WO2023093774A1 - 基于深度学习的风电集群功率预测方法 - Google Patents
基于深度学习的风电集群功率预测方法 Download PDFInfo
- Publication number
- WO2023093774A1 WO2023093774A1 PCT/CN2022/133796 CN2022133796W WO2023093774A1 WO 2023093774 A1 WO2023093774 A1 WO 2023093774A1 CN 2022133796 W CN2022133796 W CN 2022133796W WO 2023093774 A1 WO2023093774 A1 WO 2023093774A1
- Authority
- WO
- WIPO (PCT)
- Prior art keywords
- wind power
- data
- cluster
- historical
- prediction
- Prior art date
Links
- 238000000034 method Methods 0.000 title claims abstract description 46
- 238000013135 deep learning Methods 0.000 title claims abstract description 33
- 238000012549 training Methods 0.000 claims abstract description 34
- 238000007781 pre-processing Methods 0.000 claims abstract description 20
- 238000000605 extraction Methods 0.000 claims description 39
- 230000004927 fusion Effects 0.000 claims description 39
- 230000006870 function Effects 0.000 claims description 34
- 238000004590 computer program Methods 0.000 claims description 21
- 238000013528 artificial neural network Methods 0.000 claims description 17
- 238000013527 convolutional neural network Methods 0.000 claims description 13
- 230000003993 interaction Effects 0.000 claims description 12
- 230000007246 mechanism Effects 0.000 claims description 11
- 230000004913 activation Effects 0.000 claims description 9
- 238000012545 processing Methods 0.000 claims description 7
- 230000009467 reduction Effects 0.000 claims description 6
- 238000005457 optimization Methods 0.000 claims description 5
- ORILYTVJVMAKLC-UHFFFAOYSA-N Adamantane Natural products C1C(C2)CC3CC1CC2C3 ORILYTVJVMAKLC-UHFFFAOYSA-N 0.000 claims description 4
- 238000004458 analytical method Methods 0.000 claims description 4
- 238000004422 calculation algorithm Methods 0.000 claims description 4
- 238000010276 construction Methods 0.000 claims description 4
- 238000004364 calculation method Methods 0.000 claims description 3
- 230000002123 temporal effect Effects 0.000 claims 1
- 230000002159 abnormal effect Effects 0.000 description 4
- 238000001514 detection method Methods 0.000 description 4
- 238000010586 diagram Methods 0.000 description 3
- 238000010248 power generation Methods 0.000 description 3
- 230000008569 process Effects 0.000 description 3
- 238000003491 array Methods 0.000 description 2
- 238000005516 engineering process Methods 0.000 description 2
- 238000013277 forecasting method Methods 0.000 description 2
- 239000000463 material Substances 0.000 description 2
- 230000003287 optical effect Effects 0.000 description 2
- 238000009987 spinning Methods 0.000 description 2
- 238000012360 testing method Methods 0.000 description 2
- 238000013473 artificial intelligence Methods 0.000 description 1
- 230000009286 beneficial effect Effects 0.000 description 1
- 238000013075 data extraction Methods 0.000 description 1
- 238000013500 data storage Methods 0.000 description 1
- 238000011161 development Methods 0.000 description 1
- 239000000284 extract Substances 0.000 description 1
- 239000000835 fiber Substances 0.000 description 1
- 238000009472 formulation Methods 0.000 description 1
- 239000000203 mixture Substances 0.000 description 1
- 238000012986 modification Methods 0.000 description 1
- 230000004048 modification Effects 0.000 description 1
- 238000003058 natural language processing Methods 0.000 description 1
- 238000003062 neural network model Methods 0.000 description 1
- 238000010606 normalization Methods 0.000 description 1
- 238000000053 physical method Methods 0.000 description 1
- 238000007619 statistical method Methods 0.000 description 1
- 238000006467 substitution reaction Methods 0.000 description 1
- 230000009897 systematic effect Effects 0.000 description 1
Images
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F30/00—Computer-aided design [CAD]
- G06F30/20—Design optimisation, verification or simulation
- G06F30/27—Design optimisation, verification or simulation using machine learning, e.g. artificial intelligence, neural networks, support vector machines [SVM] or training a model
-
- 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
-
- 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
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F2113/00—Details relating to the application field
- G06F2113/06—Wind turbines or wind farms
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F2119/00—Details relating to the type or aim of the analysis or the optimisation
- G06F2119/06—Power analysis or power optimisation
Definitions
- the disclosure relates to the technical fields of deep learning, artificial intelligence, neural network, natural language processing and new energy, and in particular to a method, device, computer equipment and storage medium for predicting power of a wind power cluster based on deep learning.
- the problems of the past methods are: either data from a certain source, or only considering The characteristics of spatio-temporal data, or the neural network model method of constructing a single spatio-temporal data feature, and so on. These methods are not accurate enough to predict the power of wind clusters, and the error is high. At the same time, they cannot meet the requirements of the existing power grid supply and dispatch, which brings a lot of inconvenience to the power supply system, such as increased operating costs such as increased spinning reserves, and indirectly increased manpower. and other expenses. With the rapid application of deep learning technology, the use of heterogeneous data to build a deep learning fusion model will help improve and optimize the key scientific issue of wind power cluster power prediction, and provide intelligence and digitalization for the scheduling and operation of the power grid system. , Systematic innovation and upgrading, reducing various operating costs of wind power equipment.
- the present disclosure provides a wind power cluster power prediction method, device, computer equipment, and storage medium based on deep learning, aiming at avoiding false positives and misreporting phenomena in the abnormal detection process of power generation equipment, and improving the accuracy of abnormal detection of power generation equipment .
- the embodiment of the first aspect of the present disclosure proposes a method for predicting power of a wind power cluster based on deep learning.
- the method includes: obtaining historical heterogeneous data of wind power clusters and performing data preprocessing, using the preprocessed heterogeneous historical data of wind power clusters as a training set; constructing a wind power cluster power prediction network model, and using the training set to construct The wind power cluster power prediction network model is trained; and the real-time wind power cluster heterogeneous data is preprocessed and input into the trained wind power cluster power prediction network model, and the output result is used as the wind power cluster power prediction result.
- the steps of acquiring historical heterogeneous data of wind power clusters and performing data preprocessing include:
- the historical heterogeneous data of wind power clusters are real-time wind power cluster power data extracted from the SCADA system, historical wind power clusters Power data, NWP data and corresponding geographic data;
- the historical heterogeneous data of the wind power cluster is normalized
- w' represents the normalized value
- w represents the true value of the sample in the historical heterogeneous data of the selected wind power cluster
- w min represents the minimum value of the sample in the historical heterogeneous data of the selected wind power cluster
- w max represents the sample maximum value in the historical heterogeneous data of the selected wind power cluster.
- the wind power cluster power prediction network model includes a feature extraction module, a key information prediction module, a feature fusion module, and a result prediction module; wherein, the feature extraction module is a feature extraction neural network for preprocessing data Feature extraction is performed on the historical heterogeneous data of the wind power cluster; the key information prediction module is used to obtain the internal interaction characteristics of the respective characteristics of the historical heterogeneous data of the wind power cluster and the correlation features between the data; the feature fusion module is used to The characteristics of the historical heterogeneous data of the wind power cluster are fused and spliced to obtain multi-modal feature fusion information; the result prediction module is used to calculate the prediction result according to the feature fusion information, and complete the power prediction of the wind power cluster.
