CN109376960A - Load Forecasting based on LSTM neural network - Google Patents
Load Forecasting based on LSTM neural network Download PDFInfo
- Publication number
- CN109376960A CN109376960A CN201811484599.5A CN201811484599A CN109376960A CN 109376960 A CN109376960 A CN 109376960A CN 201811484599 A CN201811484599 A CN 201811484599A CN 109376960 A CN109376960 A CN 109376960A
- Authority
- CN
- China
- Prior art keywords
- neural network
- lstm neural
- training
- load
- network
- 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 45
- 238000012549 training Methods 0.000 claims abstract description 32
- 238000013480 data collection Methods 0.000 claims abstract description 13
- 238000012545 processing Methods 0.000 claims abstract description 8
- 230000006870 function Effects 0.000 claims description 11
- 230000004913 activation Effects 0.000 claims description 4
- ORILYTVJVMAKLC-UHFFFAOYSA-N Adamantane Natural products C1C(C2)CC3CC1CC2C3 ORILYTVJVMAKLC-UHFFFAOYSA-N 0.000 claims description 3
- 238000005457 optimization Methods 0.000 claims description 3
- 230000001131 transforming effect Effects 0.000 claims description 3
- 238000000034 method Methods 0.000 description 8
- 238000003062 neural network model Methods 0.000 description 6
- 239000003245 coal Substances 0.000 description 5
- 230000000694 effects Effects 0.000 description 5
- 230000005611 electricity Effects 0.000 description 5
- 238000012360 testing method Methods 0.000 description 4
- 241001269238 Data Species 0.000 description 2
- 238000005485 electric heating Methods 0.000 description 2
- 238000002474 experimental method Methods 0.000 description 2
- 230000007774 longterm Effects 0.000 description 2
- 230000015654 memory Effects 0.000 description 2
- 210000004218 nerve net Anatomy 0.000 description 2
- 230000008092 positive effect Effects 0.000 description 2
- 238000000611 regression analysis Methods 0.000 description 2
- 230000009286 beneficial effect Effects 0.000 description 1
- 238000006243 chemical reaction Methods 0.000 description 1
- 230000007812 deficiency Effects 0.000 description 1
- 238000001514 detection method Methods 0.000 description 1
- 238000010586 diagram Methods 0.000 description 1
- 238000005516 engineering process Methods 0.000 description 1
- 238000011156 evaluation Methods 0.000 description 1
- 238000010438 heat treatment Methods 0.000 description 1
- 230000007787 long-term memory Effects 0.000 description 1
- 238000010801 machine learning Methods 0.000 description 1
- 238000012544 monitoring process Methods 0.000 description 1
- 238000004088 simulation Methods 0.000 description 1
- 230000002269 spontaneous effect Effects 0.000 description 1
- 238000000714 time series forecasting Methods 0.000 description 1
- 230000009466 transformation Effects 0.000 description 1
- 238000012795 verification Methods 0.000 description 1
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06Q—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
- G06Q10/00—Administration; Management
- G06Q10/04—Forecasting or optimisation specially adapted for administrative or management purposes, e.g. linear programming or "cutting stock problem"
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- 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
-
- Y—GENERAL TAGGING OF NEW TECHNOLOGICAL DEVELOPMENTS; GENERAL TAGGING OF CROSS-SECTIONAL TECHNOLOGIES SPANNING OVER SEVERAL SECTIONS OF THE IPC; TECHNICAL SUBJECTS COVERED BY FORMER USPC CROSS-REFERENCE ART COLLECTIONS [XRACs] AND DIGESTS
- Y04—INFORMATION OR COMMUNICATION TECHNOLOGIES HAVING AN IMPACT ON OTHER TECHNOLOGY AREAS
- Y04S—SYSTEMS INTEGRATING TECHNOLOGIES RELATED TO POWER NETWORK OPERATION, COMMUNICATION OR INFORMATION TECHNOLOGIES FOR IMPROVING THE ELECTRICAL POWER GENERATION, TRANSMISSION, DISTRIBUTION, MANAGEMENT OR USAGE, i.e. SMART GRIDS
- Y04S10/00—Systems supporting electrical power generation, transmission or distribution
- Y04S10/50—Systems or methods supporting the power network operation or management, involving a certain degree of interaction with the load-side end user applications
Landscapes
- Business, Economics & Management (AREA)
- Engineering & Computer Science (AREA)
- Economics (AREA)
- Human Resources & Organizations (AREA)
- Strategic Management (AREA)
- Theoretical Computer Science (AREA)
- Physics & Mathematics (AREA)
- Health & Medical Sciences (AREA)
- Marketing (AREA)
- General Physics & Mathematics (AREA)
- General Business, Economics & Management (AREA)
- Tourism & Hospitality (AREA)
- Public Health (AREA)
- General Health & Medical Sciences (AREA)
- Primary Health Care (AREA)
- Water Supply & Treatment (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 present invention provides a kind of Load Forecasting based on LSTM neural network, and precision of prediction is high, and implementation result is good, and the power grid prediction that can satisfy under high load capacity requires.It includes the following steps, step 1, historical data acquisition and processing；Using each website as classification foundation, the network load value in prediction area is temporally arranged to obtain sequence data collection；Step 2, processing data obtain the sample of supervised learning；Sequence data collection is standardized, it is between [- 1,1], obtains sample of the training dataset as supervised learning；Step 3, the modeling training of LSTM neural network；Step 4, it according to historical data, and completes trained LSTM neural network and the load value of future time is predicted.
