CN113240217A - Photovoltaic power generation prediction method and device based on integrated prediction model - Google Patents
Photovoltaic power generation prediction method and device based on integrated prediction model Download PDFInfo
- Publication number
- CN113240217A CN113240217A CN202110784293.7A CN202110784293A CN113240217A CN 113240217 A CN113240217 A CN 113240217A CN 202110784293 A CN202110784293 A CN 202110784293A CN 113240217 A CN113240217 A CN 113240217A
- Authority
- CN
- China
- Prior art keywords
- neural network
- network model
- model
- photovoltaic power
- prediction
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Pending
Links
- 238000000034 method Methods 0.000 title claims abstract description 95
- 238000010248 power generation Methods 0.000 title claims abstract description 68
- 238000013528 artificial neural network Methods 0.000 claims abstract description 68
- 238000003062 neural network model Methods 0.000 claims abstract description 64
- 238000013527 convolutional neural network Methods 0.000 claims abstract description 55
- 230000006870 function Effects 0.000 claims description 31
- 238000005457 optimization Methods 0.000 claims description 30
- ORILYTVJVMAKLC-UHFFFAOYSA-N Adamantane Natural products C1C(C2)CC3CC1CC2C3 ORILYTVJVMAKLC-UHFFFAOYSA-N 0.000 claims description 22
- 238000011176 pooling Methods 0.000 claims description 16
- 238000002474 experimental method Methods 0.000 claims description 13
- 238000012545 processing Methods 0.000 claims description 9
- 238000004590 computer program Methods 0.000 claims description 7
- 210000002569 neuron Anatomy 0.000 claims description 7
- 230000004044 response Effects 0.000 claims description 7
- 238000012935 Averaging Methods 0.000 claims description 5
- 230000010354 integration Effects 0.000 claims description 5
- 238000010606 normalization Methods 0.000 claims description 4
- 238000007781 pre-processing Methods 0.000 claims description 4
- 150000001875 compounds Chemical class 0.000 claims description 2
- 238000007637 random forest analysis Methods 0.000 claims description 2
- 238000011425 standardization method Methods 0.000 claims description 2
- 238000012549 training Methods 0.000 description 21
- 230000000694 effects Effects 0.000 description 10
- 238000004422 calculation algorithm Methods 0.000 description 9
- 238000010586 diagram Methods 0.000 description 7
- 230000004913 activation Effects 0.000 description 5
- 230000008569 process Effects 0.000 description 3
- 230000005855 radiation Effects 0.000 description 3
- 229910052799 carbon Inorganic materials 0.000 description 2
- 238000013136 deep learning model Methods 0.000 description 2
- 230000005611 electricity Effects 0.000 description 2
- 238000011156 evaluation Methods 0.000 description 2
- 238000011160 research Methods 0.000 description 2
- 239000002699 waste material Substances 0.000 description 2
- OKTJSMMVPCPJKN-UHFFFAOYSA-N Carbon Chemical compound [C] OKTJSMMVPCPJKN-UHFFFAOYSA-N 0.000 description 1
- 230000002159 abnormal effect Effects 0.000 description 1
- 230000006978 adaptation Effects 0.000 description 1
- 230000009286 beneficial effect Effects 0.000 description 1
- 238000006243 chemical reaction Methods 0.000 description 1
- 239000003245 coal Substances 0.000 description 1
- 238000004891 communication Methods 0.000 description 1
- 230000007423 decrease Effects 0.000 description 1
- 238000013135 deep learning Methods 0.000 description 1
- 238000011161 development Methods 0.000 description 1
- 238000003912 environmental pollution Methods 0.000 description 1
- 230000007246 mechanism Effects 0.000 description 1
- 238000010295 mobile communication Methods 0.000 description 1
- 238000012986 modification Methods 0.000 description 1
- 230000004048 modification Effects 0.000 description 1
- 238000006386 neutralization reaction Methods 0.000 description 1
- 239000003208 petroleum Substances 0.000 description 1
- 239000007787 solid Substances 0.000 description 1
- 238000006467 substitution reaction Methods 0.000 description 1
- 238000012360 testing method Methods 0.000 description 1
- 230000009466 transformation Effects 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
- G06N3/045—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
- G06Q10/00—Administration; Management
- G06Q10/06—Resources, workflows, human or project management; Enterprise or organisation planning; Enterprise or organisation modelling
- G06Q10/067—Enterprise or organisation modelling
-
- 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)