- the feature extraction module is a feature extraction neural network for preprocessing data Feature extraction is performed on the historical heterogeneous data of the wind power cluster
- the key information prediction module is used to obtain the internal interaction characteristics of the respective characteristics of the historical heterogeneous data of the wind power cluster and the
- the feature extraction neural network is a CNN+BiLSTM combined neural network; CNN is good at obtaining spatiotemporal data features and combining BiLSTM to obtain forward and backward sequences for time series data, thereby completing feature extraction of heterogeneous data.
- the key information prediction module uses the attention mechanism to obtain the internal interaction characteristics of the respective characteristics of the historical heterogeneous data of the wind power cluster and the correlation features between the data; the feature fusion module merges the historical heterogeneous data of the wind power cluster The features of the multi-modal fusion feature containing the characteristics of spatio-temporal complementarity and relevance of the context are obtained.
- the result prediction module uses the fully connected layer to calculate the predicted result, uses the activation function ReLU function as the activation function of the fully connected layer, and calculates the predicted result using the normalized reduction function calculation formula (2) Get the restored power predictions:
- W pre represents the predicted output value of the wind power cluster power prediction network model
- W o represents the restored power prediction value
- w max represents the sample maximum value in the historical heterogeneous data of the selected wind power cluster
- w min represents the selected wind power cluster Sample minimum in clustered historical heterogeneous data.
- the step of training the constructed wind power cluster power prediction network model through the training set includes:
- the predicted key features are merged to obtain the fusion features with the characteristics of contextual space-time complementarity and correlation, input power prediction module, calculate the predicted results, compare with the actual wind power cluster power results, use the mean square error as the loss function, and train the network
- the optimization uses the Adam algorithm, and the network training is completed by continuously adjusting the network functions and parameters until the predicted results are consistent with the marked power results.
- the method further includes a step of displaying the prediction result; the display method includes at least one of the following: text display, voice broadcast, outbound terminal, Email, SMS reminder, smart speaker.
- the embodiment of the second aspect of the present disclosure proposes a wind power cluster power forecasting device based on deep learning, including: a data acquisition module, used to acquire historical heterogeneous data of wind power clusters and perform data preprocessing, and the preprocessed wind power cluster is heterogeneous
- the historical data is used as a training set
- the model construction module is used to construct a wind power cluster power forecasting network model, and the wind power cluster power forecasting network model constructed is trained through the training set
- the power forecasting module is used to convert real-time
- the wind power cluster heterogeneous data is preprocessed and input into the trained wind power cluster power prediction network model, and the output result is used as the wind power cluster power prediction result.
- the embodiment of the third aspect of the present disclosure proposes a computer device, including a processor and a memory for storing a computer program executable by the processor.
- the processor executes the computer program, the deep learning-based wind power of the embodiment of the first aspect is realized. Cluster Power Prediction Methods.
- the embodiment of the fourth aspect of the present disclosure proposes a non-transitory computer-readable storage medium, on which a computer program is stored.
- the computer program is executed by a processor, the deep learning-based wind power cluster power prediction method of the embodiment of the first aspect is implemented.
- the embodiment of the fifth aspect of the present disclosure proposes a computer program product, including a computer program.
- the computer program is executed by a processor, the deep learning-based wind power cluster power prediction method of the embodiment of the first aspect is implemented.
- the wind power cluster power prediction method based on deep learning constructs a wind power cluster power prediction network model, extracts features from heterogeneous data through a feature extraction network, and predicts key information based on the extracted features based on the attention mechanism.
- the modal fusion strategy is fused to generate multi-modal fusion features, and the wind power cluster power prediction is performed according to the generated multi-modal fusion features.
- Fig. 1 is a schematic flowchart of a method for predicting power of a wind power cluster based on deep learning provided by the present disclosure.
- Fig. 2 is a schematic structural diagram of a wind power cluster power forecasting network model based on a deep learning-based wind power cluster power forecasting method provided by the present disclosure.
- Fig. 3 is a schematic structural diagram of a wind power cluster power prediction device based on deep learning provided by the present disclosure.
- Fig. 4 is a schematic structural diagram of a non-transitory computer-readable storage medium provided by the present disclosure.
- FIG. 1 is a schematic flowchart of a method for predicting power of a wind power cluster based on deep learning provided by an embodiment of the present disclosure. The method includes the following steps 110 to 130 .
- Step 110 acquire the historical heterogeneous data of the wind power cluster and perform data preprocessing, and use the preprocessed heterogeneous historical data of the wind power cluster as a training set.
- the historical heterogeneous data of the wind power cluster is obtained from the SCADA system database.
- the historical heterogeneous data of wind power clusters specifically include real-time wind power cluster power data extracted from the SCADA system, historical wind power cluster power data, NWP data and corresponding geographical data. weather forecast data.
- the steps of data preprocessing after data extraction, as shown in Figure 2, specifically include:
- w' represents the normalized value
- w represents the true value of the sample in the historical heterogeneous data of the selected wind power cluster
- w min represents the sample minimum value in the historical heterogeneous data of the selected wind power cluster
- w max represents the selected The maximum value of samples in the historical heterogeneous data of wind power clusters.
- the heterogeneous data is divided into training set and test set, and the output power data of the last year is taken as the training set, or the real-time data of the SCADA system can be used as the test set.
- Step 120 Construct a wind power cluster power forecasting network model, and train the constructed wind power cluster power forecasting network model through the training set.
- FIG. 2 The network structure of the wind power cluster power prediction network model constructed in this disclosure is shown in FIG. 2 , including a feature extraction module 102 , a key information prediction module 103 , a feature fusion module 104 and a result prediction module 105 .
- the feature extraction module 102 is a feature extraction neural network, which is used to extract features from the historical heterogeneous data of wind power clusters after data preprocessing; the feature extraction neural network is CNN+BiLSTM combined neural network; it is good at obtaining spatio-temporal data features and combining them with CNN BiLSTM obtains forward and backward sequences for time series data, thereby completing feature extraction of heterogeneous data.
- the key information prediction module 103 is used to obtain the internal interaction characteristics of the respective characteristics of the historical heterogeneous data of the wind power cluster and the correlation characteristics between the data; the key information prediction module uses the attention mechanism to obtain the internal interaction characteristics and The characteristics of the correlation between the data.
- the feature fusion module 104 is used to perform feature fusion and splicing on the features of the historical heterogeneous data of the wind power cluster to obtain multimodal feature fusion information; the feature fusion module obtains the spatiotemporal context containing the context by merging the features of the historical heterogeneous data of the wind power cluster. Multimodal Fusion Features of Complementarity and Correlation Features.
- the result prediction module 105 is used to calculate the prediction result according to the feature fusion information, and complete the power prediction of the wind power cluster.
- the result prediction module 105 uses the fully connected layer to calculate the prediction score result, uses the activation function ReLU function as the activation function of the fully connected layer, calculates the predicted result through the activation function, and uses the normalized reduction function to calculate the formula (2) to obtain its original Size; the mean absolute error MAE and the root mean square error RMSE are selected as the loss function to correct the normalized reduction function calculation formula (2).
- w pre represents the predicted output value of the network model
- W o represents the restored power prediction value
- w min represents the sample minimum value in the historical heterogeneous data of the selected wind power cluster
- w max represents the value of the selected wind power cluster in the historical heterogeneous data Sample max.
