CN113205223A - Electric quantity prediction system and prediction method thereof - Google Patents

Electric quantity prediction system and prediction method thereof Download PDF

Info

Publication number
CN113205223A
CN113205223A CN202110532962.1A CN202110532962A CN113205223A CN 113205223 A CN113205223 A CN 113205223A CN 202110532962 A CN202110532962 A CN 202110532962A CN 113205223 A CN113205223 A CN 113205223A
Authority
CN
China
Prior art keywords
data
electric quantity
prediction
characteristic
prediction method
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
Application number
CN202110532962.1A
Other languages
Chinese (zh)
Inventor
温蜜
罗俊然
周绍景
Current Assignee (The listed assignees may be inaccurate. Google has not performed a legal analysis and makes no representation or warranty as to the accuracy of the list.)
Shanghai University of Electric Power
Shanghai Electric Power University
University of Shanghai for Science and Technology
Original Assignee
Shanghai Electric Power University
Priority date (The priority date 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 date listed.)
Filing date
Publication date
Application filed by Shanghai Electric Power University filed Critical Shanghai Electric Power University
Priority to CN202110532962.1A priority Critical patent/CN113205223A/en
Publication of CN113205223A publication Critical patent/CN113205223A/en
Pending legal-status Critical Current

Links

Images

Classifications

    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06Q—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
    • G06Q10/00—Administration; Management
    • G06Q10/04—Forecasting or optimisation specially adapted for administrative or management purposes, e.g. linear programming or "cutting stock problem"
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06F—ELECTRIC DIGITAL DATA PROCESSING
    • G06F17/00—Digital computing or data processing equipment or methods, specially adapted for specific functions
    • G06F17/10—Complex mathematical operations
    • G06F17/18—Complex mathematical operations for evaluating statistical data, e.g. average values, frequency distributions, probability functions, regression analysis
    • 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
    • G06N3/04—Architecture, e.g. interconnection topology
    • G06N3/047—Probabilistic or stochastic 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
    • G06N3/049—Temporal neural networks, e.g. delay elements, oscillating neurons or pulsed inputs
    • 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
    • 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 relates to an electric quantity prediction system and a prediction method thereof, belonging to the field of electric quantity prediction. Because the electric quantity measurement model is improved and the quantile regression model is introduced, the valuable electric quantity reference interval can be predicted more accurately, so that the interference of data noise is effectively reduced, the characteristics of the electric quantity are further extracted from time and space by combining the advantages of a long-term and short-term memory network and a convolutional neural network, the weight is updated by using an attention mechanism, the characteristic extraction effect is optimized, the probability prediction is realized by using the quantile regression model, and the credible electric quantity prediction reference range can be provided for researchers on the basis of ensuring the accurate point prediction.