- Business, Economics & Management (AREA)
- Physics & Mathematics (AREA)
- Theoretical Computer Science (AREA)
- Strategic Management (AREA)
- Human Resources & Organizations (AREA)
- Economics (AREA)
- General Physics & Mathematics (AREA)
- Health & Medical Sciences (AREA)
- Tourism & Hospitality (AREA)
- Marketing (AREA)
- General Health & Medical Sciences (AREA)
- General Business, Economics & Management (AREA)
- Entrepreneurship & Innovation (AREA)
- Software Systems (AREA)
- Life Sciences & Earth Sciences (AREA)
- General Engineering & Computer Science (AREA)
- Computing Systems (AREA)
- Molecular Biology (AREA)
- Evolutionary Computation (AREA)
- Quality & Reliability (AREA)
- Operations Research (AREA)
- Data Mining & Analysis (AREA)
- Computational Linguistics (AREA)
- Mathematical Physics (AREA)
- Biophysics (AREA)
- Biomedical Technology (AREA)
- Artificial Intelligence (AREA)
- Development Economics (AREA)
- Game Theory and Decision Science (AREA)
- Primary Health Care (AREA)
- Water Supply & Treatment (AREA)
- Public Health (AREA)
- Educational Administration (AREA)
- Supply And Distribution Of Alternating Current (AREA)
Abstract
The invention discloses a photovoltaic power generation prediction method and a photovoltaic power generation prediction device based on an integrated prediction model, wherein the method comprises the following steps: responding to the acquired historical data of the photovoltaic power station, and constructing a neural network model, wherein the neural network model comprises an artificial neural network model, a deep neural network model and a convolutional neural network model; constructing an integrated prediction model based on a base learner which respectively uses an artificial neural network model, a deep neural network model and a convolutional neural network model as Bagging methods; and responding to the acquired real-time data of the photovoltaic power station, and outputting a prediction result of the photovoltaic power generation power based on the integrated prediction model. An artificial neural network model, a deep neural network model and a convolutional neural network model are respectively used as base learners of the Bagging method to construct an integrated prediction model, and the real-time data of the photovoltaic power station outputs a prediction result of photovoltaic power generation power based on the integrated prediction model, so that the prediction precision of photovoltaic power generation is improved.
Description
Technical Field
The invention belongs to the technical field of photovoltaic power generation, and particularly relates to a photovoltaic power generation prediction method and device based on an integrated prediction model.
Background
Compared with traditional fossil energy such as coal, petroleum and the like, the solar energy represented energy has the characteristics of less pollution, renewability and the like, the development strength of the clean energy is increased, the large-scale utilization of the clean energy is realized, the solar energy has important significance for relieving the problems of serious environmental pollution and resource exhaustion in the world at present, and the solar energy becomes a work key point of the current power industry.
In order to realize the double-carbon target (the target of carbon neutralization in 2030 and 2060), a high-efficiency safe energy system is built, renewable energy is vigorously developed, and clean transformation of energy is promoted to be an effective measure. Solar energy is a typical renewable energy source and has high research value, but photovoltaic power generation is greatly influenced by climate and external environment, and the output of the photovoltaic power generation has the characteristic of high randomness. With the fact that the proportion of the photovoltaic in the power grid is larger and larger, in order to better connect the photovoltaic power generation into the large power grid, reduce the impact of the photovoltaic power generation on the power grid, improve the safe and accurate scheduling and electric energy quality of the power grid, and accurately predict the output of the photovoltaic power generation, the method is very important.
Disclosure of Invention
The invention provides a photovoltaic power generation prediction method and device based on an integrated prediction model, which are used for solving at least one of the technical problems.
