CN113902207A - Short-term load prediction method based on TCN-LSTM - Google Patents
Short-term load prediction method based on TCN-LSTM Download PDFInfo
- Publication number
- CN113902207A CN113902207A CN202111215767.2A CN202111215767A CN113902207A CN 113902207 A CN113902207 A CN 113902207A CN 202111215767 A CN202111215767 A CN 202111215767A CN 113902207 A CN113902207 A CN 113902207A
- Authority
- CN
- China
- Prior art keywords
- data
- lstm
- time
- tcn
- input
- 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.)
- Withdrawn
Links
- 238000000034 method Methods 0.000 title claims abstract description 32
- 238000012549 training Methods 0.000 claims abstract description 13
- 238000003062 neural network model Methods 0.000 claims abstract description 8
- 238000012545 processing Methods 0.000 claims abstract description 6
- 238000005070 sampling Methods 0.000 claims abstract description 5
- 230000002159 abnormal effect Effects 0.000 claims abstract description 4
- 238000007781 pre-processing Methods 0.000 claims abstract description 4
- 230000004913 activation Effects 0.000 claims description 8
- 238000013507 mapping Methods 0.000 claims description 8
- 230000005611 electricity Effects 0.000 claims description 7
- 210000002569 neuron Anatomy 0.000 claims description 7
- 238000010606 normalization Methods 0.000 claims description 4
- 230000009467 reduction Effects 0.000 claims description 4
- ORILYTVJVMAKLC-UHFFFAOYSA-N Adamantane Natural products C1C(C2)CC3CC1CC2C3 ORILYTVJVMAKLC-UHFFFAOYSA-N 0.000 claims description 3
- 230000008859 change Effects 0.000 claims description 3
- 230000006870 function Effects 0.000 description 11
- 238000013528 artificial neural network Methods 0.000 description 8
- 238000013527 convolutional neural network Methods 0.000 description 3
- 239000000284 extract Substances 0.000 description 3
- 238000000605 extraction Methods 0.000 description 3
- 230000008569 process Effects 0.000 description 3
- 230000000306 recurrent effect Effects 0.000 description 3
- 238000007619 statistical method Methods 0.000 description 3
- 230000001364 causal effect Effects 0.000 description 2
- 230000008034 disappearance Effects 0.000 description 2
- 230000007774 longterm Effects 0.000 description 2
- 238000010801 machine learning Methods 0.000 description 2
- 238000012706 support-vector machine Methods 0.000 description 2
- 230000009286 beneficial effect Effects 0.000 description 1
- 230000006378 damage Effects 0.000 description 1
- 230000007850 degeneration Effects 0.000 description 1
- 238000010586 diagram Methods 0.000 description 1
- 230000000694 effects Effects 0.000 description 1
- 238000004880 explosion Methods 0.000 description 1
- 238000012417 linear regression Methods 0.000 description 1
- 238000013178 mathematical model Methods 0.000 description 1
- 238000005065 mining Methods 0.000 description 1
- 238000012986 modification Methods 0.000 description 1
- 230000004048 modification Effects 0.000 description 1
- 238000011176 pooling Methods 0.000 description 1
- 238000010248 power generation Methods 0.000 description 1
- 230000007704 transition Effects 0.000 description 1
- XLYOFNOQVPJJNP-UHFFFAOYSA-N water Substances O XLYOFNOQVPJJNP-UHFFFAOYSA-N 0.000 description 1
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
- 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
-
- 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
- Engineering & Computer Science (AREA)
- Theoretical Computer Science (AREA)
- Physics & Mathematics (AREA)
- Business, Economics & Management (AREA)
- Health & Medical Sciences (AREA)
- General Physics & Mathematics (AREA)
- General Health & Medical Sciences (AREA)
- Economics (AREA)
- Software Systems (AREA)
- Biomedical Technology (AREA)
- Data Mining & Analysis (AREA)
- Molecular Biology (AREA)
- Computing Systems (AREA)
- General Engineering & Computer Science (AREA)
- Computational Linguistics (AREA)
- Mathematical Physics (AREA)
- Biophysics (AREA)
- Strategic Management (AREA)
- Artificial Intelligence (AREA)
- Human Resources & Organizations (AREA)
- Life Sciences & Earth Sciences (AREA)
- Evolutionary Computation (AREA)
- General Business, Economics & Management (AREA)
- Marketing (AREA)
- Tourism & Hospitality (AREA)
- Game Theory and Decision Science (AREA)
- Operations Research (AREA)
- Quality & Reliability (AREA)
- Entrepreneurship & Innovation (AREA)
- Development Economics (AREA)
- Public Health (AREA)
- Water Supply & Treatment (AREA)
- Primary Health Care (AREA)
- Supply And Distribution Of Alternating Current (AREA)
- Management, Administration, Business Operations System, And Electronic Commerce (AREA)
Abstract