Description

Electric quantity prediction system and prediction method thereof
Technical Field
The invention relates to the field of electric quantity prediction, in particular to an electric quantity prediction system and a prediction method thereof.
Background
The electric power company usually counts the line loss rate to reflect the management level and the economic benefit of the electric power company, but the line loss rate is more difficult to count due to the fact that the distributed energy is connected to the power grid and the phenomenon that the power is supplied and sold in different periods exists. Therefore, achieving short-term power prediction is crucial to reducing line loss statistical errors.
With the development of mathematical theory and modern technology, further improvement of the accuracy of short-term power prediction means considerable economic benefits to society, and these potential benefits promote continuous improvement of power prediction models. The regional power prediction can reduce the line loss rate statistical error, provide reasonable distributed energy system planning for the power system, and reduce the operating pressure of the power grid, which are all concerned by power enterprises. Therefore, a new method for reducing the line loss rate error and providing an intuitive reference range is needed by power enterprises.
In the related art, a statistical method is used, for example, an ARIMA (differential Integrated Moving Average Autoregressive model), which predicts future electric quantity according to a history change rule of the electric quantity, but finds a rule from itself without being combined with other characteristics for analysis, and meanwhile, the fitting effect on nonlinear data is not good. In addition, a neural network method is more used, for example, there is a method of predicting the power amount using LSTM (Long Short-Term Memory network), but correlation on a feature space is not considered, resulting in insufficient feature capture for influencing the power amount. There is also a method of predicting the amount of electricity using CNN (Convolutional Neural Networks), but it does not consider temporal context, resulting in failure to extract temporal period effects. In the electric quantity prediction, besides using a proper model, accurate data characteristics can be constructed to effectively improve the prediction precision of the model, but the existing characteristic processing method cannot well find out influential characteristics. Most of the existing related technologies are based on point prediction, but when the electric quantity fluctuates greatly, the accuracy of the point prediction cannot meet the requirement, in order to provide a reference value more reasonably, a probability interval in which the electric quantity possibly appears needs to be presented, the existing probability prediction technology is still in a starting stage, the accuracy of the prediction interval is not enough, and an overlarge or wrong interval provides a negative reference value.
Disclosure of Invention
The present invention has been made to solve the above-described problems, and an object of the present invention is to provide an electric power prediction system and a prediction method thereof.
The electric quantity prediction method provided by the invention has the characteristics that the method comprises the following steps:
step 1, acquiring data to be processed from a scheduling center, wherein the data to be processed comprises electric quantity and a first characteristic;
step 2, processing the data to be processed to obtain first data, and the specific substeps comprise:
step 2-1, performing box separation processing on data to be processed through a first method to obtain box separation processing data;
step 2-2, dividing the box processing data into a first electric quantity and a second characteristic, carrying out correlation analysis on the first electric quantity and the second characteristic through a maximum information coefficient algorithm, and screening target characteristics;
step 2-3, decomposing the first electric quantity into a plurality of subsequences by integrating an empirical mode decomposition algorithm;
2-4, performing correlation analysis on the target characteristics and the subsequences through a maximum information coefficient algorithm, and dividing the subsequences into characteristic correlation and non-characteristic correlation;
step 2-5, carrying out normalization processing on continuous variables in the first electric quantity, and carrying out one-hot encoding processing on discrete variables in the first electric quantity so as to obtain first data;
step 3, calculating and analyzing the first data to obtain the predicted electric quantity, wherein the specific substeps comprise:
step 3-1, learning the time relation of the first data by using a first network to obtain time characteristics;
step 3-2, learning the spatial relationship of the first data by using a second network to obtain spatial characteristics;
3-3, updating the weight of the time characteristic and the space characteristic by using an attention mechanism to obtain a final output matrix;
3-4, performing probability prediction on the subsequence through the first model to obtain a predicted subsequence;
and 3-5, calculating the sum of the prediction subsequences through the evaluation indexes, and analyzing to obtain the predicted electric quantity.
The power prediction method provided by the present invention may further have the following features:
the first method in step 2-1 is to divide different temperatures into different boxes, and then obtain the temperature range of each box through calculation, wherein the specific formula is as follows:
Figure BDA0003068653810000041
where d is the range of each bin, μ is the mean of the data, and n is the number of samples.
The power prediction method provided by the present invention may further have the following features:
wherein, the formula of the normalization processing in the step 2-5 is as follows:
Figure BDA0003068653810000042
xnormis a normalized value, xminIs the minimum value of x, xmaxIs the maximum value of x.
The power prediction method provided by the present invention may further have the following features:
wherein, the first network in step 3-1 is a long-term and short-term memory network, and the specific parameters are as follows:
it＝σ(Wi[ht-1，xt]+bi)
ft＝σ(Wf[ht-1，xt]+bf)
ot＝σ(Wo[ht-1，xt]+bo)
c′t＝tanh(Wc[ht-1，xt]+bc)
ct＝fcct-1+itc′t
wherein itIs an input gate, ftIs a forgetting door otIs an output gate, htIs the hidden layer output at time t, ht-1Is the hidden layer output at time t-1, ctIs cell state renewal, ctIs the output of the cell state, Wi，Wf，Wo，WcWeights of input gate, forget gate, output gate and cell state, respectively, bi，bf，bo，bcThe deviations of the input gate, the forgetting gate, the output gate and the cell state, respectively.
The power prediction method provided by the present invention may further have the following features:
wherein, the second network in the step 3-2 is a convolutional neural network, and the sub-steps of obtaining the spatial features are as follows:
step 3-2-1, defining the spatial features as a matrix F ═ Ft+1，ft+2，...，ft+n]TWherein f ist+nIs a forgetting gate.
And 3-2-2, inputting the matrix F into the convolutional neural network.
And 3-2-3, extracting the spatial features by using a filter to obtain the spatial features.
The power prediction method provided by the present invention may further have the following features:
wherein, the expression of the final output matrix calculation process in the step 3-3 is as follows:
f(HT，h)＝HTh
a＝σ(f(HT，h)Wa
V＝HaT
wherein HTIs the transpose matrix of the spatial features, h is the temporal feature, WaIs the weight and V is the final output matrix.
The power prediction method provided by the present invention may further have the following features:
the first model in the step 3-4 is a quantile regression model, and the specific formula is as follows:
Figure BDA0003068653810000051
wherein r is the quantile,
Figure BDA0003068653810000052
are the prediction set and y is the true value.
The power prediction method provided by the present invention may further have the following features:
wherein the evaluation indexes in step 3-5 are MAPE, RMSE, Pinball and Winkler.
The invention also provides an electric quantity prediction system adopting the electric quantity prediction method, which comprises a data acquisition module, a processing module, a prediction module and a control module. The data acquisition module acquires data to be processed from the scheduling center, the processing module processes the data to be processed to obtain first data, the prediction module calculates and analyzes the first data to obtain predicted electric quantity, and the control module controls the data acquisition module, the processing module and the prediction module.
Action and Effect of the invention
According to the electric quantity prediction system and the prediction method thereof, because the electric quantity measurement model is improved and the quantile regression model is introduced, the valuable electric quantity reference interval can be predicted more accurately, so that the interference of data noise is effectively reduced, the characteristics of the electric quantity are further extracted from time and space by combining the advantages of the long-term and short-term memory network and the convolutional neural network, the weight is updated by using an attention mechanism, the characteristic extraction effect is optimized, the probability prediction is realized by using the quantile regression model, and the credible electric quantity prediction reference range can be provided for researchers on the basis of ensuring the accurate point prediction.
Drawings
FIG. 1 is a flowchart illustrating the general steps of a power prediction method according to an embodiment of the present invention;
fig. 2 is a flow chart of substeps of step 2 in a power prediction method according to an embodiment of the invention;
fig. 3 is a flow chart of substeps of step 3 of the power prediction method of an embodiment of the invention;
FIG. 4 is a diagram of a model combination structure for extracting data space features by a convolutional neural network in an embodiment of the present invention; and
fig. 5 is a block diagram of a power prediction system according to an embodiment of the present invention.
Detailed Description
In order to make the technical means, creation features, achievement objectives and effects of the present invention easy to understand, the following embodiments are specifically described with reference to the accompanying drawings.
Fig. 1 is a flowchart illustrating the general steps of a power prediction method according to an embodiment of the present invention.
Fig. 2 is a flow chart of substeps of step 2 in the power prediction method according to an embodiment of the present invention.
Fig. 3 is a flow chart of substeps of step 3 in the power prediction method of the embodiment of the invention.
As shown in fig. 1, the electric quantity prediction method in this embodiment includes the following steps:
step 1, acquiring data to be processed from a scheduling center, and dividing the data to be processed into electric quantity and a first characteristic.
Step 2, processing the data to be processed to obtain first data, as shown in fig. 2, the specific sub-steps include:
step 2-1, in order to avoid the influence of temperature errors on electric quantity prediction, a first method is adopted to perform box separation processing on data to be processed to obtain box separation processing data, the temperature can be divided into certain boxes by calculating the range of each box, each box comprises a certain temperature range, and a box interval formula is as follows:
Figure BDA0003068653810000071
where d is the range of each bin, μ is the mean of the data, and n is the number of samples. And obtaining a first temperature box by adding the range of the box to the lowest temperature, obtaining a second box by adding the range of the box to the upper limit of the first box, and so on to finish temperature box separation.