Description
Technical field
The present invention relates to network load predictions, the specially Load Forecasting based on LSTM neural network.
Background technique
In recent years in order to reduce winter pollution that caused by coal burning, improve air quality, many cities of northern China start popularization, and " coal changes
Electricity " policy.The main implementation method of " coal changes electricity " technology is to be heated using air source heat pump as winter, can be big after the completion of transformation
It is big to reduce coal-fired use but growing day by day to the pressure of power grid simultaneously.Since " coal changes electricity ", network system carry than with
Toward bigger pressure, the incidence of accident increases, especially winter heating when, due in advance to the prediction of load not
Standard estimates deficiency to imminent accident, accident is caused to cause very big influence.
The mainly linear fit or regression analysis model etc. that traditional network load prediction uses, however actual electric power
Load model be it is nonlinear, the load of power grid will receive the interference of the various influence factors such as temperature, humidity, and previous accident
Early warning, emergency are predicted that do not consider historical data, early warning effect is bad, so that traditional according only to future weather data
The precision of prediction of Nonlinear Prediction Models is unable to satisfy the required precision of modern power network management system.
Summary of the invention
Aiming at the problems existing in the prior art, the present invention provides a kind of network load prediction based on LSTM neural network
Method, precision of prediction is high, and implementation result is good, and the power grid prediction that can satisfy under high load capacity requires.
The present invention is to be achieved through the following technical solutions:
Based on the Load Forecasting of LSTM neural network, include the following steps,
Step 1, historical data acquisition and processing；
Using each website as classification foundation, the network load value in prediction area is temporally arranged to obtain sequence data
Collection；
Step 2, processing data obtain the sample of supervised learning；
Sequence data collection is standardized, is in it between [- 1,1], obtains training dataset as supervision
The sample of study；
Step 3, the modeling training of LSTM neural network；
The load value var (t-n) and var (t) of the current time t and time in the past (t-n, t-1) that are concentrated with sequence data
As the list entries X of LSTM neural network prediction model, to predict the future time (t+1, t+n) as output sequence Y
Load value var (t+n), prediction step n are positive integer；
Using Keras as modeling environment, model simultaneously learning training to LSTM neural network；When training, LSTM nerve net
The network layer of network maintains state between the data of fixed line number；
LSTM neural network is compiled using mean_squared_error loss function, is completed by ADAM optimization algorithm
The training of LSTM neural network；
Step 4, it according to historical data, and completes trained LSTM neural network and the load value of future time is carried out in advance
It surveys.
Preferably, the data that load value is 0 are rejected in the sequence data collection in step 1.
Preferably, step 2, it is in it between [- 1,1] using MinMaxScaler transforming sequence data set.
Preferably, step 3, the fixed line number is the training that LSTM neural network is run before updating network weight
Number of data lines in data set.
It preferably, step 3, take hyperbolic tangent function as the activation primitive of LSTM neural network when training.
Preferably, step 3, when training, the state of LSTM neural network network layer is determined using reset_states function
It is emptied the time.
Compared with prior art, the invention has the following beneficial technical effects:
The present invention is based on the Load Forecastings of LSTM neural network, in conjunction with machine learning algorithm and big data skill
Art, by the network load prediction model based on LSTM time series, by same monitoring point different time historical data into
Row training, the simulation mankind predict the mode of accident early warning the network load numerical value at the following a certain moment.Experiments verify that in reality
The load capacity of power grid can be reacted in time and accurately in the application on border, prediction to network load and corresponding met an urgent need
Very big positive effect is arrived.