- the step of training the constructed wind power cluster power prediction network model through the training set includes:
- the predicted key features are merged to obtain the fusion features with the characteristics of contextual space-time complementarity and correlation, input power prediction module, calculate the predicted results, compare with the actual wind power cluster power results, use the mean square error as the loss function, and train the network
- the optimization uses the Adam algorithm, and the network training is completed by continuously adjusting the network functions and parameters until the predicted results are consistent with the marked power results.
- each type of heterogeneous data is Input to a feature extraction network
- the feature extraction network includes a connected CNN network model and a BiLSTM network model, and the features of each type of heterogeneous data extracted through the feature extraction network output
- the key information prediction module includes an attention mechanism network and a fully connected layer , the features of the four types of heterogeneous data output by the feature extraction network are respectively input into an attention mechanism network, the output results are input into the fully connected layer, and the key information features of each type of heterogeneous data are output.
- the key information features of the four types of heterogeneous data are input into the feature fusion module 104 for feature fusion, and after the fusion is completed, the fusion features are input into the power prediction module 105 for prediction.
- the network model is trained by inputting historical data into the constructed model, and if the output data of the wind power cluster power output by inputting the historical data is consistent with the actual data, the training is completed.
- S130 Input the preprocessed real-time heterogeneous wind power cluster data into the trained wind power cluster power prediction network model, and output the result as the wind power cluster power prediction result.
- this disclosure selects 5 years of historical data and real-time collected text data from the SCADA system, and calculates the output power prediction in the sub-cluster through the above steps, and the entire cluster power can be summed by the sub-clusters.
- the disclosure is of great significance for the prediction of power output in the station area for the formulation of dispatching plans and spinning reserve capacity of the power system.
- the abnormal detection result of the power generation equipment If the abnormal detection result of the power generation equipment is obtained, the abnormal detection result will be displayed; the display methods include at least: text display, voice broadcast, outbound terminal, email, SMS reminder, and smart speaker.
- the present disclosure also proposes a wind power cluster power prediction device based on deep learning, as shown in FIG. 3 , including a data acquisition module 310 , a model construction module 320 and a power prediction module 330 .
- the data acquisition module 310 is used to acquire the historical heterogeneous data of the wind power cluster and perform data preprocessing, and use the preprocessed heterogeneous historical data of the wind power cluster as a training set;
- the model construction module 320 is used to construct a wind power cluster power prediction network model, and train the constructed wind power cluster power prediction network model through the training set; wherein, the wind power cluster power prediction network model includes sequentially connected Feature extraction module, key information prediction module, feature fusion module and result prediction module;
- the power prediction module 330 is configured to input the preprocessed real-time heterogeneous wind power cluster data into the trained wind power cluster power prediction network model, and output the result as the wind power cluster power prediction result.
- the present disclosure also proposes a computer device, including: a memory for storing a computer program executable by the processor, and when the processor executes the computer program, the deep learning-based Wind power cluster power forecasting method.
- the non-transitory computer-readable storage medium includes a memory 410 of instructions and an interface 430 , and the above instructions can be executed by the processor 420 of the deep learning-based wind power cluster power prediction device to complete the above method.
- the storage medium may be a non-transitory computer-readable storage medium, for example, the non-transitory computer-readable storage medium may be ROM, random access memory (RAM), CD-ROM, magnetic tape, floppy disk, and optical data storage device, etc. .
- the present disclosure also proposes a non-transitory computer-readable storage medium, on which a computer program is stored, and when the computer program is executed by a processor, the power prediction of the wind power cluster based on deep learning as in the embodiment of the present disclosure is realized. method.
- the present disclosure also proposes a computer program product, including a computer program, wherein, when the computer program is executed by a processor, the deep learning-based wind power cluster power prediction method according to the embodiments of the present disclosure is implemented.
- first and second are used for descriptive purposes only, and cannot be interpreted as indicating or implying relative importance or implicitly specifying the quantity of indicated technical features.
- the features defined as “first” and “second” may explicitly or implicitly include at least one of these features.
- “plurality” means at least two, such as two, three, etc., unless otherwise specifically defined.
- a "computer-readable medium” may be any device that can contain, store, communicate, propagate or transmit a program for use in or in conjunction with an instruction execution system, device or device.
- computer-readable media include the following: electrical connection with one or more wires (electronic device), portable computer disk case (magnetic device), random access memory (RAM), Read Only Memory (ROM), Erasable and Editable Read Only Memory (EPROM or Flash Memory), Fiber Optic Devices, and Portable Compact Disc Read Only Memory (CDROM).
- the computer-readable medium may even be paper or other suitable medium on which the program can be printed, as it may be possible, for example, by optically scanning the paper or other medium, followed by editing, interpreting, or other suitable processing if necessary.
- the program is processed electronically and stored in computer memory.
- various parts of the present disclosure may be implemented in hardware, software, firmware or a combination thereof.
- various steps or methods may be implemented by software or firmware stored in memory and executed by a suitable instruction execution system.
- a suitable instruction execution system For example, if implemented in hardware as in another embodiment, it can be implemented by any one or a combination of the following techniques known in the art: a discrete Logic circuits, ASICs with suitable combinational logic gates, Programmable Gate Arrays (PGA), Field Programmable Gate Arrays (FPGA), etc.
- each functional unit in each embodiment of the present disclosure may be integrated into one processing module, each unit may exist separately physically, or two or more units may be integrated into one module.
- the above-mentioned integrated modules can be implemented in the form of hardware or in the form of software function modules. If the integrated modules are implemented in the form of software function modules and sold or used as independent products, they can also be stored in a computer-readable storage medium.
- the storage medium mentioned above may be a read-only memory, a magnetic disk or an optical disk, and the like.
