CN114444750A - Power load prediction method based on neural network - Google Patents
Power load prediction method based on neural network Download PDFInfo
- Publication number
- CN114444750A CN114444750A CN202011209003.8A CN202011209003A CN114444750A CN 114444750 A CN114444750 A CN 114444750A CN 202011209003 A CN202011209003 A CN 202011209003A CN 114444750 A CN114444750 A CN 114444750A
- Authority
- CN
- China
- Prior art keywords
- power load
- load prediction
- prediction model
- training
- data
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Pending
Links
- 238000000034 method Methods 0.000 title claims abstract description 38
- 238000013528 artificial neural network Methods 0.000 title claims abstract description 32
- 238000012549 training Methods 0.000 claims abstract description 45
- 238000012360 testing method Methods 0.000 claims abstract description 38
- 210000002569 neuron Anatomy 0.000 claims description 13
- 230000006870 function Effects 0.000 claims description 9
- 230000008569 process Effects 0.000 claims description 8
- 238000012545 processing Methods 0.000 claims description 7
- 238000010606 normalization Methods 0.000 claims description 6
- 238000012937 correction Methods 0.000 claims description 3
- 230000005284 excitation Effects 0.000 claims description 3
- 238000013277 forecasting method Methods 0.000 claims 1
- 230000009466 transformation Effects 0.000 description 4
- 230000009286 beneficial effect Effects 0.000 description 2
- 230000008859 change Effects 0.000 description 2
- 230000000694 effects Effects 0.000 description 2
- 238000011425 standardization method Methods 0.000 description 2
- 230000007547 defect Effects 0.000 description 1
- 230000010365 information processing Effects 0.000 description 1
- 238000005457 optimization Methods 0.000 description 1
- 238000010248 power generation Methods 0.000 description 1
- 238000011002 quantification Methods 0.000 description 1
- 230000000306 recurrent effect Effects 0.000 description 1
- 230000001932 seasonal effect Effects 0.000 description 1
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06Q—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
- G06Q10/00—Administration; Management
- G06Q10/04—Forecasting or optimisation specially adapted for administrative or management purposes, e.g. linear programming or "cutting stock problem"
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/02—Neural networks
- G06N3/04—Architecture, e.g. interconnection topology
- G06N3/044—Recurrent networks, e.g. Hopfield networks
-
- 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
- G06N3/04—Architecture, e.g. interconnection topology
- G06N3/045—Combinations of networks
-
- 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
- G06Q50/00—Systems or methods specially adapted for specific business sectors, e.g. utilities or tourism
- G06Q50/06—Electricity, gas or water supply
Abstract
The invention discloses a power load prediction method based on a neural network, which comprises the following steps: acquiring historical data of the power load, and establishing a sample library; randomly dividing a sample library into a training set and a testing set; taking the date Tn, the air temperature In and the weather phi n as inputs, taking the power load value t as an output, and establishing a power load prediction model based on a BP neural network; training the power load prediction model according to the training set; testing the power load prediction model after training according to the test set, and calculating the error of the power load prediction model until the power load prediction model which is qualified in testing is obtained; and inputting the power load data to be predicted into a power load prediction model which is qualified in the test, and predicting a corresponding power load value. According to the method, the related data influencing the power load are utilized to establish the power load prediction model based on the BP neural network, and the efficiency and the precision of power load prediction are improved.
Description
Technical Field
The invention relates to the technical field of information processing, in particular to a power load prediction method based on a neural network.
Background
Electric energy is the most important energy source in the world at present, but the defect that the electric energy is difficult to store is not effectively solved, and a power generation plan and a load demand are required to reach dynamic balance, so high-precision load prediction is an important guarantee for normal and safe operation and high power supply quality of a power system.
In recent years, because the neural network has strong self-learning and nonlinear fitting capabilities, the problem of power load prediction can be solved well, and the application of the neural network in the power load prediction becomes a hot point. The existing method for predicting the power load by using the neural network generally comprises the steps of compensating the predicted load output by a neural network prediction model by using a rough set theory so as to predict the power load, or predicting the short-term load by using an ant colony optimization algorithm-based recurrent neural network. However, both of them do not fully utilize information in the power load data, and the accuracy of power load prediction is low.
