Physics & Mathematics (AREA)
- General Business, Economics & Management (AREA)
- Tourism & Hospitality (AREA)
- Marketing (AREA)
- General Health & Medical Sciences (AREA)
- Software Systems (AREA)
- Primary Health Care (AREA)
- Mathematical Physics (AREA)
- Computing Systems (AREA)
- Molecular Biology (AREA)
- Evolutionary Computation (AREA)
- Data Mining & Analysis (AREA)
- Public Health (AREA)
- Water Supply & Treatment (AREA)
- Computational Linguistics (AREA)
- Biophysics (AREA)
- General Engineering & Computer Science (AREA)
- Biomedical Technology (AREA)
- Artificial Intelligence (AREA)
- Life Sciences & Earth Sciences (AREA)
- Development Economics (AREA)
- Game Theory and Decision Science (AREA)
- Entrepreneurship & Innovation (AREA)
- Operations Research (AREA)
- Quality & Reliability (AREA)
- Supply And Distribution Of Alternating Current (AREA)
- Management, Administration, Business Operations System, And Electronic Commerce (AREA)
Abstract
The invention relates to the field of load prediction of an electric power market, in particular to a short-term load prediction method based on a BP (back propagation) neural network.
Description
Technical Field
The invention relates to the field of load prediction of power markets, in particular to a short-term load prediction method based on a BP neural network.
Background
With the continuous improvement of the scale and the complexity of the electric power system, whether the load prediction is accurate or not directly influences the normal operation of the electric power system, and has a key role in effectively reducing the power generation cost and implementing the optimal control of the electric power systems in various regions. Due to the characteristics of strong randomness, low stability, complex influence factors and the like of short-term load, the neural network with the nonlinear characteristic and the strong learning capacity can greatly improve the prediction accuracy, so that the neural network is widely applied to short-term load prediction of a power system.
The traditional neural network model only considers the influence of historical load and climate factors generally, but actually, a plurality of characteristic factors influencing the load exist, and the characteristic factors cannot be completely reflected in the neural network, so that the traditional neural network model cannot achieve high prediction accuracy.
Disclosure of Invention
The invention aims to solve the problem of low prediction precision of the traditional neural network in the prior art, provides a short-term load prediction method based on the BP neural network, can extract other characteristic factors except weather factors from a time sequence, and improves corresponding objective functions and optimization algorithms aiming at a model, thereby improving the precision of load prediction of the model
In order to solve the technical problems, the invention adopts the technical scheme that: the short-term load prediction method based on the BP neural network comprises the following steps:
the method comprises the following steps: establishing a traditional neural network model, adopting a gradient descent method as an optimization algorithm for training the model, wherein the formula of an objective function is as follows:
wherein, tphDenotes the actual load at p hour on day h, yphRepresenting the predicted load at the p hour of the h day, Eav (r)' representing the objective function of load prediction, and r representing the number of iterations;
step two: on the basis of the traditional neural network, a period of adjustable time zone is selected in a section of place where the same model is predicted, the deviation between a target value and a predicted value on a time zone is used as the correction of a new prediction to form a new target function, and the formula is as follows:
where Eav (r) represents the objective function of the load prediction, tphIndicates the th day of the h dayp hours actual load, yphDenotes the predicted load at p hour on day h, tjkDenotes the actual load at the j-th hour on the k-th day, yjkRepresents the predicted load at the j hour on the k day, and alpha and beta represent non-0 natural numbers, wherein alpha<Beta, r represents the iteration times of the neural network, H is the number of training samples, and P is the output space dimension;
step three: and (5) constructing a neural network load prediction model according to the objective function in the step two, and training the neural network load prediction model.
Preferably, the formula of the relative deviation of the predicted value and the target value of the conventional neural network is as follows:
where T is the actual load and Y is the predicted load.
Preferably, the neural network is composed of an input layer, a hidden layer and an output layer, and the output of the mth neuron of a single sample in the hidden layer is:
wherein, wimFor weights of input layer to hidden layer, xiFor input features and consisting of predicting the load per hour n days before the day and predicting the weather per hour on the day, bmIs the bias of the input layer to the hidden layer;
the output of a single sample at the p-th neuron in the output layer is:
wherein, wmpFor the weight from hidden layer to output layer, hmOutput of the mth neuron of the hidden layer, bpIs the bias of the hidden layer to the output layer.
Preferably, the output layer adds an average value of deviations between a target value and a predicted value of a day from α to β in the current day as a correction on the basis of a Y value output by a conventional neural network, and obtains an output characteristic formula as follows:
wherein Z isphOutput load at p hour of h day, tjkDenotes the actual load at the j-th hour on the k-th day, yjkRepresents the predicted load at the j-th hour on the k-th day, yphThe predicted load at the P-th hour on the h-th day is represented, and P is the output space dimension.