Landscapes
- Engineering & Computer Science (AREA)
- Theoretical Computer Science (AREA)
- Physics & Mathematics (AREA)
- Evolutionary Computation (AREA)
- Software Systems (AREA)
- Artificial Intelligence (AREA)
- General Physics & Mathematics (AREA)
- General Engineering & Computer Science (AREA)
- Computing Systems (AREA)
- Mathematical Physics (AREA)
- General Health & Medical Sciences (AREA)
- Molecular Biology (AREA)
- Computational Linguistics (AREA)
- Biophysics (AREA)
- Biomedical Technology (AREA)
- Data Mining & Analysis (AREA)
- Life Sciences & Earth Sciences (AREA)
- Health & Medical Sciences (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Medical Informatics (AREA)
- Computer Hardware Design (AREA)
- Geometry (AREA)
- Management, Administration, Business Operations System, And Electronic Commerce (AREA)
- Supply And Distribution Of Alternating Current (AREA)
Abstract
本公开提出一种基于深度学习的风电集群功率预测方法。该方法包括:获取风电集群历史异构数据并进行数据预处理,将预处理后的风电集群异构历史数据作为训练集;构建风电集群功率预测网络模型,并通过所述训练集对构建的所述风电集群功率预测网络模型进行训练;和将实时的风电集群异构数据预处理后输入训练完成的所述风电集群功率预测网络模型中,输出结果作为风电集群功率预测结果。
Description
相关申请的交叉引用
本申请基于申请号为202111421723.5、申请日为2021年11月26日的中国专利申请提出,并要求该中国专利申请的优先权,该中国专利申请的全部内容在此引入本申请作为参考。
本公开涉及深度学习、人工智能、神经网络、自然语言处理和新能源领域技术领域,尤其涉及一种基于深度学习的风电集群功率预测方法、装置、计算机设备及存储介质。
伴随深度学习与智能风电场技术相结合发展,对电力系统供电调度及运行提出了更大的挑战,例如:如何对风电集群功率进行精准预测,保障电力系统稳定及调度运行的最佳运营方案等问题,特别是目前大规模集群内的区域风电功率预测准确性问题关系整体供电系统的安全运营。因此风电集群功率精准预测对新能源供电的智能运行及调度至关重要。当前预测风电功率方法有物理方法、统计方法这两种为主,但是这些方法普遍有局限性,比如风电集群功率预测不准确,过去方法的问题有:要么对某一来源数据,要么仅考虑到时空数据的特征,要么构建单一时空数据特征的神经网络模型方法,等等。这些方法对风力集群功率预测不够准、误差偏高,同时,也不能满足现有电网供应及调度的要求,给供电系统带来诸多不便,例如增加旋转备用量等运营成本增加,间接增加了人力等支出。随着深度学习技术的快速落地应用,利用异构数据构建深度学习融合模型有助于风电集群功率预测这一关键性科学问题的提高及优化,为电网系统的调度、运行等方面智能化、数字化、系统化创新升级,降低风力发电设备各项运营成本。
发明内容
本公开提供一种基于深度学习的风电集群功率预测方法、装置、计算机设备及存储介质,旨在规避发电设备异常检测过程中的漏报误报、错报现象,提升预测发电设备异常检测准确率。
本公开第一方面实施例提出一种基于深度学习的风电集群功率预测方法。所述方法包括:获取风电集群历史异构数据并进行数据预处理,将预处理后的风电集群异构历史数据作为训练集;构建风电集群功率预测网络模型,并通过所述训练集对构建的所述风电集群功率预测网络模型进行训练;和将实时的风电集群异构数据预处理后输入训练完成的所述风电集群功率预测网络模型中,输出结果作为风电集群功率预测结果。
在一些实施例中,获取风电集群历史异构数据并进行数据预处理的步骤包括:
对不同数据格式的风电集群历史异构数据进行数据格式解析和标量纲处理,转换为统一 格式;其中,所述风电集群历史异构数据是从SCADA系统提取的实时风电集群功率数据、历史风电集群功率数据、NWP数据及对应的地理数据;
依据公式(1)对风电集群历史异构数据进行归一化;
其中,w’代表归一化后的值,w代表所选风电集群历史异构数据中的样本真值;代表样本真值,w
min代表所选风电集群历史异构数据中的样本最小值;w
max代表所选风电集群历史异构数据中的样本最大值。
在一些实施例中,风电集群功率预测网络模型包括特征提取模块、关键信息预测模块、特征融合模块和结果预测模块;其中,所述特征提取模块,为特征提取神经网络,用于对数据预处理后的风电集群历史异构数据进行特征提取;所述关键信息预测模块,用于得到风电集群历史异构数据各自特征内部交互特征及数据间的关联性特征;所述特征融合模块,用于对风电集群历史异构数据的特征进行特征融合拼接,得到多模态特征融合信息;所述结果预测模块,用于根据所述特征融合信息计算预测结果,完成风电集群功率预测。
在一些实施例中,特征提取神经网络为CNN+BiLSTM结合神经网络;借助CNN擅长获得时空数据特征及结合BiLSTM对时序数据获得前后向序列,从而完成异构数据的特征提取。
在一些实施例中,关键信息预测模块利用注意力机制,得到风电集群历史异构数据各自特征内部交互特征及数据间的关联性特征;所述特征融合模块通过合并所述风电集群历史异构数据的特征,得到含有上下文的时空互补性及关联性特点的多模态融合特征。
在一些实施例中,所述结果预测模块采用全连接层计算预测结果,采用激活函数ReLU函数作为所述全连接层的激活函数,计算得到预测结果用归一化还原函数计算公式(2)而获得还原后的功率预测值:
w
o=w
pre(w
max-w
min)+w
min (2)
其中W
pre代表所述风电集群功率预测网络模型的预测输出值,W
o代表还原后的功率预测值,w
max代表所选风电集群历史异构数据中的样本最大值,w
min代表所选风电集群历史异构数据中的样本最小值。
在一些实施例中,通过所述训练集对构建的所述风电集群功率预测网络模型进行训练的步骤包括:
将预处理后的训练集数据输入特征提取模块的特征提取神经网络,借助CNN擅长获得时空数据特征及结合BiLSTM对时序数据获得前后向序列,从而完成异构数据的特征提取;
利用注意力机制得到异构数据各自特征内部交互特征及数据间的关联性特征,输入关键信息预测模块的全连接层,得到预测关键特征;
将预测关键特征进行合并操作,得到含有上下文时空互补性和关联性特点的融合特征,输入功率预测模块,计算预测结果,与实际风电集群功率结果进行对比,以均方误差作为损失函数,网络训练优化采用Adam算法,通过不断调整网络函数和参数,直至预测结果与标记的功率结果一致时,完成网络训练。
在一些实施例中,在获取风电集群功率预测结果之后,所述方法还包括将预测结果进行展示的步骤;展示方式包括以下各项中的至少一种:文本显示、语音播报、外呼终端、邮件、短信提醒、智能音箱。
本公开第二方面实施例提出一种基于深度学习的风电集群功率预测装置,包括:数据获取模块,用于获取风电集群历史异构数据并进行数据预处理,将预处理后的风电集群异构历史数据作为训练集;模型构建模块,用于构建风电集群功率预测网络模型,并通过所述训练集对构建的所述风电集群功率预测网络模型进行训练;和功率预测模块,用于将实时的风电集群异构数据预处理后输入训练完成的所述风电集群功率预测网络模型中,输出结果作为风电集群功率预测结果。
本公开第三方面实施例提出一种计算机设备,包括处理器和用于存储所述处理器可执行计算机程序的存储器,处理器执行计算机程序时,实现第一方面实施例的基于深度学习的风电集群功率预测方法。
本公开第四方面实施例提出一种非临时性计算机可读存储介质,其上存储有计算机程序,计算机程序被处理器执行时实现第一方面实施例的基于深度学习的风电集群功率预测方法。
本公开第五方面实施例提出一种计算机程序产品,包括计算机程序,计算机程序被处理器执行时实现第一方面实施例的基于深度学习的风电集群功率预测方法。