And 2-2, dividing the box processing data into a first electric quantity and a second characteristic, carrying out correlation analysis on the first electric quantity and the second characteristic through a maximum information coefficient algorithm, and screening target characteristics. The quality of the features determines the upper limit of the model result, but too many features not only can not improve the prediction precision, but also can reduce the model efficiency. Through correlation analysis, redundant features can be eliminated, and therefore the overall efficiency of the model is improved.
And 2-3, in order to avoid possible noise interference, decomposing the first electric quantity into a plurality of subsequences by integrating an empirical mode decomposition algorithm.
And 2-4, because the effective characteristics can further help the model to more accurately learn the characteristics of the data, performing correlation analysis on the target characteristics and the subsequences by using a maximum information coefficient algorithm, and dividing the subsequences into characteristic correlation and non-characteristic correlation.
Step 2-5, normalization processing is carried out on the continuous variable in the first electric quantity, the dimension problem of data can be effectively solved, the model can be trained better, and the formula of the normalization processing is as follows:
Figure BDA0003068653810000081
wherein x isnormIs a normalized value, xminIs the minimum value of x, xmaxIs the maximum value of x.
In addition, the discrete variable in the first electric quantity is subjected to one-hot encoding processing, so that first data are obtained;
step 3, calculating and analyzing the first data to obtain the predicted electric quantity, as shown in fig. 3, the specific sub-steps include:
step 3-1, learning the time relation of the first data by using a long-short term memory network to obtain a time characteristic, wherein a parameter formula of the long-short term memory network is as follows:
it＝σ(Wi[ht-1，xt]+bi)
ft＝σ(Wf[ht-1，xt]+bf)
ot＝σ(Wo[ht-1，xt]+bo)
c′t＝tanh(Wc[ht-1，xt]+bc)
ct＝fcct-1+itc′t
wherein itIs an input gate, ftIs a forgetting door otIs an output gate, htIs the hidden layer output at time t, ht-1Is the hidden layer output at time t-1, ct' is cell State renewal, ctIs the output of the cell state, Wi，Wf，Wo，WcWeights of input gate, forget gate, output gate and cell state, respectively, bi，bf，bo，bcThe deviations of the input gate, the forgetting gate, the output gate and the cell state, respectively.
Through the calculation, the hidden layer output h is obtained[t+1，t+2，...，t+n]Denoted as h.
Fig. 4 is a diagram of a model combination structure for extracting data space features by the convolutional neural network in this embodiment.
Step 3-2, learning the spatial relationship of the first data by using a convolutional neural network to obtain a spatial feature, which is defined as a matrix F ═ F as shown in fig. 2t+1，ft+2，...，ft+n]TInputting the matrix F into the convolutional neural network, and performing feature extraction by using K filters, wherein the filter is defined as K1×dmWherein dm is a characteristic number. The result obtained by the convolutional neural network is denoted as Hi，jHereinafter, it is represented by H;
step 3-3, use notesThe weight updating of the time characteristic and the space characteristic is carried out by the gravity mechanism, and H is firstly transposed into HTAnd obtaining the matrix V after weight updating by the following operations:
f(HT，h)＝HTh
a＝σ(f(HT，h)Wa
V＝HaT
combining the matrixes V and H to obtain a final output matrix H';
and 3-4, performing probability prediction on the subsequence through a quantile regression model to obtain a predicted subsequence, wherein a specific formula is as follows:
Figure BDA0003068653810000101
wherein r is the quantile,
Figure BDA0003068653810000102
the point prediction method is a prediction group, y is a true value, and when r is 0, the requirement of a point prediction result can be satisfied.
And 3-5, calculating the sum of the predictor sequences through evaluation indexes MAPE, RMSE, Pinball and Winkler, and analyzing to obtain the predicted electric quantity.
Fig. 5 is a block diagram of a power prediction system according to an embodiment of the present invention.
As shown in fig. 5, the power prediction system 100 according to the embodiment of the present invention includes a data obtaining module 10, a processing module 20, a prediction module 30, and a control module 40.
The data obtaining module 10 obtains the data to be processed by the method of step 1.
The processing module 20 processes the first data by the method of step 2.
The prediction module 30 calculates and analyzes the first data by the method of step 3 to obtain the predicted electric quantity.
The control module 40 controls the data acquisition module 10, the processing module 20 and the prediction module 30 to operate.
Effects and effects of the embodiments
According to the electric quantity prediction system and the prediction method thereof, as the electric quantity measurement model is improved and the quantile regression model is introduced, the valuable electric quantity reference interval can be predicted more accurately, so that the interference of data noise is effectively reduced, the characteristics of the electric quantity are further extracted from time and space by combining the advantages of the long-short term memory network and the convolutional neural network, the weight is updated by using the attention mechanism, the characteristic extraction effect is optimized, the probability prediction is realized by using the quantile regression model, and on the basis of ensuring accurate point prediction, a credible electric quantity prediction reference range can be provided for researchers.
The above embodiments are preferred examples of the present invention, and are not intended to limit the scope of the present invention.