In a first aspect, the invention provides a photovoltaic power generation prediction method based on an integrated prediction model, which comprises the following steps: step 1, responding to acquired photovoltaic power station historical data, and constructing a neural network model, wherein the neural network model comprises an artificial neural network model, a deep neural network model and a convolutional neural network model, and the constructing of the neural network model specifically comprises the following steps: step 1.1, constructing an artificial neural network model, searching and performing multiple experiments by a grid search method to obtain an optimal hyper-parameter required by modeling of the artificial neural network model, finally determining the number of nodes of a hidden layer, inputting the number of nodes and outputting the optimal hyper-parameter combination of the number of nodes, selecting an Adam optimization function, loading a first sequential model into the artificial neural network model, and inputting a layer into the artificial neural network model one by an add () method; step 1.2, constructing a deep neural network model, searching and performing multiple experiments by a grid search method to obtain an optimal hyper-parameter required by model modeling, finally determining the optimal hyper-parameter combination of the number of hidden layer nodes, the number of input nodes and the number of output nodes, selecting an Adam optimization function, loading a second sequential model in the deep neural network model, and inputting layer into the deep neural network model one by an add () method; step 1.3, constructing a convolutional neural network model, setting the number of layers of a convolutional layer, a pooling layer and a full-link layer, simultaneously setting related parameters of corresponding convolutional cores, pooling windows and the number of neurons, setting a flatten layer for a connection part of the pooling layer and the full-link layer for feature extension, finally setting the number of nodes of a hidden layer, the number of input nodes and the number of output nodes when the convolutional neural network model feeds data, selecting an Adam optimization function, loading the Adam optimization function into a third sequential model, and inputting layer into the convolutional neural network model one by one through an add () method; step 2, constructing an integrated prediction model based on the artificial neural network model, the deep neural network model and the convolutional neural network model which are respectively used as basis learners of a Bagging method; and 3, responding to the acquired real-time data of the photovoltaic power station, and outputting a prediction result of the photovoltaic power generation power based on the integrated prediction model.
In a second aspect, the present invention provides a photovoltaic power generation prediction apparatus based on an integrated prediction model, including: the building module is configured to build a neural network model in response to the acquired photovoltaic power station historical data, wherein the neural network model comprises an artificial neural network model, a deep neural network model and a convolutional neural network model, and the building of the neural network model specifically comprises the following steps: constructing an artificial neural network model, searching and performing multiple experiments by a grid search method to obtain an optimal hyper-parameter required by modeling of the artificial neural network model, finally determining the optimal hyper-parameter combination of the number of hidden layer nodes, the number of input nodes and the number of output nodes, selecting an Adam optimization function, loading a first sequential model into the artificial neural network model, and inputting layer into the artificial neural network model one by an add () method; constructing a deep neural network model, searching and performing multiple experiments by a grid search method to obtain an optimal hyper-parameter required by model modeling, finally determining an optimal hyper-parameter combination of a hidden layer node number, an input node number and an output node number, selecting an Adam optimization function, loading a second sequential model in the deep neural network model, and inputting a layer into the deep neural network model one by an add () method; constructing a convolutional neural network model, setting the number of layers of a convolutional layer, a pooling layer and a full-link layer, setting related parameters of the number of corresponding convolutional cores, pooling windows and neurons, setting a flat layer for a connection part of the pooling layer and the full-link layer for feature extension, finally setting the number of nodes of a hidden layer, the number of input nodes and the number of output nodes when the convolutional neural network model feeds data, selecting an Adam optimization function, loading the Adam optimization function into a third sequential model, and inputting layer into the convolutional neural network model one by one through an add () method; an integration module configured to construct an integrated prediction model based on the artificial neural network model, the deep neural network model, and the convolutional neural network model as basis learners of a Bagging method, respectively; and the output module is configured to respond to the acquired real-time data of the photovoltaic power station and output a prediction result of the photovoltaic power generation power based on the integrated prediction model.