本公开提供的基于深度学习的风电集群功率预测方法,构建风电集群功率预测网络模型,通过特征提取网络对异构数据进行特征提取,将提取的特征基于注意力机制进行关键信息预测后,采用多模态融合策略融合生成多模态融合特征,根据生成的多模态融合特征进行风电集群功率预测。通过本公开,能够提高预测风电集群功率的精准性和稳定性,有利于电网系统运行调度及系统优化工作。
本公开上述的和/或附加的方面和优点从下面结合附图对实施例的描述中将变得明显和容易理解,其中:
图1是本公开提供的一种基于深度学习的风电集群功率预测方法的流程示意图。
图2是本公开提供的一种基于深度学习的风电集群功率预测方法的风电集群功率预测网络模型的结构示意图。
图3是本公开提供的一种基于深度学习的风电集群功率预测装置的结构示意图。
图4是本公开提供的一种非临时性计算机可读存储介质的结构示意图。
下面详细描述本公开的实施例,实施例的示例在附图中示出,其中自始至终相同或类似的标号表示相同或类似的元件或具有相同或类似功能的元件。下面通过参考附图描述的实施例是示例性的,旨在用于解释本公开,而不能理解为对本公开的限制。
图1为本公开实施例所提供的一种基于深度学习的风电集群功率预测方法的流程示意图。该方法包括以下步骤110至步骤130。
步骤110,获取风电集群历史异构数据并进行数据预处理,将预处理后的风电集群异构历史数据作为训练集。
本公开中,风电集群历史异构数据取自SCADA系统数据库。风电集群历史异构数据具体包括从SCADA系统提取的实时风电集群功率数据、历史风电集群功率数据、NWP数据及对应的地理数据,NWP是对应采集实时风电集群功率数据和历史风电集群功率数据时的天气预报数据。提取数据后进行数据预处理的步骤,如图2中所示,具体包括:
对不同数据格式的风电集群历史异构数据进行数据格式解析和标量纲处理,转换为统一格式。
转换格式后,进行归一化处理,依据公式(1)对风电集群历史异构数据进行归一化;
其中,w’代表归一化后的值,w代表所选风电集群历史异构数据中的样本真值;w
min代表所选风电集群历史异构数据中的样本最小值;w
max代表所选风电集群历史异构数据中的样本最大值。
异构数据划分为训练集和测试集,取最近1年输出功率数据为训练集,也可以SCADA系统实时数据为测试集。
数据预处理完成后进入步骤120。
步骤120:构建风电集群功率预测网络模型,并通过所述训练集对构建的所述风电集群功率预测网络模型进行训练。
针对所收集的异构数据如何获得有效融合特征中含有上下文的时空互补性及关联性,进而提高预测风电集群功率准确性等关键问题,构建了基于异构数据和深多模态深度学习的风电集群功率预测模型。本公开构建的风电集群功率预测网络模型的网络结构如图2所示,包括特征提取模块102、关键信息预测模块103、特征融合模块104和结果预测模块105。
特征提取模块102,为特征提取神经网络,用于对数据预处理后的风电集群历史异构数据进行特征提取;特征提取神经网络为CNN+BiLSTM结合神经网络;借助CNN擅长获得 时空数据特征及结合BiLSTM对时序数据获得前后向序列,从而完成异构数据的特征提取。
关键信息预测模块103,用于得到风电集群历史异构数据各自特征内部交互特征及数据间的关联性特征;关键信息预测模块利用注意力机制,得到风电集群历史异构数据各自特征内部交互特征及数据间的关联性特征。
特征融合模块104,用于对风电集群历史异构数据的特征进行特征融合拼接,得到多模态特征融合信息;特征融合模块通过合并所述风电集群历史异构数据的特征,得到含有上下文的时空互补性及关联性特点的多模态融合特征。
结果预测模块105,用于根据特征融合信息计算预测结果,完成风电集群功率预测。
结果预测模块105采用全连接层计算预测得分结果,采用激活函数ReLU函数作为全连接层的激活函数,通过激活函数计算得到预测结果,并用归一化还原函数计算公式(2)而获得其原有大小;选择平均绝对误差MAE和均方根误差RMSE作为损失函数,用于校正归一化还原函数计算公式(2)。
w
o=w
pre(w
max-w
min)+w
min (2)
其中w
pre代表网络模型预测输出值,W
o代表还原后的功率预测值,w
min代表所选风电集群历史异构数据中的样本最小值,w
max代表所选风电集群历史异构数据中的样本最大值。
通过所述训练集对构建的所述风电集群功率预测网络模型进行训练的步骤包括:
将预处理后的训练集数据输入特征提取模块的特征提取神经网络,借助CNN擅长获得时空数据特征及结合BiLSTM对时序数据获得前后向序列,从而完成异构数据的特征提取;
利用注意力机制得到异构数据各自特征内部交互特征及数据间的关联性特征,输入关键信息预测模块的全连接层,得到预测关键特征;
将预测关键特征进行合并操作,得到含有上下文时空互补性和关联性特点的融合特征,输入功率预测模块,计算预测结果,与实际风电集群功率结果进行对比,以均方误差作为损失函数,网络训练优化采用Adam算法,通过不断调整网络函数和参数,直至预测结果与标记的功率结果一致时,完成网络训练。
如图2所示,分别对四类异构数据进行采集,采集完成后进行预处理,本公开中分别针对每一类异构数据进行预处理,预处理完成后,将每一类异构数据输入至一特征提取网络,特征提取网络包括连接的CNN网络模型和BiLSTM网络模型,通过特征提取网络输出提取的每一类异构数据的特征;关键信息预测模块包括注意力机制网络和全连接层,将特征提取网络输出的四类异构数据的特征分别输入至一注意力机制网络,输出结果输入全连接层,输出每类异构数据的关键信息特征。四类异构数据的关键信息特征输入特征融合模块104进行特征融合,融合完成后将融合特征输入功率预测模块105进行预测。在本公开中,通过将历史数据输入构建的模型中,对网络模型进行训练,若通过输入历史数据输出的风电集群功率 数据与实际数据一致,则训练完成。
S130:将实时的风电集群异构数据预处理后输入训练完成的所述风电集群功率预测网络模型中,输出结果作为风电集群功率预测结果。
具体的,本公开从SCADA系统中选取5年的历史数据及实时采集的文本数据,通过以上步骤计算子集群内输出功率预测,整个集群功率则将子集群进行和运算即可。本公开为场站区域出力的预测对电力系统制定调度计划、旋转备用容量都具有重要的意义。
若得到发电设备异常检测结果,则将异常检测结果进行展示;展示方式至少包括:文本显示、语音播报、外呼终端、邮件、短信提醒、智能音箱。
为了实现上述实施例,本公开还提出一种基于深度学习的风电集群功率预测装置,如图3所示,包括数据获取模块310、模型构建模块320和功率预测模块330。
数据获取模块310,用于获取风电集群历史异构数据并进行数据预处理,将预处理后的风电集群异构历史数据作为训练集;
模型构建模块320,用于构建风电集群功率预测网络模型,并通过所述训练集对构建的所述风电集群功率预测网络模型进行训练;其中,所述风电集群功率预测网络模型包括依序连接的特征提取模块、关键信息预测模块、特征融合模块和结果预测模块;
功率预测模块330,用于将实时的风电集群异构数据预处理后输入训练完成的所述风电集群功率预测网络模型中,输出结果作为风电集群功率预测结果。
为了实现上述实施例,本公开还提出一种计算机设备,包括:处理器用于存储所述处理器可执行计算机程序的存储器,处理器执行计算机程序时,实现如本公开实施例的基于深度学习的风电集群功率预测方法。
如图4所示,非临时性计算机可读存储介质包括指令的存储器410,接口430,上述指令可由基于深度学习的风电集群功率预测装置的处理器420执行以完成上述方法。可选地,存储介质可以是非临时性计算机可读存储介质,例如,非临时性计算机可读存储介质可以是ROM、随机存取存储器(RAM)、CD-ROM、磁带、软盘和光数据存储设备等。
为了实现上述实施例,本公开还提出一种非临时性计算机可读存储介质,其上存储有计算机程序,计算机程序被处理器执行时实现如本公开实施例的基于深度学习的风电集群功率预测方法。
为了实现上述实施例,本公开还提出一种计算机程序产品,包括计算机程序,其中,所述计算机程序被处理器执行时实现如本公开实施例的基于深度学习的风电集群功率预测方法。
在本说明书的描述中,参考术语“一个实施例”、“一些实施例”、“示例”、“具体示例”、或“一些示例”等的描述意指结合该实施例或示例描述的具体特征、结构、材料或者特点包含 于本公开的至少一个实施例或示例中。在本说明书中,对上述术语的示意性表述不必须针对的是相同的实施例或示例。而且,描述的具体特征、结构、材料或者特点可以在任一个或多个实施例或示例中以合适的方式结合。此外,在不相互矛盾的情况下,本领域的技术人员可以将本说明书中描述的不同实施例或示例以及不同实施例或示例的特征进行结合和组合。