The invention discloses a short-term load prediction method based on TCN-LSTM, which comprises the following steps: acquiring sample data; preprocessing sample data, eliminating abnormal data, filling incomplete data, and performing standardized processing; training a neural network model; the method specifically comprises the following steps: taking power load data and continuous non-time sequence data as input, taking power preset loads of each sampling point on a prediction day as output, and training a neural network model combining a time convolution network and an LSTM to obtain a short-term load prediction model based on TCN-LSTM; the method provides a more accurate prediction result for the power grid system and also provides a more reliable basis for the power grid system to flexibly adjust the power supply quantity.
Description
Technical Field
The invention relates to the technical field of load prediction of a power system, in particular to a short-term load prediction method based on TCN-LSTM.
Background
The load prediction of the power system is the basis for planning and stable, safe and economic operation of the power system. The load prediction can be classified into long-term prediction, medium-term prediction, short-term prediction, and ultra-short-term prediction according to the prediction duration. Different prediction types have different application purposes to the power grid. The Short Term Load Forecasting (STLF) generally refers to load forecasting from 1 hour to 1 week later from the current moment, and is suitable for thermal power distribution, water and fire coordination and the like. The reliable prediction result is beneficial to improving the utilization rate of the power generation equipment and reducing the operation cost of the power network. With the advance of market reformation of the power grid, the influence of effective short-term load prediction on the real-time electricity price is more obvious. However, as the size of the power grid is continuously enlarged and the load diversity is increased, efficient and accurate short-term load prediction becomes more difficult. This requires that the ultra-short term load prediction method be fast and accurate.
At present, methods for load prediction can be mainly divided into two major categories, namely, traditional statistical methods and emerging machine learning methods. The statistical method comprises a multiple linear regression model, a Kalman filter model, a time sequence model and the like, the established model has a relatively definite mathematical form, and the quality of a prediction result is determined by the assumption of data distribution and the rationality of the model. Due to the characteristics of complexity and nonlinearity of the power load, it is difficult to make distribution assumptions which are more in line with reality and establish a clear mathematical model. Most statistical methods are not ideal for short term load prediction. The machine learning method includes Fuzzy Inference System (FIS), Artificial Neural Network (ANN), Support Vector Machine (SVM), and the like. The methods can better process the nonlinear problem, so the prediction accuracy is improved, but other problems exist, such as the lack of self-learning capability, the inability to process large-scale data, the destruction of the time sequence characteristics of the data, manual characteristic selection and the like.
The RNN model is provided with a memory unit, can extract the time sequence relation of power load data, but is easy to generate the problem of gradient disappearance or explosion in the training process. A Long-short time memory (LSTM) network is an improved recurrent neural network that can simultaneously learn the non-linear and time-series characteristics of data and solve the problem of disappearance of the ordinary RNN gradient. CNN networks can effectively improve feature extraction on data by pooling with convolution. However, in the presence of a single CNN network and an LSTM network, in order to input data for a long time in sequence or in a multidimensional manner, the problems of loss of sequence characteristic information, disorder of structural information among data and insufficient multidimensional characteristic mining still exist. The combined prediction model predicts the short-term power load by combining various models and methods, and better meets the actual demand of short-term power load prediction by combining the characteristics and advantages of different models, and the prediction precision of the combined model is higher than that of a single model under general conditions.