Detailed description of the invention
Fig. 1 is the flow diagram of the method for the invention.
Fig. 2 is the load chart of acquisition time section historical data described in present example.
Fig. 3 is the models fitting effect under 1500epoch.
Specific embodiment
Below with reference to specific embodiment, the present invention is described in further detail, it is described be explanation of the invention and
It is not to limit.
In order to enable those skilled in the art to better understand the solution of the present invention, below in conjunction in the embodiment of the present invention
Attached drawing, technical scheme in the embodiment of the invention is clearly and completely described, it is clear that described embodiment is only
The embodiment of a part of the invention, instead of all the embodiments.Based on the embodiments of the present invention, ordinary skill people
The model that the present invention protects all should belong in member's every other embodiment obtained without making creative work
It encloses.
It should be noted that description and claims of this specification and term " first " in above-mentioned attached drawing, "
Two " etc. be to be used to distinguish similar objects, without being used to describe a particular order or precedence order.It should be understood that using in this way
Data be interchangeable under appropriate circumstances, so as to the embodiment of the present invention described herein can in addition to illustrating herein or
Sequence other than those of description is implemented.In addition, term " includes " and " having " and their any deformation, it is intended that cover
Cover it is non-exclusive include, for example, the process, method, system, product or equipment for containing a series of steps or units are not necessarily limited to
Step or unit those of is clearly listed, but may include be not clearly listed or for these process, methods, product
Or other step or units that equipment is intrinsic.
The present invention is based on the Load Forecastings of LSTM neural network, as shown in Figure 1, it includes the following steps,
Step 1, historical data acquisition and processing；
Using each website as classification foundation, the network load value in prediction area is temporally arranged to obtain sequence data
Collection, and reject the time that wherein load value is 0；
Since the number of days in some months in data acquisition is different, it may appear that this paper that some load values are 0, in number
Directly cast out during Data preprocess；Other data were used as the 1st day according to November 1, until second year 2 months 28
Day, for all data summarizations to load chart as shown in Fig. 2, number of days is abscissa, network load value is ordinate.
It can be found that network load is difficult to find determining linear relationship from Fig. 2, but on the whole, 40-80 days left
Right load value be it is highest, illustrate electric heating utilization rate this period be it is highest, be consistent with really comparing.
Step 2, processing data obtain the sample of supervised learning；
Sequence data collection is standardized, using MinMaxScaler transforming sequence data set make its be in [- 1,
1] between, as training dataset.
Data obtained in step 1 are the sequence data collections arranged sequentially in time, but as supervised learning
Sample, need to be created that the list entries X and label y for predicting.For time series forecasting, current time is t, future
(t+1, t+n), past observation (t-n, t-1) are used to predict.Herein using prediction modeling, i.e. input data is var (t-1)
With var (t), var (t+1) variable is predicted, by taking Single-step Prediction as an example, then sequence data collection can be treated to be lattice as shown in table 1 below
Formula.
Data set in 1 Single-step Prediction of table
| var1(t-1) | var(t) | |
| 0 | 0 | 176.67 |
| 1 | 176.67 | 164.36 |
| 2 | 164.36 | 171.39 |
| 3 | 171.39 | 170.51 |
| 4 | 170.51 | 163.48 |
| 5 | 163.48 | 165.24 |
Since the activation primitive of LSTM default is hyperbolic tangent function (tanh), the output valve of this function is in -1 and 1
Between, therefore also need to be standardized all data.
This is influence of the fairness by test data set information in order to avoid the experiment, and model may be made to predict
When be in a disadvantageous position.It is in it between [- 1,1] using MinMaxScaler conversion data collection herein, then after being standardized
Data set is as shown in table 2.