Preferably, the neural network load prediction model optimizes the weight and bias through the objective function, and the update formula of the weight and bias from the input layer to the hidden layer is as follows:
updating weight and bias from hidden layer to output layer:
where r denotes the number of iterations of the neural network, wimAs weights of the input layer to the hidden layer, bmIs the bias of the input layer to the hidden layer; w is ampWeights for the hidden layer to the output layer, bpIs the bias of the hidden layer to the output layer; ε is the learning rate, which ranges from (0, 1).
Preferably, the neural network training is performed by substituting the short-term load data of the PJM (Pennsylvania-New Jersey-Maryland, an independent system operator, responsible for the operation and management of the 13 state and Columbia district electrical power systems in the United states) into the network load prediction model.
Compared with the prior art, the beneficial effects are: the invention takes the deviation of the target value and the predicted value as the correction of a new prediction to form a new target function, thereby establishing a new neural network load prediction model with deviation feedback, improving the corresponding optimization algorithm and improving the accuracy of load prediction.
Drawings
FIG. 1 is a flow chart of a short-term load prediction method based on a BP neural network according to the present invention;
FIG. 2 is a diagram of a model of a neural network of the present invention;
FIG. 3 is a flow chart of the training of the neural network predictive model of the present invention.
Detailed Description
The drawings are for illustrative purposes only and are not to be construed as limiting the patent; for the purpose of better illustrating the embodiments, certain features of the drawings may be omitted, enlarged or reduced, and do not represent the size of an actual product; it will be understood by those skilled in the art that certain well-known structures in the drawings and descriptions thereof may be omitted. The positional relationships depicted in the drawings are for illustrative purposes only and are not to be construed as limiting the present patent.
The same or similar reference numerals in the drawings of the embodiments of the present invention correspond to the same or similar components; in the description of the present invention, it should be understood that if there are terms such as "upper", "lower", "left", "right", "long", "short", etc., indicating orientations or positional relationships based on the orientations or positional relationships shown in the drawings, it is only for convenience of description and simplicity of description, but does not indicate or imply that the device or element referred to must have a specific orientation, be constructed in a specific orientation, and be operated, and therefore, the terms describing the positional relationships in the drawings are only used for illustrative purposes and are not to be construed as limitations of the present patent, and specific meanings of the terms may be understood by those skilled in the art according to specific situations.
The technical scheme of the invention is further described in detail by the following specific embodiments in combination with the attached drawings:
examples
Fig. 1 to 3 show an embodiment of a short-term load prediction method based on a BP neural network, comprising the following steps:
the first step is as follows: establishing a traditional neural network model, wherein the formula of an objective function is as follows:
wherein, tphDenotes the actual load at p hour on day h, yphRepresenting the predicted load at the p hour of the h day, Eav (r)' representing the objective function of load prediction, and r representing the number of iterations;
step two: on the basis of the traditional neural network, a period of adjustable time zone is selected in a section of place where the same model is predicted, the deviation between a target value and a predicted value on a time zone is used as the correction of a new prediction to form a new target function, and the formula is as follows:
where Eav (r) represents the objective function of the load prediction, tphDenotes the actual load at p hour on day h, yphDenotes the predicted load at p hour on day h, tjkDenotes the actual load at the j-th hour on the k-th day, yjkRepresents the predicted load at the j hour on the k day, and alpha and beta represent non-0 natural numbers, wherein alpha<Beta, r represents the iteration times of the neural network, H is the number of training samples, and P is the output space dimension;
step three: and (5) constructing a neural network load prediction model according to the objective function in the step two, and training the neural network load prediction model.
Preferably, the formula of the relative deviation of the predicted value and the target value of the conventional neural network is as follows:
where T is the actual load and Y is the predicted load.
Preferably, the neural network is composed of an input layer, a hidden layer and an output layer, and the output of the mth neuron of a single sample in the hidden layer is:
wherein, wimFor weights of input layer to hidden layer, xiFor input features and consisting of predicting the load per hour n days before the day and predicting the weather per hour on the day, bmIs the bias of the input layer to the hidden layer;
the output of a single sample at the p-th neuron in the output layer is:
wherein, wmpFor the weight from hidden layer to output layer, hmOutput of the mth neuron of the hidden layer, bpIs the bias of the hidden layer to the output layer.
Preferably, the output layer adds an average value of deviations between a target value and a predicted value of a day from α to β in the current day as a correction on the basis of a Y value output by a conventional neural network, and obtains an output characteristic formula as follows:
wherein Z isphOutput load at p hour of h day, tjkDenotes the actual load at the j-th hour on the k-th day, yjkRepresents the predicted load at the j-th hour on the k-th day, yphThe predicted load at the P-th hour on the h-th day is represented, and P is the output space dimension.