Disclosure of Invention
1. The technical problem to be solved is as follows:
aiming at the technical problems, the invention provides a short-term load prediction method based on TCN-LSTM, which builds a TCN-GRU combined prediction model by utilizing the characteristic extraction capability of a time convolution network on time series data and the nonlinear fitting capability of an LSTM neural network.
2. The technical scheme is as follows:
a short-term load prediction method based on TCN-LSTM is characterized in that: the method comprises the following steps:
the method comprises the following steps: acquiring historical power load data of a plurality of acquisition points in an area to be predicted within preset time as sample data; the recording interval of the data is in the order of minutes, including between 1 and 60 minutes; the time span of the historical load data is K times of the period in which the repeated change of the load can be observed, and K is an integer greater than or equal to 1;
step two: preprocessing sample data, eliminating abnormal data, filling incomplete data, and performing standardized processing;
step three: training a neural network model; the method specifically comprises the following steps: taking power load data and continuous non-time sequence data as input, taking power preset loads of each sampling point on a prediction day as output, and training a neural network model combining a time convolution network and an LSTM to obtain a short-term load prediction model based on TCN-LSTM; the non-time sequence data comprises date, week, daily average temperature and daily average electricity price;
the short-term load prediction model based on the TCN-LSTM comprises an input 1, a time convolution network layer, an input 2 and an LSTM network layer; the input 1 inputs a time convolution network layer, and the output of the time convolution network layer is combined with the input 2 and then input
The time convolution network layer comprises 2 layers of residual error units, and each layer of residual error unit comprises 2 convolution units and 1 nonlinear mapping; the convolution unit uses a ReLU function as an activation function and performs normalization operation on the weight of a convolution kernel; the size of the convolution kernel is 2, a Dropout coefficient with the coefficient of 0.4 is set, and Dropout setting can randomly select part of neurons to be inactivated, so that over-fitting training is prevented, and the convergence speed of the model is increased; setting the expansion coefficient to be (1, 2, 4, 8, 16, 32); the filter is 128; the input and output of the residual error unit have different dimensionality, and can not be directly added, and a convolution layer with 1 multiplied by 1 in the residual error mapping is added for dimensionality reduction.
Inputting 2 continuous non-time sequence data in a preset time period; the non-time series data comprises date and daily non-time series data; the daily non-time sequence data of the preset time period comprise 7-dimensional data generated after daily average temperature, daily average electricity price and date type are subjected to one-hot coding; the non-time sequence data is 9-dimensional data;
the LSTM network layer comprises 3 layers of LSTM units; the method specifically comprises the following steps: firstly, connecting the output g of the TCN model with the output x2 of the input 2, and ensuring the dimensionality and the number of samples of data to be the same to obtain data C of formula (1)t;
ct=f(x(2,t),gt)
(1) In the formula: : f represents a connection operation;
the output layer of the LSTM network layer is a power load prediction result Y for predicting 96 time points of a dayOThe number of neurons in the output layer is set to 96; the layer network takes a Sigmoid function as an activation function.
Further, the method also comprises the step of optimizing the parameters of the model by adopting an Adam optimizer.
3. Has the advantages that:
the short-term load prediction model based on the TCN-LSTM can improve the prediction accuracy, provide a more accurate prediction result for a power grid system and provide a more reliable basis for the power grid system to flexibly adjust the power supply quantity.
Drawings
FIG. 1 is a schematic diagram of the TCN-LSTM based short term load prediction model of the present invention.
Detailed Description
The present invention will be described in detail with reference to the accompanying drawings.