In summary, how to effectively solve the problems that the prior art cannot fully utilize information in the power load data and the power load prediction accuracy is low is a problem that needs to be solved urgently by those skilled in the art at present.
Disclosure of Invention
The invention provides a power load prediction method based on a neural network, which solves the problem of low prediction precision of the existing power load.
The invention achieves the above purpose through the following technical scheme:
a neural network-based power load prediction method includes:
acquiring historical data of the power load, and establishing a sample library;
randomly dividing a sample library into a training set and a testing set;
taking the date Tn, the air temperature In and the weather phi n as inputs, taking the power load value t as an output, and establishing a power load prediction model based on a BP neural network;
training the power load prediction model according to the training set;
testing the power load prediction model after training according to the test set, and calculating the error of the power load prediction model until the power load prediction model which is qualified in testing is obtained;
and inputting the power load data to be predicted into a power load prediction model which is qualified in the test, and predicting a corresponding power load value.
Further, after the steps of obtaining historical data of the power load and establishing a sample library, the method further comprises the following steps:
and processing the sample library according to the characteristic attributes.
Further, the characteristic attributes include date, temperature, and weather.
Further, after the step of processing the sample library according to the characteristic attributes, the method further comprises the following steps:
the data of the sample library is converted into the same format.
Further, after the step of converting the data of the sample library into the same format, the method further comprises the following steps:
normalizing the data of the sample library by adopting a min-max standardization method, wherein for each attribute, minA and maxA are respectively set as the minimum value and the maximum value of the attribute A, and an original value x of A is mapped into a value x' in an interval [0,1] through min-max standardization, and the formula is as follows: new data is (original data-min)/(max-min).
Further, the training set and the test set are unbalanced sample sets, the number of the training sets accounts for 2/3 of the number of the sample banks, and the number of the test sets accounts for 1/3 of the number of the sample banks.
Furthermore, the number of neurons of an input layer of the BP neural network in the power load prediction model is 3, the number of neurons of an output layer is 1, the number of neurons of a hidden layer is 7, the neurons of the hidden layer adopt an S-type transformation function, and the output layer is a linear transformation function.
Further, in the step of training the power load prediction model according to the training set,
and the BP neural network trains the power load prediction model by adopting a small batch gradient descent algorithm.
Further, the specific process of training the power load prediction model according to the training set is as follows:
the power load data are transmitted to the hidden layer from the input layer, the hidden layer transmits the processed result to the output layer through the weight and the excitation function, the result of the output layer is compared with the expected value to obtain an error, then the feedback correction is carried out on the weight in the power load prediction model through reverse pushing, and corresponding data generated regularly are trained for multiple times on the power load prediction model to complete learning.
By adopting the technical scheme, the invention mainly has the following technical effects: the method comprises the steps of establishing a power load prediction model based on a BP neural network by using relevant data influencing power loads, training the power load value by using date, temperature and weather as input and using the power load value as output, predicting the power load value by using the trained power load prediction model, and improving the efficiency and the precision of power load prediction.
Detailed Description
The technical solutions in the embodiments of the present invention will be described clearly and completely below, and it is obvious that the described embodiments are only a part of the embodiments of the present invention, and not all embodiments. All other embodiments, which can be derived by a person skilled in the art from the embodiments given herein without making any creative effort, shall fall within the protection scope of the present invention.
The first embodiment is as follows:
the invention provides a power load prediction method based on a neural network, which comprises the following steps:
step 11: acquiring historical data of the power load, and establishing a sample library;
step 12: randomly dividing a sample library into a training set and a testing set;
step 13: taking the date Tn, the air temperature In and the weather phi n as inputs, taking the power load value t as an output, and establishing a power load prediction model based on a BP neural network;
step 14: training the power load prediction model according to the training set;
step 15: testing the power load prediction model after training according to the test set, and calculating the error of the power load prediction model until the power load prediction model which is qualified in testing is obtained;
step 16: and inputting the power load data to be predicted into a power load prediction model which is qualified in the test, and predicting a corresponding power load value.
As described above in step 11, historical data of the power load is obtained, and a sample library is created, wherein raw data generally has problems and must be processed for analysis in order to improve the data quality of the power load.