Preferably, the neural network load prediction model optimizes the weight and bias through the objective function, and the update formula of the weight and bias from the input layer to the hidden layer is as follows:
updating weight and bias from hidden layer to output layer:
where r denotes the number of iterations of the neural network, wimAs weights of the input layer to the hidden layer, bmIs the bias of the input layer to the hidden layer; w is ampWeights for the hidden layer to the output layer, bpIs the bias of the hidden layer to the output layer; ε is the learning rate, which ranges from (0, 1).
Preferably, the neural network training is performed by substituting the short-term load data of the PJM (Pennsylvania-New Jersey-Maryland, an independent system operator, responsible for the operation and management of the 13 state and Columbia district electrical power systems in the United states) into the network load prediction model.
The beneficial effect of this embodiment is: the invention takes the deviation of the target value and the predicted value as the correction of a new prediction to form a new target function, thereby establishing a new neural network load prediction model with deviation feedback, improving the corresponding optimization algorithm and improving the accuracy of load prediction.
It should be understood that the above-described embodiments of the present invention are merely examples for clearly illustrating the present invention, and are not intended to limit the embodiments of the present invention. Other variations and modifications will be apparent to persons skilled in the art in light of the above description. And are neither required nor exhaustive of all embodiments. Any modification, equivalent replacement, and improvement made within the spirit and principle of the present invention should be included in the protection scope of the claims of the present invention.
Claims (6)
1. A short-term load prediction method based on a BP neural network is characterized by comprising the following steps:
the method comprises the following steps: establishing a traditional neural network model, wherein the formula of an objective function is as follows:
wherein, tphDenotes the actual load at p hour on day h, yphRepresenting the predicted load at the p hour of the h day, Eav (r)' representing the objective function of load prediction, and r representing the number of iterations;
step two: on the basis of the traditional neural network, a period of adjustable time zone is selected in a section of place where the same model is predicted, the deviation between a target value and a predicted value on a time zone is used as the correction of a new prediction to form a new target function, and the formula is as follows:
where Eav (r) represents the objective function of the load prediction, tphDenotes the actual load at p hour on day h, yphDenotes the predicted load at p hour on day h, tjkDenotes the actual load at the j-th hour on the k-th day, yjkRepresents the predicted load at the j hour on the k day, and alpha and beta represent non-0 natural numbers, wherein alpha<Beta, r represents the iteration times of the neural network, H is the number of training samples, and P is the output space dimension;
step three: and (5) constructing a neural network load prediction model according to the objective function in the step two, and training the neural network load prediction model.
2. The method of claim 1, wherein the formula of the relative deviation of the conventional neural network prediction value from the target value is as follows:
where T is the actual load and Y is the predicted load.
3. The method of claim 2, wherein the neural network comprises an input layer, a hidden layer and an output layer, and the output of the mth neuron of a single sample in the hidden layer is:
wherein, wimFor weights of input layer to hidden layer, xiFor input features and consisting of predicting the load per hour n days before the day and predicting the weather per hour on the day, bmIs the bias of the input layer to the hidden layer;
the output of a single sample at the p-th neuron in the output layer is:
wherein, wmpFor the weight from hidden layer to output layer, hmOutput of the mth neuron of the hidden layer, bpIs the bias of the hidden layer to the output layer.
4. The method of claim 3, wherein the output layer adds an average of deviations between the predicted values and the target values of a days a to β days before the current day as a correction to the Y value output by the conventional neural network, and obtains an output characteristic formula as follows:
wherein Z isphOutput load at p hour of h day, tjkDenotes the actual load at the j-th hour on the k-th day, yjkRepresents the predicted load at the j-th hour on the k-th day, yphThe predicted load at the P-th hour on the h-th day is represented, and P is the output space dimension.
5. The method of claim 4, wherein the neural network load prediction model optimizes weights and biases through the objective function, and the update formula of the input layer-to-hidden layer weights and biases is as follows:
updating weight and bias from hidden layer to output layer:
where r denotes the number of iterations of the neural network, wimAs weights of the input layer to the hidden layer, bmIs the bias of the input layer to the hidden layer; w is ampWeights for the hidden layer to the output layer, bpIs the bias of the hidden layer to the output layer; ε is the learning rate, which ranges from (0, 1).
6. The method for short-term load prediction based on BP neural network as claimed in any of claims 1-5, wherein the neural network training is performed by substituting the PJM short-term load data into the network load prediction model.