The TCN model is a simple and universal convolutional neural network architecture for solving the time series problem. The TCN model is composed of a group of residual error units, each residual error unit is a small neural network with residual error connection, feedback and convergence of a deep network can be accelerated through the residual error connection, and the 'degeneration phenomenon' caused by the increase of network layers is solved. The residual unit contains 2 convolution units and nonlinear mapping. In the convolution unit, one-dimensional expansion causal convolution is firstly carried out, sampling intervals are adjusted through expansion coefficients, a larger Reception Field (RF) is achieved, namely an area range where characteristics on a convolution layer can be seen, a network can memorize enough long historical information, only the input before t time is convolved to obtain the output at t time, and it is guaranteed that future information cannot be leaked; then, carrying out normalization processing on the weight, and using a ReLU function as an activation function; and finally, adopting Dropout operation, and randomly discarding the neurons according to a certain probability to achieve the purposes of preventing overfitting and accelerating the model training speed. The non-linear mapping is to perform dimensionality reduction on data of high dimensionality when the input and output of the residual unit have different dimensionalities.
The recurrent neural network is a neural network with a good processing effect on time sequence data, but has a long-term dependence problem that a model cannot utilize data information in an earlier stage in a sequence as the length of an input sequence is increased. To solve this problem, Hochreiter et al, 1997, proposed an LSTM network with an added input gate itForgetting door ftOutput gate otThree logic units controlling the memory unitAnd (6) outputting. The LSTM network is characterized by the state transitions (c) of two chronologically adjacent memory cellst-1To ct) The activation function, the input gate and the forgetting gate are jointly used for controlling.
The TCN model extracts past data through one-dimensional causal convolution, guarantees time sequence, accelerates convergence speed through residual connection, and expands convolution to achieve time sequence feature extraction. The LSTM model is used as a variant of a recurrent neural network, has nonlinear fitting capability, can effectively extract data characteristics, and simultaneously obtains faster convergence rate.
Fig. 1 shows a short-term load prediction method based on TCN-LSTM, which is characterized in that: the method comprises the following steps:
the method comprises the following steps: acquiring historical power load data of a plurality of acquisition points in an area to be predicted within preset time as sample data; the recording interval of the data is in the order of minutes, including between 1 and 60 minutes; the time span of the historical load data is K times of the period in which the repeated change of the load can be observed, and K is an integer greater than or equal to 1;
step two: preprocessing sample data, eliminating abnormal data, filling incomplete data, and performing standardized processing;
step three: training a neural network model; the method specifically comprises the following steps: taking power load data and continuous non-time sequence data as input, taking power preset loads of each sampling point on a prediction day as output, and training a neural network model combining a time convolution network and an LSTM to obtain a short-term load prediction model based on TCN-LSTM; the non-time sequence data comprises date, week, daily average temperature and daily average electricity price;
the short-term load prediction model based on the TCN-LSTM comprises an input 1, a time convolution network layer, an input 2 and an LSTM network layer; the input 1 inputs a time convolution network layer, and the output of the time convolution network layer is combined with the input 2 and then input
The time convolution network layer comprises 2 layers of residual error units, and each layer of residual error unit comprises 2 convolution units and 1 nonlinear mapping; the convolution unit uses a ReLU function as an activation function and performs normalization operation on the weight of a convolution kernel; the size of the convolution kernel is 2, a Dropout coefficient with the coefficient of 0.4 is set, and Dropout setting can randomly select part of neurons to be inactivated, so that over-fitting training is prevented, and the convergence speed of the model is increased; setting the expansion coefficient to be (1, 2, 4, 8, 16, 32); the filter is 128; the input and output of the residual error unit have different dimensionality, and can not be directly added, and a convolution layer with 1 multiplied by 1 in the residual error mapping is added for dimensionality reduction.
Inputting 2 continuous non-time sequence data in a preset time period; the non-time series data comprises date and daily non-time series data; the daily non-time sequence data of the preset time period comprise 7-dimensional data generated by encoding daily average temperature, daily average electricity price and date type by oge-hot; the non-time sequence data is 9-dimensional data;
the LSTM network layer comprises 3 layers of LSTM units; the method specifically comprises the following steps: firstly, connecting the output g of the TCN model with the output x2 of the input 2, and ensuring the dimensionality and the number of samples of data to be the same to obtain data C of formula (1)t;
ct=f(x(2,t),gt)
(1) In the formula: : f represents a connection operation;
the output layer of the LSTM network layer is a power load prediction result Y for predicting 96 time points of a day0The number of neurons in the output layer is set to 96; the layer network takes a Sigmoid function as an activation function.