As in step 12 above, the sample library is randomly divided into training sets and test sets, where the training sets and the test sets are unbalanced sample sets, the number of training sets is 2/3 times the number of sample libraries, and the number of test sets is 1/3 times the number of sample libraries. The training set is used for network training, and the testing set is used for testing the generalization ability of the network.
As described above In step 13, the power load prediction model based on the BP neural network is established using the date Tn, the air temperature In, and the weather φ n as inputs and the power load value t as an output, wherein the date, the air temperature, and the weather are characteristic information of the power load prediction model. The number of neurons of an input layer of a BP neural network in the power load prediction model is 3, the number of neurons of an output layer is 1, the number of neurons of a hidden layer is 7, the neurons of the hidden layer adopt an S-shaped transformation function, and the output layer is a linear transformation function.
As the step 14, the power load prediction model is trained according to the training set, wherein the specific process is as follows: the power load data are transmitted to the hidden layer from the input layer, the hidden layer transmits the processed result to the output layer through the weight and the excitation function, the result of the output layer is compared with the expected value to obtain an error, then the feedback correction is carried out on the weight in the power load prediction model through reverse pushing, and corresponding data generated regularly are trained for multiple times on the power load prediction model to complete learning. The BP neural network adopts an error back propagation learning algorithm, a proper learning algorithm needs to be selected to facilitate network training, the BP neural network of the embodiment adopts a small-batch gradient descent algorithm, and the training algorithm has the advantages of fast convergence and high precision.
And step 15, testing the power load prediction model after training according to the test set, and calculating the error of the power load prediction model until the power load prediction model which is qualified in testing is obtained, wherein the generalization capability is evaluated by calculating the error between the predicted value and the true value of the test set. The smaller the relative error is, the better the performance of the power load prediction model is, if the generalization ability meets the requirement, the trained power load prediction model can be used for prediction, otherwise, the above step 14 needs to be returned to continue training until the generalization ability meets the requirement.
As shown in step 16, the power load data to be predicted is input into the power load prediction model qualified in training, and the corresponding power load value is predicted, wherein the specific process is as follows: and inputting data such as date, temperature, weather and the like into the power load prediction model after a series of processing, wherein the output of the power load prediction model is a prediction result.
Example two:
the embodiment of the invention provides a power load prediction method based on a neural network, which comprises the following steps:
step 21: acquiring historical data of the power load, and establishing a sample library; for the specific description of this step, reference may be made to the specific description of step 11, which is not described herein again;
step 22: processing the sample library according to the characteristic attributes;
step 23: converting the data of the sample library into the same format;
step 24: carrying out normalization processing on the data of the sample library by adopting a min-max standardization method;
step 25: randomly dividing a sample library into a training set and a testing set; for the specific description of this step, reference may be made to the specific description of step 12, which is not described herein again;
step 26: taking the date Tn, the air temperature In and the weather phi n as inputs, taking the power load value t as an output, and establishing a power load prediction model based on a BP neural network; for the specific description of this step, reference may be made to the specific description of step 13, which is not described herein again;
step 27: training the power load prediction model according to a training set; for the specific description of this step, reference may be made to the specific description of step 14, which is not described herein again;
step 28: testing the power load prediction model after training according to the test set, and calculating the error of the power load prediction model until the power load prediction model which is qualified in testing is obtained; for the specific description of this step, reference may be made to the specific description of step 15, which is not described herein again;
step 29: inputting the power load data to be predicted into a power load prediction model which is qualified in the test, and predicting a corresponding power load value; for the detailed description of this step, reference may be made to the detailed description of step 16, which is not described herein again;
as shown in step 22, the sample library is processed according to the characteristic attributes, wherein the relevant characteristic attributes of the power load prediction are determined, and the characteristic samples are screened according to the characteristic attributes, which include various influence factors that may influence the prediction result of the object to be predicted. The load of the power system is influenced by more factors, not only by factors such as power load demand, weather conditions, seasonality and regions, but also by factors such as national economy, politics and residential habits. Since the influence of the date, temperature, and weather factors on the power load is greater than the influence of other factors on the power load, which is beneficial to improving the prediction accuracy, in this embodiment, the date, temperature, and weather are selected as the characteristic information of the power load prediction model. When the air temperature changes within a certain range, the influence on the load change is basically similar, so that quantized values between [0,1] are given to different air temperatures. And aiming at the weather characteristics, according to historical load data and seasonal characteristics, a quantitative value from large to small between [0,1] is given from clear weather to cloudy weather to rainstorm. The date types are mainly processed in a quantification mode according to working days and rest days.