In a third aspect, an electronic device is provided, comprising: the photovoltaic power generation prediction method comprises at least one processor and a memory which is in communication connection with the at least one processor, wherein the memory stores instructions executable by the at least one processor, and the instructions are executed by the at least one processor to enable the at least one processor to perform the steps of the integrated prediction model-based photovoltaic power generation prediction method according to any embodiment of the invention.
In a fourth aspect, the present invention also provides a computer program product comprising a computer program stored on a non-transitory computer readable storage medium, the computer program comprising program instructions which, when executed by a computer, cause the computer to perform the steps of the integrated prediction model based photovoltaic power generation prediction method of any of the embodiments of the present invention.
According to the photovoltaic power generation prediction method and device based on the integrated prediction model, the artificial neural network model, the deep neural network model and the convolutional neural network model are respectively used as base learners of a Bagging method to construct the integrated prediction model, the photovoltaic power station real-time data output the prediction result of photovoltaic power generation power based on the integrated prediction model, the prediction precision of photovoltaic power generation is improved, data processing, photovoltaic power generation prediction, modeling, optimization and other process references can be provided for power stations with poor photovoltaic power generation prediction capability, more waste of electric quantity and insufficient self power generation capability, and the cost control and data processing level of the power stations are improved.
Drawings
In order to more clearly illustrate the technical solutions of the embodiments of the present invention, the drawings needed to be used in the description of the embodiments are briefly introduced below, and it is obvious that the drawings in the following description are some embodiments of the present invention, and it is obvious for those skilled in the art to obtain other drawings based on the drawings without creative efforts.
Fig. 1 is a flowchart of a photovoltaic power generation prediction method based on an integrated prediction model according to an embodiment of the present invention;
fig. 2 is a graph of irradiance versus power of a photovoltaic power generation prediction method based on an integrated prediction model according to an embodiment of the present invention;
fig. 3 is a graph of a relationship between temperature and power of a photovoltaic power generation prediction method based on an integrated prediction model according to an embodiment of the present invention;
fig. 4 is a graph of a relationship between humidity and power of a photovoltaic power generation prediction method based on an integrated prediction model according to an embodiment of the present invention;
FIG. 5 is a diagram of a DNN model architecture according to an embodiment of the present invention;
FIG. 6 is a diagram of an ANN model architecture according to an embodiment of the present invention;
fig. 7 is a diagram of a CNN model structure according to an embodiment of the present invention;
FIG. 8 is a chart comparing accuracy of different optimization algorithms corresponding to DNN models provided by an example of the present invention.
Fig. 9 is a chart comparing accuracy of CNN models corresponding to different optimization algorithms provided by the embodiment of the present invention.
Fig. 10 is a block diagram of a photovoltaic power generation prediction apparatus based on an integrated prediction model according to an embodiment of the present invention;
fig. 11 is a schematic structural diagram of an electronic device according to an embodiment of the present invention.
Detailed Description
In order to make the objects, technical solutions and advantages of the embodiments of the present invention clearer, the technical solutions in the embodiments of the present invention will be clearly and completely described below with reference to the drawings in the embodiments of the present invention, and it is obvious that the described embodiments are some, but not all, embodiments of the present invention. All other embodiments, which can be derived by a person skilled in the art from the embodiments given herein without making any creative effort, shall fall within the protection scope of the present invention.
Referring to fig. 1, a flowchart of a photovoltaic power generation prediction method based on an integrated prediction model according to the present application is shown.