Data set after table 2 standardizes
| var1(t-1) | var(t) | |
| 0 | -1 | -0.3133 |
| 1 | 0.640847 | -0.73136 |
| 2 | 0.526516 | -0.49261 |
| 3 | 0.591808 | -0.5225 |
| 4 | 0.583635 | -0.76125 |
| 5 | 0.518343 | -0.70148 |
Step 3, the modeling training of LSTM neural network；
The load value var (t-n) and var (t) of the current time t and time in the past (t-n, t-1) that are concentrated with sequence data
As the list entries X of prediction model, to predict the load value var (t+ of the future time (t+1, t+n) as output sequence Y
N), prediction step n is positive integer；
It take hyperbolic tangent function as the activation primitive of LSTM neural network
Using Keras as modeling environment, model simultaneously learning training to LSTM neural network；When training, LSTM nerve net
The network layer of network maintains state between the data of fixed line number, and the fixed line number is the LSTM before updating network weight
The number of data lines that the training data of neural network operation is concentrated.
Determine that the state of LSTM neural network network layer is emptied the time using reset_states function,
LSTM neural network is compiled using mean_squared_error loss function, is completed by ADAM optimization algorithm
The training of LSTM neural network.
Shot and long term memory network (LSTM) is a kind of special RNN, can learn and remember longer sequence, be not relying on pre-
First specified window lag observed value is as input；T-th of value, Ke Yiji can be predicted by t-1 value before a sequence
Pervious information is recalled to understand current content, solves the problems, such as that the gradient easily occurred in RNN is withered away.
Step 4, it according to historical data, and completes trained LSTM neural network and the load value of future time is carried out in advance
It surveys.
The present invention is finally by network load number true after acquisition Changping District, Beijing 2015-2017 " coal changes electricity "
According to experimental verification is carried out, show that the model can react the load capacity of power grid in time and accurately in actual application,
Very big positive effect is played to accident early warning, emergency.
The training data sample number used for preceding 90 network load value, test sample using 28 days 2 months electricity
Net load value.Final evaluation index is the mean square deviation of predicted value and actual value.Mean square deviation can be with for mean error
More data patterns not described by model, such as periodicity are detected except linear trend, for this paper the problem of is pre-
Network load is surveyed, itself is there is also certain periodicity, for example winter increases with the utilization rate of the weather is growing cold electric heating, power grid
Load can increase with it, and serve as the coldest time, and the load of power grid can slowly be restored to reduced levels.
Regression analysis, the fitting of test data are carried out in test data set using the LSTM neural network model after training
Effect is as shown, abscissa is number of days, and ordinate is network load value, and blue is true network load value, and orange is pre-
The network load value of survey.
When training number is 1500epoch, fitting effect is as shown in figure 3, of the present invention be based on LSTM neural network
Load Forecasting, the specific load value predicted and true value have a gap, but the entirety of predicted value curve a
The variation tendency raised and reduced has been able to accurately reflect really load curve b very much, in order to obtain closer to true
Real prediction data can be by LSTM neural network model that increase the quantity training of epoch more complicated.
Model accuracy is improved by improving the complexity of training pattern, and true network load value is compared, it can
Utilize LSTM neural network model to find out, on the basis of 90 days training datas of history, can to the 28 following day datas into
Row accurately prediction.
It is as shown in table 3 with multinomial model comparing result in the prior art:
3 LSTM neural network model of table and multinomial model comparing result
| RMSE | |
| LSTM model 50epoch | 5.38 |
| LSTM model 1500epoch | 7.066 |
| Fitting of a polynomial | 188.99 |
Simultaneously in traditional neural network model, connected entirely to output layer again from input layer to hidden layer, every layer
Between node be connectionless.But this network is helpless to the prediction of sequence data, therefore circulation occurs
Neural network (RNN) can predict t-th of value by t-1 value before a sequence, can remember pervious information to understand
Current content.Long Memory Neural Networks LSTM (LongShort-TermMemoryNeuralNetwork) in short-term is a kind of RNN
Specific type, the long-term information of memory that can be spontaneous, rather than go to learn.In the present invention, the power grid at a certain moment
Load can or can not overload detection often closely related with the network load value before this moment, artificial generally by going through
History data carry out early warning come the network load that future may occur, while in order to remember long-term information, final to use
LSTM neural network model simulates the mankind to the mode of accident forecast, passes through true history network load data training mould
Type, to predict the network load at the following a certain moment.