此外,术语“第一”、“第二”仅用于描述目的,而不能理解为指示或暗示相对重要性或者隐含指明所指示的技术特征的数量。由此,限定有“第一”、“第二”的特征可以明示或者隐含地包括至少一个该特征。在本公开的描述中,“多个”的含义是至少两个,例如两个,三个等,除非另有明确具体的限定。
流程图中或在此以其他方式描述的任何过程或方法描述可以被理解为,表示包括一个或更多个用于实现定制逻辑功能或过程的步骤的可执行指令的代码的模块、片段或部分,并且本公开的优选实施方式的范围包括另外的实现,其中可以不按所示出或讨论的顺序,包括根据所涉及的功能按基本同时的方式或按相反的顺序,来执行功能,这应被本公开的实施例所属技术领域的技术人员所理解。
在流程图中表示或在此以其他方式描述的逻辑和/或步骤,例如,可以被认为是用于实现逻辑功能的可执行指令的定序列表,可以具体实现在任何计算机可读介质中,以供指令执行系统、装置或设备(如基于计算机的系统、包括处理器的系统或其他可以从指令执行系统、装置或设备取指令并执行指令的系统)使用,或结合这些指令执行系统、装置或设备而使用。就本说明书而言,"计算机可读介质"可以是任何可以包含、存储、通信、传播或传输程序以供指令执行系统、装置或设备或结合这些指令执行系统、装置或设备而使用的装置。计算机可读介质的更具体的示例(非穷尽性列表)包括以下:具有一个或多个布线的电连接部(电子装置),便携式计算机盘盒(磁装置),随机存取存储器(RAM),只读存储器(ROM),可擦除可编辑只读存储器(EPROM或闪速存储器),光纤装置,以及便携式光盘只读存储器(CDROM)。另外,计算机可读介质甚至可以是可在其上打印所述程序的纸或其他合适的介质,因为可以例如通过对纸或其他介质进行光学扫描,接着进行编辑、解译或必要时以其他合适方式进行处理来以电子方式获得所述程序,然后将其存储在计算机存储器中。
应当理解,本公开的各部分可以用硬件、软件、固件或它们的组合来实现。在上述实施方式中,多个步骤或方法可以用存储在存储器中且由合适的指令执行系统执行的软件或固件来实现。如,如果用硬件来实现和在另一实施方式中一样,可用本领域公知的下列技术中的任一项或他们的组合来实现:具有用于对数据信号实现逻辑功能的逻辑门电路的离散逻辑电路,具有合适的组合逻辑门电路的专用集成电路,可编程门阵列(PGA),现场可编程门阵列(FPGA)等。
本技术领域的普通技术人员可以理解实现上述实施例方法携带的全部或部分步骤是可 以通过程序来指令相关的硬件完成,所述的程序可以存储于一种计算机可读存储介质中,该程序在执行时,包括方法实施例的步骤之一或其组合。
此外,在本公开各个实施例中的各功能单元可以集成在一个处理模块中,也可以是各个单元单独物理存在,也可以两个或两个以上单元集成在一个模块中。上述集成的模块既可以采用硬件的形式实现,也可以采用软件功能模块的形式实现。所述集成的模块如果以软件功能模块的形式实现并作为独立的产品销售或使用时,也可以存储在一个计算机可读取存储介质中。
上述提到的存储介质可以是只读存储器,磁盘或光盘等。尽管上面已经示出和描述了本公开的实施例,可以理解的是,上述实施例是示例性的,不能理解为对本公开的限制,本领域的普通技术人员在本公开的范围内可以对上述实施例进行变化、修改、替换和变型。
Claims (19)
- 一种基于深度学习的风电集群功率预测方法,包括:获取风电集群历史异构数据并进行数据预处理,将预处理后的风电集群异构历史数据作为训练集;构建风电集群功率预测网络模型,并通过所述训练集对构建的所述风电集群功率预测网络模型进行训练;和将实时的风电集群异构数据预处理后输入训练完成的所述风电集群功率预测网络模型中,输出结果作为风电集群功率预测结果。
- 根据权利要求2所述的基于深度学习的风电集群功率预测方法,其中,所述风电集群功率预测网络模型包括特征提取模块、关键信息预测模块、特征融合模块和结果预测模块;其中,所述特征提取模块,为特征提取神经网络,用于对数据预处理后的风电集群历史异构数据进行特征提取;所述关键信息预测模块,用于得到风电集群历史异构数据各自特征内部交互特征及数据间的关联性特征;所述特征融合模块,用于对风电集群历史异构数据的特征进行特征融合拼接,得到多模态特征融合信息;所述结果预测模块,用于根据所述特征融合信息计算预测结果,完成风电集群功率预测。
- 根据权利要求3所述的基于深度学习的风电集群功率预测方法,其中,所述特征提取神经网络为CNN+BiLSTM结合神经网络;借助CNN擅长获得时空数据特征及结合BiLSTM对时序数据获得前后向序列,从而完成异构数据的特征提取。
- 根据权利要求3所述的基于深度学习的风电集群功率预测方法,其中,所述关键信 息预测模块利用注意力机制,得到风电集群历史异构数据各自特征内部交互特征及数据间的关联性特征;所述特征融合模块通过合并所述风电集群历史异构数据的特征,得到含有上下文的时空互补性及关联性特点的多模态融合特征。
- 根据权利要求3所述的基于深度学习的风电集群功率预测方法,其中,所述结果预测模块采用全连接层计算预测结果,采用激活函数ReLU函数作为所述全连接层的激活函数,计算得到预测结果用归一化还原函数计算公式(2)而获得还原后的功率预测值:w o=w pre(w max-w min)+w min (2)其中W pre代表所述风电集群功率预测网络模型的预测输出值,W o代表还原后的功率预测值,w max代表所选风电集群历史异构数据中的样本最大值,w min代表所选风电集群历史异构数据中的样本最小值。
- 根据权利要求4所述的基于深度学习的风电集群功率预测方法,其中,通过所述训练集对构建的所述风电集群功率预测网络模型进行训练的步骤包括:将预处理后的训练集数据输入所述特征提取模块的特征提取神经网络,借助CNN擅长获得时空数据特征及结合BiLSTM对时序数据获得前后向序列,从而完成异构数据的特征提取;利用注意力机制得到异构数据各自特征内部交互特征及数据间的关联性特征,输入关键信息预测模块的全连接层,得到预测关键特征;将预测关键特征进行合并操作,得到含有上下文时空互补性和关联性特点的融合特征,输入功率预测模块,计算预测结果,与实际风电集群功率结果进行对比,以均方误差作为损失函数,网络训练优化采用Adam算法,通过不断调整网络函数和参数,直至预测结果与标记的功率结果一致时,完成网络训练。
- 根据权利要求1所述的基于深度学习的风电集群功率预测方法,在获取风电集群功率预测结果之后,还包括将预测结果进行展示的步骤;展示方式包括以下各项中的至少一种:文本显示、语音播报、外呼终端、邮件、短信提醒、智能音箱。
- 一种基于深度学习的风电集群功率预测装置,包括:数据获取模块,用于获取风电集群历史异构数据并进行数据预处理,将预处理后的风电集群异构历史数据作为训练集;模型构建模块,用于构建风电集群功率预测网络模型,并通过所述训练集对构建的所述风电集群功率预测网络模型进行训练;和功率预测模块,用于将实时的风电集群异构数据预处理后输入训练完成的所述风电集群功率预测网络模型中,输出结果作为风电集群功率预测结果。
- 一种计算机设备,包括:处理器;和用于存储所述处理器可执行计算机程序的存储器,其中,所述处理器被配置为执行所述计算机程序,实现以下步骤:获取风电集群历史异构数据并进行数据预处理,将预处理后的风电集群异构历史数据作为训练集;构建风电集群功率预测网络模型,并通过所述训练集对构建的所述风电集群功率预测网络模型进行训练;和将实时的风电集群异构数据预处理后输入训练完成的所述风电集群功率预测网络模型中,输出结果作为风电集群功率预测结果。
- 根据权利要求11所述的计算机设备,其中,所述风电集群功率预测网络模型包括特征提取模块、关键信息预测模块、特征融合模块和结果预测模块;其中,所述特征提取模块,为特征提取神经网络,用于对数据预处理后的风电集群历史异构数据进行特征提取;所述关键信息预测模块,用于得到风电集群历史异构数据各自特征内部交互特征及数据间的关联性特征;所述特征融合模块,用于对风电集群历史异构数据的特征进行特征融合拼接,得到多模态特征融合信息;所述结果预测模块,用于根据所述特征融合信息计算预测结果,完成风电集群功率预测。
- 根据权利要求12所述的计算机设备,其中,所述特征提取神经网络为CNN+BiLSTM结合神经网络;借助CNN擅长获得时空数据特征及结合BiLSTM对时序数据获得前后向序列,从而完成异构数据的特征提取。
- 根据权利要求12所述的计算机设备,其中,所述关键信息预测模块利用注意力机制,得到风电集群历史异构数据各自特征内部交互特征及数据间的关联性特征;所述特征融 合模块通过合并所述风电集群历史异构数据的特征,得到含有上下文的时空互补性及关联性特点的多模态融合特征。