As shown in fig. 1, in step 1, in response to acquired photovoltaic power station historical data, a neural network model is constructed, where the neural network model includes an artificial neural network model, a deep neural network model, and a convolutional neural network model, and the constructing of the neural network model specifically includes the following steps: step 1.1, constructing an artificial neural network model, searching and performing multiple experiments by a grid search method to obtain an optimal hyper-parameter required by modeling of the artificial neural network model, finally determining the number of nodes of a hidden layer, inputting the number of nodes and outputting the optimal hyper-parameter combination of the number of nodes, selecting an Adam optimization function, loading a first sequential model into the artificial neural network model, and inputting a layer into the artificial neural network model one by an add () method; step 1.2, constructing a deep neural network model, searching and performing multiple experiments by a grid search method to obtain an optimal hyper-parameter required by model modeling, and finally determining an optimal hyper-parameter combination of a hidden layer node number, an input node number and an output node number, wherein an optimization algorithm is selection of an activation function, a second sequential model is loaded in the deep neural network model, and layers are input into the deep neural network model one by an add () method; step 1.3, constructing a convolutional neural network model, setting the number of layers of a convolutional layer, a pooling layer and a full-link layer, simultaneously setting related parameters of corresponding convolutional cores, pooling windows and the number of neurons, setting a flatten layer for a connection part of the pooling layer and the full-link layer for feature extension, finally setting the number of nodes of a hidden layer, the number of input nodes and the number of output nodes when the convolutional neural network model feeds data, loading the nodes into a third sequential model, and inputting layers into the convolutional neural network model one by one through an add () method.
In step 2, an integrated prediction model is constructed based on the artificial neural network model, the deep neural network model and the convolutional neural network model which are respectively used as basis learners of the Bagging method.
In the embodiment, the integrated prediction model is constructed based on the artificial neural network model, the deep neural network model and the convolutional neural network model, so that the problems that the neural network of the existing single prediction model has high variance, the result is not easy to reproduce, the result of the model is sensitive to the initialization parameter abnormity and the like are solved, the precision, the robustness and the generalization capability of the prediction model are improved, and the problem of model selection is solved to a certain extent.
In step 3, in response to the acquired real-time data of the photovoltaic power station, a prediction result of the photovoltaic power generation power is output based on the integrated prediction model.
In the method of the embodiment, an Artificial Neural Network (ANN) is adopted to build a 3-layer Artificial Neural Network model, a Deep Neural Network (DNN) is adopted to build an 8-layer Deep Neural Network model, a Convolutional Neural Network (CNN) is adopted to build a 7-layer Convolutional Neural Network model, all-connection modes are adopted to connect adjacent layers of the Neural Networks, the Artificial Neural Network model, the Deep Neural Network model and the Convolutional Neural Network model are respectively used as a base learner of a Bagging method to build an integrated prediction model, so that real-time data of a photovoltaic power station outputs a prediction result of photovoltaic power generation power based on the integrated prediction model, the prediction precision of photovoltaic power generation is improved, and data processing and photovoltaic power generation prediction can be provided for power stations with poor photovoltaic power generation prediction capability, more electricity waste and insufficient self power generation capability, And the modeling, the optimization and other processes are referred to, so that the cost control and data processing level of the system is improved.
Specifically, various methods for model building are fully considered, and a high-quality deep learning model is built for improving the prediction precision. And in the data set division stage, a self-help method is adopted, a grid search method is adopted for searching for the optimal batch _ size and epoch, an optimization algorithm Adam function is added at last, and a loss function is minimized by improving a training mode.
In some optional embodiments, before step 1, preprocessing the photovoltaic power plant historical data, the preprocessing comprising: filling missing variable values of characteristic data in the historical data of the photovoltaic power station based on a random forest; calculating and visualizing the importance of the feature data based on the XGBOOST method, selecting the required feature data, and performing data normalization processing on the related feature data, wherein the data normalization processing is a Z-score standardization method:
in the formula (I), the compound is shown in the specification,represents the mean of the total number of samples,the standard deviation of the sample is shown as,represents the observed value of the sample and,representing the normalized variables.
In some optional embodiments, the power generation history data comprises instantaneous power history data, and the meteorological history data comprises irradiance history data, ambient temperature history data, humidity history data, and wind speed history data.
In a specific embodiment, according to the photovoltaic power generation prediction method based on the integrated prediction model, data of a photovoltaic power station history are divided into a training set accounting for 70% and a testing set accounting for 30%, a weather feature is used for training a photovoltaic power prediction model in the research, the feature data are normalized to facilitate model adaptation, and then an integrated learning model is built.