Claims (6)
1. the Load Forecasting based on LSTM neural network, which is characterized in that include the following steps,
Step 1, historical data acquisition and processing；
Using each website as classification foundation, the network load value in prediction area is temporally arranged to obtain sequence data collection；
Step 2, processing data obtain the sample of supervised learning；
Sequence data collection is standardized, it is between [- 1,1], obtains training dataset as supervised learning
Sample；
Step 3, the modeling training of LSTM neural network；
With load value var (t-n) and var (t) conduct of current time t and time in the past (t-n, t-1) that sequence data is concentrated
The list entries X of LSTM neural network prediction model, to predict the load of the future time (t+1, t+n) as output sequence Y
Value var (t+n), prediction step n are positive integer；
Using Keras as modeling environment, model simultaneously learning training to LSTM neural network；When training, LSTM neural network
Network layer maintains state between the data of fixed line number；
LSTM neural network is compiled using mean_squared_error loss function, LSTM mind is completed by ADAM optimization algorithm
Training through network；
Step 4, it according to historical data, and completes trained LSTM neural network and the load value of future time is predicted.
2. the Load Forecasting according to claim 1 based on LSTM neural network, which is characterized in that step 1
In sequence data collection in reject load value be 0 data.
3. the Load Forecasting according to claim 1 based on LSTM neural network, which is characterized in that step 2,
It is between [- 1,1] using MinMaxScaler transforming sequence data set.
4. the Load Forecasting according to claim 1 based on LSTM neural network, which is characterized in that step 3,
The fixed line number is the number of data lines that the training data that LSTM neural network is run before updating network weight is concentrated.
5. the Load Forecasting according to claim 1 based on LSTM neural network, which is characterized in that step 3,
It take hyperbolic tangent function as the activation primitive of LSTM neural network when training.
6. the Load Forecasting according to claim 1 based on LSTM neural network, which is characterized in that step 3,
When training, determine that the state of LSTM neural network network layer is emptied the time using reset_states function.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201811484599.5A CN109376960A (en) | 2018-12-06 | 2018-12-06 | Load Forecasting based on LSTM neural network |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201811484599.5A CN109376960A (en) | 2018-12-06 | 2018-12-06 | Load Forecasting based on LSTM neural network |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN109376960A true CN109376960A (en) | 2019-02-22 |
Family
ID=65375808
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201811484599.5A Pending CN109376960A (en) | 2018-12-06 | 2018-12-06 | Load Forecasting based on LSTM neural network |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN109376960A (en) |
Cited By (20)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN109742777A (en) * | 2019-03-04 | 2019-05-10 | 曹麾 | Low-voltage platform area peak load balance intelligence control subsystem |
| CN109842141A (en) * | 2019-03-04 | 2019-06-04 | 曹麾 | Low-voltage platform area peak load balances intelligent management |
| CN109842140A (en) * | 2019-03-04 | 2019-06-04 | 曹麾 | High-voltage distribution network peak load balances intelligent management-control method |
| CN109904865A (en) * | 2019-03-04 | 2019-06-18 | 曹麾 | High-voltage distribution network peak load balance intelligence control main system |
| CN109993359A (en) * | 2019-03-26 | 2019-07-09 | 华南理工大学 | A kind of Tendency Prediction method based on sophisticated machine study |
| CN110135643A (en) * | 2019-05-17 | 2019-08-16 | 国网山东省电力公司莱芜供电公司 | Consider the Short-term Load Forecast method of steel forward price and Spot Price factor |
| CN110222897A (en) * | 2019-06-11 | 2019-09-10 | 国网上海市电力公司 | A kind of distribution network reliability analysis method |
| CN110263866A (en) * | 2019-06-24 | 2019-09-20 | 苏州智睿新能信息科技有限公司 | A kind of power consumer load setting prediction technique based on deep learning |
| CN110334879A (en) * | 2019-07-11 | 2019-10-15 | 华北电力大学 | Power grid bus reactive load forecasting method and device |
| CN110413406A (en) * | 2019-06-27 | 2019-11-05 | 莫毓昌 | A kind of task load forecasting system and method |