Claims (9)

1. A method for predicting an amount of electricity, comprising the steps of:
step 1, acquiring data to be processed from a scheduling center, wherein the data to be processed comprises electric quantity and a first characteristic;
step 2, processing the data to be processed to obtain first data, and the specific substeps comprise:
step 2-1, performing box separation processing on the data to be processed through a first method to obtain box separation processing data;
step 2-2, dividing the box processing data into a first electric quantity and a second characteristic, carrying out correlation analysis on the first electric quantity and the second characteristic through a maximum information coefficient algorithm, and screening target characteristics;
step 2-3, decomposing the first electric quantity into a plurality of subsequences by integrating an empirical mode decomposition algorithm;
2-4, performing correlation analysis on the target characteristics and the subsequences through a maximum information coefficient algorithm, and dividing the subsequences into characteristic correlation and non-characteristic correlation;
step 2-5, performing normalization processing on continuous variables in the first electric quantity, and performing one-hot encoding processing on discrete variables in the first electric quantity to obtain first data;
step 3, calculating and analyzing the first data to obtain the predicted electric quantity, wherein the specific substeps comprise:
step 3-1, learning the time relation of the first data by using a first network to obtain time characteristics;
step 3-2, learning the spatial relationship of the first data by using a second network to obtain spatial characteristics;
3-3, updating the weights of the time characteristic and the space characteristic by using an attention mechanism to obtain a final output matrix;
3-4, performing probability prediction on the subsequence through a first model to obtain a predicted subsequence;
and 3-5, calculating the sum of the prediction subsequence through the evaluation index, and analyzing to obtain the predicted electric quantity.
2. The electricity quantity prediction method according to claim 1, characterized in that:
in step 2-1, the first method is to divide different temperatures into different boxes, and then obtain the temperature range of each box through calculation, and the specific formula is as follows:
Figure FDA0003068653800000021
where d is the range of each bin, μ is the mean of the data, and n is the number of samples.
3. The electricity quantity prediction method according to claim 1, characterized in that:
in step 2-5, the normalization processing formula is as follows:
Figure FDA0003068653800000022
wherein x isnormIs made ofNormalized value, xminIs the minimum value of x, xmaxIs the maximum value of x.
4. The electricity quantity prediction method according to claim 1, characterized in that:
in step 3-1, the first network is a long-term and short-term memory network, and the specific parameters are as follows:
it＝σ(Wi[ht-1，xt]+bi)
ft＝σ(Wf[ht-1，xt]+bf)
ot＝σ(Wo[ht-1，xt]+bo)
c′t＝tanh(Wc[ht-1，xt]+bc)
ct＝fcct-1+itc′t
wherein itIs an input gate, ftIs a forgetting door otIs an output gate; h istIs the hidden layer output at time t; h ist-1Is the hidden layer output at time t-1; c. Ct' is cell status renewal; c. CtIs the cell state output; wi，Wf，Wo，WcThe weights of the input gate, the forgetting gate, the output gate and the cell state are respectively; bi，bf，bo，bcThe deviations of the input gate, the forgetting gate, the output gate and the cell state, respectively.
5. The electricity quantity prediction method according to claim 1, characterized in that:
in step 3-2, the second network is a convolutional neural network, and the sub-step of obtaining the spatial features is as follows:
step 3-2-1, defining the spatial features as a matrix F ═ Ft+1，ft+2，...，ft+n]TWherein f ist+nIs a forgetting gate;
step 3-2-2, inputting the matrix F into the convolutional neural network;
and 3-2-3, extracting the spatial features by using a filter to obtain the spatial features.
6. The electricity quantity prediction method according to claim 1, characterized in that:
in step 3-3, the expression of the final output matrix calculation process is as follows:
f(HT，h)＝HTh
a＝σ(f(HT，h)Wa
V＝HaT
HTis the transpose of the spatial feature, h is the temporal feature, WaIs the weight and V is the final output matrix.
7. The electricity quantity prediction method according to claim 1, characterized in that:
in step 3-4, the first model is a quantile regression model, and the specific formula is as follows:
Figure FDA0003068653800000041
wherein r is the quantile,
Figure FDA0003068653800000042
are the prediction set and y is the true value.
8. The electricity quantity prediction method according to claim 1, characterized in that:
in step 3-5, the evaluation indexes are MAPE, RMSE, Pinball and Winkler.
9. An electric quantity prediction system for predicting electric quantity by using the electric quantity prediction method according to any one of claims 1 to 8, comprising:
a data acquisition module, a processing module, a prediction module and a control module,
the data acquisition module acquires data to be processed from a scheduling center,
the processing module processes the data to be processed to obtain first data,
the prediction module calculates and analyzes the first data to obtain predicted electric quantity,
the control module controls the data acquisition module, the processing module and the prediction module.
CN202110532962.1A 2021-05-17 2021-05-17 Electric quantity prediction system and prediction method thereof Pending CN113205223A (en)