- 根据权利要求12所述的计算机设备,其中,所述结果预测模块采用全连接层计算预测结果,采用激活函数ReLU函数作为所述全连接层的激活函数,计算得到预测结果用归一化还原函数计算公式(2)而获得还原后的功率预测值:w o=w pre(w max-w min)+w min (2)其中W pre代表所述风电集群功率预测网络模型的预测输出值,W o代表还原后的功率预测值,w max代表所选风电集群历史异构数据中的样本最大值,w min代表所选风电集群历史异构数据中的样本最小值。
- 根据权利要求13所述的计算机设备,其中,所述处理器还被配置为:将预处理后的训练集数据输入所述特征提取模块的特征提取神经网络,借助CNN擅长获得时空数据特征及结合BiLSTM对时序数据获得前后向序列,从而完成异构数据的特征提取;利用注意力机制得到异构数据各自特征内部交互特征及数据间的关联性特征,输入关键信息预测模块的全连接层,得到预测关键特征;将预测关键特征进行合并操作,得到含有上下文时空互补性和关联性特点的融合特征,输入功率预测模块,计算预测结果,与实际风电集群功率结果进行对比,以均方误差作为损失函数,网络训练优化采用Adam算法,通过不断调整网络函数和参数,直至预测结果与标记的功率结果一致时,完成网络训练。
- 根据权利要求10所述的计算机设备,其中,所述处理器还被配置为:将预测结果进行展示;展示方式包括以下各项中的至少一种:文本显示、语音播报、外呼终端、邮件、短信提醒、智能音箱。
- 一种非临时性计算机可读存储介质,其上存储有计算机程序,其中,所述计算机程序被处理器执行时实现如权利要求1至8中任一项所述的方法。
- 一种计算机程序产品,包括计算机程序,其中,所述计算机程序被处理器执行时实现如权利要求1至8中任一项所述的方法。
Applications Claiming Priority (2)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN202111421723.5 | 2021-11-26 | ||
CN202111421723.5A CN114330097A (zh) | 2021-11-26 | 2021-11-26 | 一种异构数据与深度学习的风电集群功率预测方法及设备 |
Publications (1)
Publication Number | Publication Date |
---|---|
WO2023093774A1 true WO2023093774A1 (zh) | 2023-06-01 |
Family
ID=81047058
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
PCT/CN2022/133796 WO2023093774A1 (zh) | 2021-11-26 | 2022-11-23 | 基于深度学习的风电集群功率预测方法 |
Country Status (2)
Country | Link |
---|---|
CN (1) | CN114330097A (zh) |
WO (1) | WO2023093774A1 (zh) |
Cited By (12)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN116822370A (zh) * | 2023-07-11 | 2023-09-29 | 东南大学溧阳研究院 | 数据驱动下风电集群超短期出力预测方法 |
CN117013534A (zh) * | 2023-09-26 | 2023-11-07 | 宁德时代新能源科技股份有限公司 | 发电功率预测方法、功率预测模型训练方法、装置和设备 |
CN117110700A (zh) * | 2023-08-23 | 2023-11-24 | 易集康健康科技(杭州)有限公司 | 一种射频电源脉冲功率检测方法及系统 |
CN117394308A (zh) * | 2023-09-24 | 2024-01-12 | 中国华能集团清洁能源技术研究院有限公司 | 一种多时间尺度风力发电功率预测方法、系统及电子设备 |
CN117498320A (zh) * | 2023-11-01 | 2024-02-02 | 南方电网电力科技股份有限公司 | 一种波浪发电功率预测方法和系统 |
CN117578428A (zh) * | 2023-11-22 | 2024-02-20 | 广西电网有限责任公司 | 风机功率递归预测方法、装置、计算机设备和存储介质 |
CN117708716A (zh) * | 2024-02-05 | 2024-03-15 | 敏博科技(武汉)有限公司 | 基于回归和时序融合的光伏发电功率/量预测方法及设备 |
CN117748482A (zh) * | 2023-12-13 | 2024-03-22 | 正泰集团研发中心(上海)有限公司 | 发电功率预测方法、装置、电子设备及存储介质 |
CN118094469A (zh) * | 2024-02-22 | 2024-05-28 | 国网智能电网研究院有限公司 | 多源异构数据的融合方法、装置、设备及存储介质 |
CN118195276A (zh) * | 2024-05-16 | 2024-06-14 | 绍兴柯桥春生水务科技有限公司 | 一种区域水资源优化配置方法及系统 |
CN118208374A (zh) * | 2024-04-19 | 2024-06-18 | 华电(福建)风电有限公司 | 用于海上风电的工况外载加载一体化监测系统与方法 |
CN118428493A (zh) * | 2024-07-05 | 2024-08-02 | 杭州数亮科技股份有限公司 | 基于机器学习的人工智能数据分析方法及系统 |
Families Citing this family (4)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN114330495A (zh) * | 2021-11-26 | 2022-04-12 | 中国华能集团清洁能源技术研究院有限公司 | 一种基于深度学习融合模型的风电功率预测方法及设备 |
CN114330097A (zh) * | 2021-11-26 | 2022-04-12 | 中国华能集团清洁能源技术研究院有限公司 | 一种异构数据与深度学习的风电集群功率预测方法及设备 |
CN114970980A (zh) * | 2022-05-07 | 2022-08-30 | 哈尔滨工业大学(深圳) | 一种基于深度学习的风电功率预测方法及系统 |
CN115204522B (zh) * | 2022-09-13 | 2023-01-24 | 华北电力大学 | 区域发电功率预测方法、装置、设备及存储介质 |
Citations (4)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN106875033A (zh) * | 2016-12-26 | 2017-06-20 | 华中科技大学 | 一种基于动态自适应的风电集群功率预测方法 |
US20190178229A1 (en) * | 2016-08-02 | 2019-06-13 | Wobben Properties Gmbh | Method for sending a controller setpoint for a power generator, device therefor and system containing the same |
CN112529282A (zh) * | 2020-12-03 | 2021-03-19 | 清华大学 | 基于时空图卷积神经网络的风电场集群短期功率预测方法 |
CN114330097A (zh) * | 2021-11-26 | 2022-04-12 | 中国华能集团清洁能源技术研究院有限公司 | 一种异构数据与深度学习的风电集群功率预测方法及设备 |
-
2021
- 2021-11-26 CN CN202111421723.5A patent/CN114330097A/zh active Pending
-
2022
- 2022-11-23 WO PCT/CN2022/133796 patent/WO2023093774A1/zh unknown
Patent Citations (4)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
US20190178229A1 (en) * | 2016-08-02 | 2019-06-13 | Wobben Properties Gmbh | Method for sending a controller setpoint for a power generator, device therefor and system containing the same |