Further, the method also comprises the step of optimizing the parameters of the model by adopting an Adam optimizer.
Although the present invention has been described with reference to the preferred embodiments, it should be understood that various changes and modifications can be made therein by those skilled in the art without departing from the spirit and scope of the invention as defined by the appended claims.
Claims (2)
1. A short-term load prediction method based on TCN-LSTM is characterized in that: the method comprises the following steps:
the method comprises the following steps: acquiring historical power load data of a plurality of acquisition points in an area to be predicted within preset time as sample data; the recording interval of the data is in the order of minutes, including between 1 and 60 minutes; the time span of the historical load data is K times of the period in which the repeated change of the load can be observed, and K is an integer greater than or equal to 1;
step two: preprocessing sample data, eliminating abnormal data, filling incomplete data, and performing standardized processing;
step three: training a neural network model; the method specifically comprises the following steps: taking power load data and continuous non-time sequence data as input, taking power preset loads of each sampling point on a prediction day as output, and training a neural network model combining a time convolution network and an LSTM to obtain a short-term load prediction model based on TCN-LSTM; the non-time sequence data comprises date, week, daily average temperature and daily average electricity price;
the short-term load prediction model based on the TCN-LSTM comprises an input 1, a time convolution network layer, an input 2 and an LSTM network layer; the input 1 inputs a time convolution network layer, and the output of the time convolution network layer is combined with the input 2 and then input
The time convolution network layer comprises 2 layers of residual error units, and each layer of residual error unit comprises 2 convolution units and 1 nonlinear mapping; the convolution unit uses a ReLU function as an activation function and performs normalization operation on the weight of a convolution kernel; the size of the convolution kernel is 2, a Dropout coefficient with the coefficient of 0.4 is set, and Dropout setting can randomly select part of neurons to be inactivated, so that over-fitting training is prevented, and the convergence speed of the model is increased; setting the expansion coefficient to be (1, 2, 4, 8, 16, 32); the filter is 128; the input and output of the residual error unit have different dimensionalities, the addition operation can not be directly carried out, and the dimensionality reduction is carried out on the convolution layer with 1 multiplied by 1 added in the residual error mapping;
inputting 2 continuous non-time sequence data in a preset time period; the non-time series data comprises date and daily non-time series data; the daily non-time sequence data of the preset time period comprise 7-dimensional data generated after daily average temperature, daily average electricity price and date type are subjected to one-hot coding; the non-time sequence data is 9-dimensional data;
the LSTM network layer comprises 3 layers of LSTM units; the method specifically comprises the following steps: firstly, connecting the output g of the TCN model with the output x2 of the input 2, and ensuring the dimensionality and the number of samples of data to be the same to obtain data C of formula (1)t;
ct=f(x(2,t),gt)
(1) The formula is as follows: f represents a connection operation;
the output layer of the LSTM network layer is a power load prediction result Y for predicting 96 time points of a dayOThe number of neurons in the output layer is set to 96; the layer network takes a Sigmoid function as an activation function.
2. The TCN-LSTM based short-term load prediction method of claim 1, wherein: further comprising optimizing parameters of the model using an Adam optimizer.