As shown in step 23, the data in the sample library is converted into the same format, wherein the original data is generally obtained from various actual application systems (multiple databases, multiple file systems), and the formats of these systems are different, so that the data needs to be converted into the same format.
As shown in step 24, the data in the sample library is normalized by the min-max normalization method, where minA and maxA are respectively set as the minimum value and the maximum value of the attribute a for each attribute, and an original value x of a is mapped to a value x' in the interval [0,1] by the min-max normalization, and the formula is: the new data is (original data-minimum)/(maximum-minimum), and the normalization process is beneficial to eliminating the influence when different attributes of the sample have different magnitudes: 1. the difference in magnitude will result in the property of the magnitude being dominant; 2. the difference in order of magnitude will result in a slow convergence speed of the iteration.
The invention provides a power load prediction method based on a neural network, which mainly has the following technical effects: the method comprises the steps of establishing a power load prediction model based on a BP neural network by using relevant data influencing power loads, training the power load value by using date, temperature and weather as input and using the power load value as output, predicting the power load value by using the trained power load prediction model, and improving the efficiency and the precision of power load prediction.
It should be noted that the terms "comprises," "comprising," or any other variation thereof, are intended to cover a non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements does not include only those elements but may include other elements not expressly listed or inherent to such process, method, article, or apparatus. Without further limitation, an element defined by the phrase "comprising … …" does not exclude the presence of other identical elements in a process, method, article, or apparatus that comprises the element.
The foregoing detailed description of the embodiments of the present application has been presented to illustrate the principles and implementations of the present application, and the above description of the embodiments is only provided to help understand the method and the core concept of the present application; meanwhile, for a person skilled in the art, according to the idea of the present application, there may be a change in the specific implementation manner and the application scope, and the above-mentioned embodiments are only preferred embodiments of the present invention, and do not limit the technical solution of the present invention, and as long as the technical solution can be implemented on the basis of the above-mentioned embodiments without creative work, the technical solution should be considered to fall within the protection scope of the claims of the present patent application, and in sum, the content of the present specification should not be construed as a limitation to the present application.
Claims (9)
1. A power load prediction method based on a neural network is characterized by comprising the following steps:
acquiring historical data of a power load, and establishing a sample library;
randomly dividing a sample library into a training set and a testing set;
taking the date Tn, the air temperature In and the weather phi n as inputs, taking the power load value t as an output, and establishing a power load prediction model based on a BP neural network;
training the power load prediction model according to the training set;
testing the power load prediction model after training according to the test set, and calculating the error of the power load prediction model until the power load prediction model which is qualified in testing is obtained;
and inputting the power load data to be predicted into a power load prediction model which is qualified in the test, and predicting a corresponding power load value.
2. The method of claim 1, wherein after the steps of obtaining historical data of the electrical load and building a sample library, the method further comprises:
and processing the sample library according to the characteristic attributes.
3. The neural network-based power load forecasting method according to claim 2, wherein the characteristic attributes include date, temperature and weather.
4. The method of claim 2, wherein after the step of processing the sample library according to the characteristic attributes, the method further comprises:
the data of the sample library is converted into the same format.
5. The method of claim 4, further comprising, after the step of converting the data in the sample library into the same format:
normalizing the data of the sample library by adopting a min-max normalization method, wherein minA and maxA are respectively set as the minimum value and the maximum value of the attribute A for each attribute, and an original value x of A is mapped into a value x' in an interval [0,1] through min-max normalization, and the formula is as follows: new data is (original data-min)/(max-min).
6. The method of claim 1, wherein the training set and the testing set are unbalanced sample sets, the number of training sets is 2/3 times the number of sample banks, and the number of testing sets is 1/3 times the number of sample banks.