Priority Applications (1)

Application Number Priority Date Filing Date Title
CN202110532962.1A CN113205223A (en) 2021-05-17 2021-05-17 Electric quantity prediction system and prediction method thereof

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
CN202110532962.1A CN113205223A (en) 2021-05-17 2021-05-17 Electric quantity prediction system and prediction method thereof

Publications (1)

Publication Number Publication Date
CN113205223A true CN113205223A (en) 2021-08-03

Family

ID=77031483

Family Applications (1)

Application Number Title Priority Date Filing Date
CN202110532962.1A Pending CN113205223A (en) 2021-05-17 2021-05-17 Electric quantity prediction system and prediction method thereof

Country Status (1)

Country Link
CN (1) CN113205223A (en)

Cited By (3)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN113762225A (en) * 2021-11-09 2021-12-07 博兴兴业精细化工产业发展有限公司 Automatic monitoring alarm system for chemical workshop
CN114118855A (en) * 2021-12-06 2022-03-01 国网江苏省电力有限公司苏州供电分公司 CNN-based method for calculating benchmarking values of line loss rate of transformer area
CN116227543A (en) * 2023-05-09 2023-06-06 南方电网科学研究院有限责任公司 Detection method and detection device for abnormal electricity consumption, electronic device and electronic equipment

Citations (5)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN107292453A (en) * 2017-07-24 2017-10-24 国网江苏省电力公司电力科学研究院 A kind of short-term wind power prediction method based on integrated empirical mode decomposition Yu depth belief network
CN107392363A (en) * 2017-07-12 2017-11-24 河海大学 A kind of CEEMD and random forest short-term wind power prediction method
CN107704953A (en) * 2017-09-20 2018-02-16 河海大学 The short-term wind-electricity power probability density Forecasting Methodology of EWT quantile estimate forests
CN110909919A (en) * 2019-11-07 2020-03-24 哈尔滨工程大学 Photovoltaic power prediction method of depth neural network model with attention mechanism fused
CN112241802A (en) * 2019-07-18 2021-01-19 内蒙古电力(集团)有限责任公司内蒙古电力科学研究院分公司 Interval prediction method for wind power