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN202111215767.2A CN113902207A (en) | 2021-10-19 | 2021-10-19 | Short-term load prediction method based on TCN-LSTM |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN202111215767.2A CN113902207A (en) | 2021-10-19 | 2021-10-19 | Short-term load prediction method based on TCN-LSTM |
Publications (1)
Publication Number | Publication Date |
---|---|
CN113902207A true CN113902207A (en) | 2022-01-07 |
Family
ID=79192834
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN202111215767.2A Withdrawn CN113902207A (en) | 2021-10-19 | 2021-10-19 | Short-term load prediction method based on TCN-LSTM |
Country Status (1)
Country | Link |
---|---|
CN (1) | CN113902207A (en) |
Cited By (3)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN114298446A (en) * | 2022-03-11 | 2022-04-08 | 南京工业大学 | Wind power prediction method and device based on digital-analog dual-drive and storage medium |
CN115014445A (en) * | 2022-08-10 | 2022-09-06 | 中国农业大学 | Smart fishery multi-dimensional panoramic perception monitoring method, system and device |
CN115936185A (en) * | 2022-11-15 | 2023-04-07 | 国网江苏省电力有限公司苏州供电分公司 | Short-term power load and carbon emission prediction method and system based on DCNN-LSTM-AE-AM |
-
2021
- 2021-10-19 CN CN202111215767.2A patent/CN113902207A/en not_active Withdrawn
Cited By (6)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN114298446A (en) * | 2022-03-11 | 2022-04-08 | 南京工业大学 | Wind power prediction method and device based on digital-analog dual-drive and storage medium |
CN114298446B (en) * | 2022-03-11 | 2022-06-03 | 南京工业大学 | Wind power prediction method and device based on digital-analog dual-drive and storage medium |
CN115014445A (en) * | 2022-08-10 | 2022-09-06 | 中国农业大学 | Smart fishery multi-dimensional panoramic perception monitoring method, system and device |
CN115014445B (en) * | 2022-08-10 | 2022-11-15 | 中国农业大学 | Smart fishery multi-dimensional panoramic perception monitoring method, system and device |
US11758887B1 (en) | 2022-08-10 | 2023-09-19 | China Agricultural University | Method, system and apparatus for intelligently monitoring aquafarm with multi-dimensional panoramic perception |
CN115936185A (en) * | 2022-11-15 | 2023-04-07 | 国网江苏省电力有限公司苏州供电分公司 | Short-term power load and carbon emission prediction method and system based on DCNN-LSTM-AE-AM |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
CN109685252B (en) | Building energy consumption prediction method based on cyclic neural network and multi-task learning model | |
CN113902207A (en) | Short-term load prediction method based on TCN-LSTM | |
CN108448610B (en) | Short-term wind power prediction method based on deep learning | |
Elsaraiti et al. | Solar power forecasting using deep learning techniques | |
CN110751318A (en) | IPSO-LSTM-based ultra-short-term power load prediction method | |
CN113902202A (en) | Short-term load prediction model and method based on double attention mechanism and LSTM | |
Islam et al. | LSTM-based electrical load forecasting for Chattogram city of Bangladesh | |
Yongchun | Application of Elman neural network in short-term load forecasting | |
CN114519471A (en) | Electric load prediction method based on time sequence data periodicity | |
Li et al. | Application of ARIMA and LSTM in relative humidity prediction | |
Xiao et al. | Online sequential extreme learning machine algorithm for better predispatch electricity price forecasting grids | |
Liu et al. | Research on wind speed prediction of wind power system based on GRU deep learning | |
Zhou et al. | Short-term wind power prediction based on EMD-LSTM | |
CN115018179A (en) | CNN-BiGRU-based ultra-short-term power load prediction method | |
Wilms et al. | On the necessity of exogenous variables for load, pv and wind day-ahead forecasts using recurrent neural networks | |
Prema et al. | LSTM based Deep Learning model for accurate wind speed prediction | |
CN109146191B (en) | Short-term power load prediction method and system based on generic implication model | |
CN116885699A (en) | Power load prediction method based on dual-attention mechanism | |
Zuo | Integrated forecasting models based on LSTM and TCN for short-term electricity load forecasting | |
CN113393119B (en) | Stepped hydropower short-term scheduling decision method based on scene reduction-deep learning | |
Fan et al. | Ultra-short-term bus load forecasting method based on multi-source data and hybrid neural network | |
Emamian et al. | Solar power forecasting with LSTM network ensemble | |
Zhang et al. | Solar-mixer: An efficient end-to-end model for long-sequence photovoltaic power generation time series forecasting | |
Xu et al. | Short-term gas load forecast based on tcn-bigru | |
Zivkovic et al. | Optimized recurrent neural networks with attention for wind farm energy generation forecasting |
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 | ||
WW01 | Invention patent application withdrawn after publication |
Application publication date: 20220107 |
|
WW01 | Invention patent application withdrawn after publication |