On the basis of a traditional neural network-based generated power prediction model, an integrated learning model based on deep learning is constructed to predict a photovoltaic generated power system prototype. The power prediction system mainly comprises three modules, namely a data preprocessing module, three respective modeling modules of the base learners and an integrated prediction module.
When the photovoltaic power station generates electricity, historical data such as meteorological data and instantaneous power are collected by using a sensor installed in the system, the collected data generally belong to dirty data, and the common problems of missing values, abnormal values and other data exist, so that the original dirty data need to be preprocessed, and the data are cleaned, normalized and the like and then are arranged into a format required by a model. And then putting the three neural network models into the fitting training respectively, after the models are trained, inputting meteorological data needing to be predicted at a certain moment into the three neural network models respectively to obtain 3 outputs, and then using an averaging method to obtain a final predicted instantaneous power value for the 3 outputs. The instantaneous generated power is predicted by integrating models of three DNN, ANN, CNN neural networks and using this method.
As shown in fig. 2, 3 and 4, the influence of key factors influencing photovoltaic power output is discussed by using historical actual operation data of a photovoltaic power station of a certain solar research center. By selecting and analyzing typical characteristics with large influence on photovoltaic power generation, the relation between the input variable and the photovoltaic power generation power is mined, and a foundation is laid for power prediction. It can be seen from the figure that the solar radiation intensity (irradiation) and the photovoltaic power output have a positive correlation trend, and the larger the solar radiation intensity is, the more sufficient the photovoltaic panel receives the light, the larger the photovoltaic system output is. The air temperature is closely related to the radiation degree, generally, the air temperature is higher, the irradiance is higher, and the air temperature and the power are positively related at the moment; in practice, the higher the temperature is, the lower the photovoltaic output efficiency is, because the temperature of the photovoltaic panel assembly inevitably increases due to the increase of the outside temperature (temperature-environment), and the conversion efficiency of the photovoltaic panel assembly also decreases due to the increase of the temperature inside the assembly, which is the manifestation of negative correlation between the air temperature and the power. Humidity (humidity) represents a physical quantity of the degree of dryness of the atmosphere, which does not directly affect the power generation amount, but is closely related to irradiance, and generally speaking, the higher the humidity is, the lower the irradiance is, and the lower the photovoltaic power output is. The humidity versus power relationship and irradiance are diametrically opposed.
As shown in fig. 5, for the selection of the optimal parameters, through a grid search method search and multiple experiments, the optimal hyper-parameters required for DNN model modeling are successfully obtained, and finally, the optimal hyper-parameter combinations are determined to be batch _ Size =1000, epoch =200, the number of hidden layer (input) nodes =19, the number of input (hidden) nodes =9, the number of output (output) nodes =1, the optimization algorithm is an Adam function, and the activation functions are ELU and softmax. A Sequential (Sequential) model is loaded among the models. By adding a sequential model, a linear stacking of multiple network layers is performed, which is called "one-way-to-black". During the course of the experiment, there are two methods, one is to construct a Sequential (Sequential) model by passing the list of a layer to the model, and the other is to add layers to the model one by the add () method. The add () method inputs the layer into the model, and the DNN model collectively establishes 8 layers of networks (1 layer of input layer, one layer of output layer, and six hidden layers), and in order to prevent the overfitting phenomenon, a dropout mechanism is adopted to randomly discard some neurons according to the above. And (4) constructing a DNN model by adopting a neuron full-connection mode and utilizing the optimal hyper-parameter combination to obtain an experimental result predicted by the DNN model.
As shown in fig. 6, the ANN model and the DNN model are consistent, for the selection of the optimal parameters, the optimal hyper-parameters required for modeling the ANN model are successfully obtained through the grid search method search and multiple experiments, and the optimal hyper-parameter combinations are finally determined to be batch _ Size =3000, epoch =300, the number of hidden layer nodes =19, the number of input nodes =9, the number of output nodes =1, the optimization algorithm is an Adam function, and the activation function is an ELU. Like the DNN model, the ANN model loads a Sequential (Sequential) model, and layers are added to the model one by the add () method, thereby establishing a layer-3 network. And (4) building an ANN model by adopting the optimal hyper-parameter combination provided by the invention to obtain an experimental result predicted by the ANN model.