| CN110703101A (en) * | 2019-09-12 | 2020-01-17 | 北京交通大学 | Lithium ion battery inter-partition cycle capacity decline prediction method |
| CN110889491A (en) * | 2019-11-21 | 2020-03-17 | 福建工程学院 | Weather factor-based power load prediction method and prediction system |
| CN110969296A (en) * | 2019-11-25 | 2020-04-07 | 国网河北省电力有限公司经济技术研究院 | Electric heating load prediction method and device and terminal equipment |
| CN111401667A (en) * | 2020-06-03 | 2020-07-10 | 广东电网有限责任公司东莞供电局 | Power utilization scheduling method and device for factory, computer equipment and storage medium |
| CN112036542A (en) * | 2019-06-04 | 2020-12-04 | 山东华软金盾软件股份有限公司 | CPU occupancy prediction method and system based on deep learning |
| CN112418496A (en) * | 2020-11-10 | 2021-02-26 | 国网四川省电力公司经济技术研究院 | Power distribution station energy storage configuration method based on deep learning |
| CN112734106A (en) * | 2021-01-08 | 2021-04-30 | 深圳市国电科技通信有限公司 | Method and device for predicting energy load |
| CN113726559A (en) * | 2021-08-09 | 2021-11-30 | 国网福建省电力有限公司 | Artificial intelligence network-based security analysis early warning model |
| CN115361318A (en) * | 2022-07-20 | 2022-11-18 | 贵州大学 | LSTM edge calculation flow prediction method for dynamic load balancing in complex environment |
| CN115526300A (en) * | 2022-11-14 | 2022-12-27 | 南京邮电大学 | Sequence rearrangement method based on cyclic neural network |
Citations (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN106960252A (en) * | 2017-03-08 | 2017-07-18 | 深圳市景程信息科技有限公司 | Methods of electric load forecasting based on long Memory Neural Networks in short-term |
-
2018
- 2018-12-06 CN CN201811484599.5A patent/CN109376960A/en active Pending
Patent Citations (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN106960252A (en) * | 2017-03-08 | 2017-07-18 | 深圳市景程信息科技有限公司 | Methods of electric load forecasting based on long Memory Neural Networks in short-term |
Cited By (29)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN109842141A (en) * | 2019-03-04 | 2019-06-04 | 曹麾 | Low-voltage platform area peak load balances intelligent management |
| CN109842140A (en) * | 2019-03-04 | 2019-06-04 | 曹麾 | High-voltage distribution network peak load balances intelligent management-control method |
| CN109904865A (en) * | 2019-03-04 | 2019-06-18 | 曹麾 | High-voltage distribution network peak load balance intelligence control main system |
| CN109742777A (en) * | 2019-03-04 | 2019-05-10 | 曹麾 | Low-voltage platform area peak load balance intelligence control subsystem |
| CN109904865B (en) * | 2019-03-04 | 2024-04-26 | 曹麾 | Intelligent peak-valley load balance management and control main system of high-voltage distribution network |
| CN109842140B (en) * | 2019-03-04 | 2023-01-17 | 曹麾 | Intelligent management and control method for peak-valley load balance of high-voltage distribution network |
| CN109993359A (en) * | 2019-03-26 | 2019-07-09 | 华南理工大学 | A kind of Tendency Prediction method based on sophisticated machine study |
| CN110135643A (en) * | 2019-05-17 | 2019-08-16 | 国网山东省电力公司莱芜供电公司 | Consider the Short-term Load Forecast method of steel forward price and Spot Price factor |
| CN112036542A (en) * | 2019-06-04 | 2020-12-04 | 山东华软金盾软件股份有限公司 | CPU occupancy prediction method and system based on deep learning |
| CN110222897A (en) * | 2019-06-11 | 2019-09-10 | 国网上海市电力公司 | A kind of distribution network reliability analysis method |
| CN110263866A (en) * | 2019-06-24 | 2019-09-20 | 苏州智睿新能信息科技有限公司 | A kind of power consumer load setting prediction technique based on deep learning |
| CN110263866B (en) * | 2019-06-24 | 2023-11-10 | 苏州智睿新能信息科技有限公司 | Power consumer load interval prediction method based on deep learning |
| CN110413406A (en) * | 2019-06-27 | 2019-11-05 | 莫毓昌 | A kind of task load forecasting system and method |
| CN110334879A (en) * | 2019-07-11 | 2019-10-15 | 华北电力大学 | Power grid bus reactive load forecasting method and device |
| CN110703101A (en) * | 2019-09-12 | 2020-01-17 | 北京交通大学 | Lithium ion battery inter-partition cycle capacity decline prediction method |
| CN110703101B (en) * | 2019-09-12 | 2021-01-05 | 北京交通大学 | Lithium ion battery inter-partition cycle capacity decline prediction method |
| CN110889491A (en) * | 2019-11-21 | 2020-03-17 | 福建工程学院 | Weather factor-based power load prediction method and prediction system |
| CN110969296A (en) * | 2019-11-25 | 2020-04-07 | 国网河北省电力有限公司经济技术研究院 | Electric heating load prediction method and device and terminal equipment |
| CN110969296B (en) * | 2019-11-25 | 2023-11-07 | 国网河北省电力有限公司经济技术研究院 | Electric heating load prediction method and device and terminal equipment |