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201910819380.4A CN110619429A (en) | 2019-08-31 | 2019-08-31 | Short-term load prediction method based on BP neural network |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201910819380.4A CN110619429A (en) | 2019-08-31 | 2019-08-31 | Short-term load prediction method based on BP neural network |
Publications (1)
Publication Number | Publication Date |
---|---|
CN110619429A true CN110619429A (en) | 2019-12-27 |
Family
ID=68922903
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN201910819380.4A Pending CN110619429A (en) | 2019-08-31 | 2019-08-31 | Short-term load prediction method based on BP neural network |
Country Status (1)
Country | Link |
---|---|
CN (1) | CN110619429A (en) |
Citations (4)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN103020743A (en) * | 2012-12-27 | 2013-04-03 | 中国科学院电工研究所 | Ultra-short-term wind speed forecasting method for wind power plant |
CN103295081A (en) * | 2013-07-02 | 2013-09-11 | 上海电机学院 | Electrical power system load prediction method based on back propagation (BP) neural network |
CN105354620A (en) * | 2014-08-19 | 2016-02-24 | 中国科学院上海高等研究院 | Method for predicting fan generation power |
CN109376950A (en) * | 2018-11-19 | 2019-02-22 | 国网陕西省电力公司电力科学研究院 | A kind of polynary Load Forecasting based on BP neural network |
-
2019
- 2019-08-31 CN CN201910819380.4A patent/CN110619429A/en active Pending
Patent Citations (4)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN103020743A (en) * | 2012-12-27 | 2013-04-03 | 中国科学院电工研究所 | Ultra-short-term wind speed forecasting method for wind power plant |
CN103295081A (en) * | 2013-07-02 | 2013-09-11 | 上海电机学院 | Electrical power system load prediction method based on back propagation (BP) neural network |
CN105354620A (en) * | 2014-08-19 | 2016-02-24 | 中国科学院上海高等研究院 | Method for predicting fan generation power |
CN109376950A (en) * | 2018-11-19 | 2019-02-22 | 国网陕西省电力公司电力科学研究院 | A kind of polynary Load Forecasting based on BP neural network |
Non-Patent Citations (1)
Title |
---|
李灿: "基于改进BP神经网络的负荷预测问题研究", 《中国优秀博硕士学位论文全文数据库(硕士)工程科技II辑》 * |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
CN106600059B (en) | Intelligent power grid short-term load prediction method based on improved RBF neural network | |
JP5888640B2 (en) | Photovoltaic power generation prediction apparatus, solar power generation prediction method, and solar power generation prediction program | |
CN107590567A (en) | Recurrent neural network short-term load prediction method based on information entropy clustering and attention mechanism | |
CN110298501A (en) | Electric load prediction technique based on long Memory Neural Networks in short-term | |
CN111308896B (en) | Nonlinear system self-adaptive optimal control method based on variable error | |
CN107527116B (en) | Short-term load prediction method based on support vector regression | |
CN107730006A (en) | A kind of nearly zero energy consumption controller of building based on regenerative resource big data deep learning | |
Xiao et al. | Rational and self-adaptive evolutionary extreme learning machine for electricity price forecast | |
CN105574586B (en) | General-purpose aircraft boat material needing forecasting method based on MPSO-BP networks | |
CN110110380B (en) | Piezoelectric actuator hysteresis nonlinear modeling method and application | |
Tan et al. | Support vector machine algorithm for artificial intelligence optimization | |
CN116526473A (en) | Particle swarm optimization LSTM-based electrothermal load prediction method | |
Ko et al. | Feedforward error learning deep neural networks for multivariate deterministic power forecasting | |
Motlagh et al. | A new architecture for modeling and prediction of dynamic systems using neural networks: Application in Tehran stock exchange | |
CN110619429A (en) | Short-term load prediction method based on BP neural network | |
Tang et al. | Model and method based on GA for nonlinear programming problems with fuzzy objective and resources | |
CN111524348A (en) | Long-short term traffic flow prediction model and method | |
CN111369072A (en) | Nuclear minimum mean square time sequence online prediction model based on sparsification method | |
CN111461462B (en) | Daily load prediction method based on trellis Net-LSTM | |
Kayakuş | The Estimation of Turkey's Energy Demand Through Artificial Neural Networks and Support Vector Regression Methods | |
CN115660347A (en) | Energy storage system optimal configuration method facing multiple constraints and nonlinear targets | |
CN114444763A (en) | Wind power prediction method based on AFSA-GNN | |
CN113591391A (en) | Power load control device, control method, terminal, medium and application | |
Xiang et al. | Application of discrete Grey Wolf Algorithm in balanced transport problem | |
CN114387028A (en) | Intelligent analysis method for commodity demand of online shopping platform |
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 | ||
RJ01 | Rejection of invention patent application after publication |
Application publication date: 20191227 |