CN106875033A (zh) * | 2016-12-26 | 2017-06-20 | 华中科技大学 | 一种基于动态自适应的风电集群功率预测方法 |
CN112529282A (zh) * | 2020-12-03 | 2021-03-19 | 清华大学 | 基于时空图卷积神经网络的风电场集群短期功率预测方法 |
CN114330097A (zh) * | 2021-11-26 | 2022-04-12 | 中国华能集团清洁能源技术研究院有限公司 | 一种异构数据与深度学习的风电集群功率预测方法及设备 |
Cited By (16)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN116822370B (zh) * | 2023-07-11 | 2024-02-20 | 东南大学溧阳研究院 | 数据驱动下风电集群超短期出力预测方法 |
CN116822370A (zh) * | 2023-07-11 | 2023-09-29 | 东南大学溧阳研究院 | 数据驱动下风电集群超短期出力预测方法 |
CN117110700A (zh) * | 2023-08-23 | 2023-11-24 | 易集康健康科技(杭州)有限公司 | 一种射频电源脉冲功率检测方法及系统 |
CN117110700B (zh) * | 2023-08-23 | 2024-06-04 | 易集康健康科技(杭州)有限公司 | 一种射频电源脉冲功率检测方法及系统 |
CN117394308A (zh) * | 2023-09-24 | 2024-01-12 | 中国华能集团清洁能源技术研究院有限公司 | 一种多时间尺度风力发电功率预测方法、系统及电子设备 |
CN117013534A (zh) * | 2023-09-26 | 2023-11-07 | 宁德时代新能源科技股份有限公司 | 发电功率预测方法、功率预测模型训练方法、装置和设备 |
CN117013534B (zh) * | 2023-09-26 | 2024-02-20 | 宁德时代新能源科技股份有限公司 | 发电功率预测方法、功率预测模型训练方法、装置和设备 |
CN117498320A (zh) * | 2023-11-01 | 2024-02-02 | 南方电网电力科技股份有限公司 | 一种波浪发电功率预测方法和系统 |
CN117578428A (zh) * | 2023-11-22 | 2024-02-20 | 广西电网有限责任公司 | 风机功率递归预测方法、装置、计算机设备和存储介质 |
CN117748482A (zh) * | 2023-12-13 | 2024-03-22 | 正泰集团研发中心(上海)有限公司 | 发电功率预测方法、装置、电子设备及存储介质 |
CN117708716A (zh) * | 2024-02-05 | 2024-03-15 | 敏博科技(武汉)有限公司 | 基于回归和时序融合的光伏发电功率/量预测方法及设备 |
CN117708716B (zh) * | 2024-02-05 | 2024-05-10 | 敏博科技(武汉)有限公司 | 基于回归和时序融合的光伏发电功率预测方法及设备 |
CN118094469A (zh) * | 2024-02-22 | 2024-05-28 | 国网智能电网研究院有限公司 | 多源异构数据的融合方法、装置、设备及存储介质 |
CN118208374A (zh) * | 2024-04-19 | 2024-06-18 | 华电(福建)风电有限公司 | 用于海上风电的工况外载加载一体化监测系统与方法 |
CN118195276A (zh) * | 2024-05-16 | 2024-06-14 | 绍兴柯桥春生水务科技有限公司 | 一种区域水资源优化配置方法及系统 |
CN118428493A (zh) * | 2024-07-05 | 2024-08-02 | 杭州数亮科技股份有限公司 | 基于机器学习的人工智能数据分析方法及系统 |
Also Published As
Publication number | Publication date |
---|---|
CN114330097A (zh) | 2022-04-12 |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
WO2023093774A1 (zh) | 基于深度学习的风电集群功率预测方法 | |
WO2023093010A1 (zh) | 基于深度学习融合模型的风电功率预测方法及设备 | |
WO2023087525A1 (zh) | 多模态深度学习的发电设备异常一体化识别方法及设备 | |
CN111143576A (zh) | 一种面向事件的动态知识图谱构建方法和装置 | |
WO2021208696A1 (zh) | 用户意图分析方法、装置、电子设备及计算机存储介质 | |
CN115879535A (zh) | 一种自动驾驶感知模型的训练方法、装置、设备和介质 | |
CN110807566A (zh) | 人工智能模型评测方法、装置、设备及存储介质 | |
CN112434718B (zh) | 基于深度图的新冠肺炎多模态特征提取融合方法及系统 | |
Chan et al. | Question-answering dialogue system for emergency operations | |
CN112002410A (zh) | 传染病疾病状态预测方法及装置、存储介质、电子设备 | |
CN110110095A (zh) | 一种基于长短期记忆循环神经网络的电力指令文本匹配方法 | |
Malakouti | Improving the prediction of wind speed and power production of SCADA system with ensemble method and 10-fold cross-validation | |
CN117236676A (zh) | 一种基于多模态事件抽取的rpa流程挖掘方法和装置 | |
CN114647732B (zh) | 一种面向弱监督文本分类系统、方法和装置 | |
CN113220951B (zh) | 一种基于智能内容的医学临床支持方法和系统 | |
CN113435582B (zh) | 基于句向量预训练模型的文本处理方法及相关设备 | |
CN110032643A (zh) | 一种建筑维修工单分析方法、装置、存储介质及客户端 | |
CN113111135A (zh) | 一种知识图谱构建方法及装置 | |
CN114358371A (zh) | 一种基于深度学习的光伏短期功率预测方法及设备 | |
Jia et al. | STCM-Net: A symmetrical one-stage network for temporal language localization in videos | |
CN115062617A (zh) | 基于提示学习的任务处理方法、装置、设备及介质 | |
CN114330493A (zh) | 一种CNN+BiLSTM+Attention风电超短期功率预测方法及系统 | |
Ajagunsegun et al. | Machine learning-based system for managing energy efficiency of public buildings: An approach towards smart cities | |
CN112700862A (zh) | 目标科室的确定方法、装置、电子设备及存储介质 | |
Qin | Research on the application of intelligent speech recognition technology in medical big data fog computing system |
Legal Events
Date | Code | Title | Description |
---|---|---|---|
121 | Ep: the epo has been informed by wipo that ep was designated in this application |
Ref document number: 22897857 Country of ref document: EP Kind code of ref document: A1 |
|
NENP | Non-entry into the national phase |
Ref country code: DE |