Patent Citations (5)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN107392363A (en) * 2017-07-12 2017-11-24 河海大学 A kind of CEEMD and random forest short-term wind power prediction method
CN107292453A (en) * 2017-07-24 2017-10-24 国网江苏省电力公司电力科学研究院 A kind of short-term wind power prediction method based on integrated empirical mode decomposition Yu depth belief network
CN107704953A (en) * 2017-09-20 2018-02-16 河海大学 The short-term wind-electricity power probability density Forecasting Methodology of EWT quantile estimate forests
CN112241802A (en) * 2019-07-18 2021-01-19 内蒙古电力(集团)有限责任公司内蒙古电力科学研究院分公司 Interval prediction method for wind power
CN110909919A (en) * 2019-11-07 2020-03-24 哈尔滨工程大学 Photovoltaic power prediction method of depth neural network model with attention mechanism fused

Non-Patent Citations (3)

* Cited by examiner, † Cited by third party
Title
于一潇: "时空特征深度挖掘的风电集群发电功率短期概率预测方法研究", 《中国优秀硕士学位论文全文数据库 工程科技Ⅱ辑》 *
庞昊等: "基于时间卷积网络分位数回归的短期负荷概率密度预测方法", 《电网技术》 *
魏健等: "基于注意力机制的CNN-LSTM 短期电力负荷预测方法", 《华北电力大学学报》 *

Cited By (5)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN113762225A (en) * 2021-11-09 2021-12-07 博兴兴业精细化工产业发展有限公司 Automatic monitoring alarm system for chemical workshop
CN113762225B (en) * 2021-11-09 2022-02-11 博兴兴业精细化工产业发展有限公司 Automatic monitoring alarm system for chemical workshop
CN114118855A (en) * 2021-12-06 2022-03-01 国网江苏省电力有限公司苏州供电分公司 CNN-based method for calculating benchmarking values of line loss rate of transformer area
CN116227543A (en) * 2023-05-09 2023-06-06 南方电网科学研究院有限责任公司 Detection method and detection device for abnormal electricity consumption, electronic device and electronic equipment
CN116227543B (en) * 2023-05-09 2023-08-04 南方电网科学研究院有限责任公司 Detection method and detection device for abnormal electricity consumption, electronic device and electronic equipment

Similar Documents

Publication Publication Date Title
CN113205223A (en) Electric quantity prediction system and prediction method thereof
CN111260136A (en) Building short-term load prediction method based on ARIMA-LSTM combined model
Hassan et al. A hybrid of multiobjective Evolutionary Algorithm and HMM-Fuzzy model for time series prediction
CN113554466B (en) Short-term electricity consumption prediction model construction method, prediction method and device
CN111008726B (en) Class picture conversion method in power load prediction
CN112434848B (en) Nonlinear weighted combination wind power prediction method based on deep belief network
CN111738520A (en) System load prediction method fusing isolated forest and long-short term memory network
CN112734128A (en) 7-day power load peak value prediction method based on optimized RBF
CN113298319B (en) Traffic speed prediction method based on skip map attention gating cycle network
CN112363896A (en) Log anomaly detection system
CN112329990A (en) User power load prediction method based on LSTM-BP neural network
CN112232604A (en) Prediction method for extracting network traffic based on Prophet model
CN116090602A (en) Power load prediction method and system
CN115269758A (en) Passenger-guidance-oriented road network passenger flow state deduction method and system
CN112990587A (en) Method, system, equipment and medium for accurately predicting power consumption of transformer area
CN115358461A (en) Natural gas load prediction method, device, equipment and medium
CN115375205A (en) Method, device and equipment for determining water user portrait
CN116187835A (en) Data-driven-based method and system for estimating theoretical line loss interval of transformer area
CN114596726B (en) Parking berth prediction method based on interpretable space-time attention mechanism
Wang et al. A Transformer-based multi-entity load forecasting method for integrated energy systems
CN116663842A (en) Digital management system and method based on artificial intelligence
CN114118568A (en) Multi-attention-based power consumption prediction method for convolutional neural network
CN114492967A (en) Urban rail station passenger flow prediction method and medium based on CEEMDAN and BLSTM combined model
CN114037138A (en) Subway short-time arrival passenger flow prediction system based on double-layer decomposition and deep learning and implementation method
CN113011512A (en) Traffic generation prediction method and system based on RBF neural network model

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: 20210803