7. The method of claim 1, wherein the number of neurons in the input layer, the number of neurons in the output layer, the number of neurons in the hidden layer, and the number of neurons in the output layer of the BP neural network in the power load prediction model are 3, 1, and 7, respectively, and the neurons in the hidden layer adopt an S-type transform function and the output layer is a linear transform function.
8. The method according to claim 1, wherein in the step of training the power load prediction model according to the training set, the BP neural network trains the power load prediction model by using a small batch gradient descent algorithm.
9. The method according to claim 8, wherein the specific process of training the power load prediction model according to the training set is as follows:
the power load data are transmitted to the hidden layer from the input layer, the hidden layer transmits the processed result to the output layer through the weight and the excitation function, the result of the output layer is compared with the expected value to obtain an error, then the feedback correction is carried out on the weight in the power load prediction model through reverse pushing, and corresponding data generated regularly are trained for multiple times on the power load prediction model to complete learning.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202011209003.8A CN114444750A (en) | 2020-11-03 | 2020-11-03 | Power load prediction method based on neural network |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202011209003.8A CN114444750A (en) | 2020-11-03 | 2020-11-03 | Power load prediction method based on neural network |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN114444750A true CN114444750A (en) | 2022-05-06 |
Family
ID=81361119
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202011209003.8A Pending CN114444750A (en) | 2020-11-03 | 2020-11-03 | Power load prediction method based on neural network |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN114444750A (en) |
Cited By (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN115589000A (en) * | 2022-12-12 | 2023-01-10 | 广东电网有限责任公司阳江供电局 | Area voltage adjustment optimization method and system based on typical load classification |
-
2020
- 2020-11-03 CN CN202011209003.8A patent/CN114444750A/en active Pending
Cited By (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN115589000A (en) * | 2022-12-12 | 2023-01-10 | 广东电网有限责任公司阳江供电局 | Area voltage adjustment optimization method and system based on typical load classification |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN110909912B (en) | Park electric power system net load combination prediction method based on self-adaptive error feedback | |
| CN110414788B (en) | Electric energy quality prediction method based on similar days and improved LSTM | |
| Li et al. | A novel grey forecasting model and its application in forecasting the energy consumption in Shanghai | |
| CN109242265B (en) | Urban water demand combined prediction method based on least square sum of errors | |
| CN102426674A (en) | Power system load prediction method based on Markov chain | |
| CN110135635A (en) | A kind of region electric power saturation load forecasting method and system | |
| CN102509173A (en) | Markov chain based method for accurately forecasting power system loads | |
| CN108595892A (en) | Soft-measuring modeling method based on time difference model | |
| CN111008870A (en) | Regional logistics demand prediction method based on PCA-BP neural network model | |
| Ishikawa et al. | Short‐Term Wind Power Prediction for Wind Turbine via Kalman Filter Based on JIT Modeling | |
| CN116227637A (en) | Active power distribution network oriented refined load prediction method and system | |
| CN110298765B (en) | Power distribution network power consumption abnormality detection method based on objective correlation factors | |
| CN115545333A (en) | Method for predicting load curve of multi-load daily-type power distribution network | |
| CN114444750A (en) | Power load prediction method based on neural network | |
| CN111652422A (en) | Heat supply system load prediction method, device and system based on building classification | |
| CN114819395A (en) | Industry medium and long term load prediction method based on long and short term memory neural network and support vector regression combination model | |
| CN114329347A (en) | Method and device for predicting metering error of electric energy meter and storage medium | |
| CN112330030B (en) | System and method for predicting requirements of expansion materials | |
| Wu | The comparison of forecasting analysis based on the ARIMA-LSTM hybrid models | |
| Guo et al. | Wind speed forecasting of genetic neural model based on rough set theory | |
| CN117273195A (en) | Steam heating system demand load prediction method | |
| CN114997470A (en) | Short-term power load prediction method based on LSTM neural network | |
| Hua | Power quality prediction of active distribution network based on CNN-LSTM deep learning model | |
| CN114912700A (en) | Factory workshop electric power energy consumption assessment method and system | |
| Li | The Study of Short Term Wind Power Prediction Based on MV-LSTM |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| WD01 | Invention patent application deemed withdrawn after publication |
Application publication date: 20220506 |