| CN111401667A (en) * | 2020-06-03 | 2020-07-10 | 广东电网有限责任公司东莞供电局 | Power utilization scheduling method and device for factory, computer equipment and storage medium |
| CN112418496B (en) * | 2020-11-10 | 2022-04-22 | 国网四川省电力公司经济技术研究院 | Power distribution station energy storage configuration method based on deep learning |
| CN112418496A (en) * | 2020-11-10 | 2021-02-26 | 国网四川省电力公司经济技术研究院 | Power distribution station energy storage configuration method based on deep learning |
| CN112734106A (en) * | 2021-01-08 | 2021-04-30 | 深圳市国电科技通信有限公司 | Method and device for predicting energy load |
| CN113726559B (en) * | 2021-08-09 | 2023-10-27 | 国网福建省电力有限公司 | Based on artificial intelligence network safety analysis early warning system |
| CN113726559A (en) * | 2021-08-09 | 2021-11-30 | 国网福建省电力有限公司 | Artificial intelligence network-based security analysis early warning model |
| CN115361318A (en) * | 2022-07-20 | 2022-11-18 | 贵州大学 | LSTM edge calculation flow prediction method for dynamic load balancing in complex environment |
| CN115361318B (en) * | 2022-07-20 | 2023-06-09 | 贵州大学 | LSTM edge calculation flow prediction method for dynamic load balancing in complex environment |
| CN115526300A (en) * | 2022-11-14 | 2022-12-27 | 南京邮电大学 | Sequence rearrangement method based on cyclic neural network |
| CN115526300B (en) * | 2022-11-14 | 2023-06-02 | 南京邮电大学 | Sequence rearrangement method based on cyclic neural network |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN109376960A (en) | Load Forecasting based on LSTM neural network | |
| CN113962364B (en) | Multi-factor power load prediction method based on deep learning | |
| CN110046743B (en) | Public building energy consumption prediction method and system based on GA-ANN | |
| CN109659933A (en) | A kind of prediction technique of power quality containing distributed power distribution network based on deep learning model | |
| CN113205203A (en) | CNN-LSTM-based building energy consumption prediction method and system | |
| CN102183621A (en) | Aquaculture dissolved oxygen concentration online forecasting method and system | |
| CN112465256A (en) | Building power consumption prediction method and system based on Stacking model fusion | |
| CN113554466A (en) | Short-term power consumption prediction model construction method, prediction method and device | |
| CN113516271A (en) | Wind power cluster power day-ahead prediction method based on space-time neural network | |
| CN116796168A (en) | CNN-BiLSTM high-altitude multi-factor power transmission line audible noise prediction method based on multi-head attention mechanism | |
| CN113469266A (en) | Electricity stealing behavior detection method based on improved deep convolutional neural network | |
| CN112633556A (en) | Short-term power load prediction method based on hybrid model | |
| CN112183877A (en) | Photovoltaic power station fault intelligent diagnosis method based on transfer learning | |
| CN115859099A (en) | Sample generation method and device, electronic equipment and storage medium | |
| Gao et al. | A multifactorial framework for short-term load forecasting system as well as the jinan’s case study | |
| CN115860286A (en) | Air quality prediction method and system based on time sequence door mechanism | |
| CN114611757A (en) | Electric power system short-term load prediction method based on genetic algorithm and improved depth residual error network | |
| Ibrahim et al. | LSTM neural network model for ultra-short-term distribution zone substation peak demand prediction | |
| Guo et al. | Short-Term Water Demand Forecast Based on Deep Neural Network:(029) | |
| KR101057430B1 (en) | How to predict the heat demand of district heating | |
| CN117033923A (en) | Method and system for predicting crime quantity based on interpretable machine learning | |
| CN115481788B (en) | Phase change energy storage system load prediction method and system | |
| CN111614489A (en) | Transient recording fault data cleaning method and system based on sparse self-coding | |
| CN113723670B (en) | Photovoltaic power generation power short-term prediction method with variable time window | |
| CN115907131A (en) | Method and system for building electric heating load prediction model in northern area |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| RJ01 | Rejection of invention patent application after publication |
Application publication date: 20190222 |
|
| RJ01 | Rejection of invention patent application after publication |