As shown in fig. 7, the convolution layer (convolution) of the CNN model is set to 2 layers, the pooling layer (max-pool) is set to 1 layer, a flat layer is set, the fully connected layer (dense) is set to 1 layer, and dropout is set to 1 layer. batch _ Size =2000, epoch = 100. The convolution kernel size is 3 × 3 and the number of convolution kernels is set to 64. As with the DNN model, the CNN model loads a Sequential model, with each layer added to the model individually by the add () method.
And integrating the output results of the three deep learning models, and obtaining a final experimental result by using an averaging method.
The prediction model based on the integrated ANN model, the CNN model and the DNN model compares the advantages and disadvantages of the models under different parameters, finds the most appropriate activation function, trains parameters such as batch size, training times, hidden layer node number and the like, after the ANN-based learner, the DNN-based learner and the CNN-based learner are built, integrates the 3 base learners by adopting a bagging integration method, divides the data set by adopting a self-service method and serves as the input of the integrated model, and outputs the final result of the results output by the three base learners by adopting an averaging method.
The influence of different values of the parameters of the DNN model, the ANN model and the CNN model on the model effect is shown below. And (3) displaying the selection process of the parameters by changing the loss rate (loss) under the condition that the parameters of the activation function, the training batch size and the training times are different in value.
When the value of the batch _ size of the DNN model is 1000 and the value of the epoch (training time) is 200, the loss rate of the model is the lowest, so when training the model, the training batch size is set to 1000 and the training time is set to 200, which can achieve the best effect. As shown in table 1:
TABLE 1 model loss Rate for different combinations of DNN model training batch size and training times
When the value of batch _ Size of the CNN model is 2000 and the value of epoch is 100, the loss rate of the model is the lowest, so that when the model is trained, the Size of the training batch is set to 2000, and the training times is set to 100, thereby achieving the optimal effect. As shown in table 2:
TABLE 2 model loss Rate for different combinations of training batch size and training times for CNN model
When the value of batch _ Size of the ANN model is 3000 and the value of epoch is 300, the loss rate of the model is the lowest, so that when the model is trained, the Size of a training batch is set to 3000, and the training times are set to 300, so that the optimal effect can be achieved. As shown in table 3:
TABLE 3 model loss Rate for different combinations of ANN model training batch size and training times
As shown in fig. 8 and fig. 9, model loss rates of different optimization algorithms for DNN models and CNN are obtained, which are specifically shown in the accompanying drawings: the loss rate of the algorithm excluding the SGD and RMSprop optimization is relatively high, other optimization algorithms are almost the same, the loss rate is below 0.025, and the use effect of the Adam optimization function is slightly better than that of the other 6 optimization functions. Therefore, in the constructed DNN and CNN models, the Adam function is adopted to optimize the models so as to achieve the optimal effect.
After the ANN base learner, the DNN base learner and the CNN base learner are built, the 3 base learners are integrated by adopting a bagging integration method, and the final result is output by averaging the results output by the three base learners. The values obtained by evaluating the indices of the regression models of the integrated model and the three base learners are shown below. As shown in table 4:
TABLE 4 model evaluation index
The regression prediction effect of the three base learners can meet certain requirements. In the three base learners, the effect of the artificial neural network is slightly worse than that of the other three models, the evaluation index value is larger, and the representative error is larger; the model effect of the deep neural network is better than that of the artificial neural network, but is poorer than that of the convolution neural network; the convolutional neural network performs the best in the three base learners, and the loss rate is the lowest; the integrated model constructed by the three base learners is better than the independent regression prediction of the three models, so that the integrated model is the optimal choice for model building and the prediction precision is higher.
Referring to fig. 10, a block diagram of a photovoltaic power generation prediction apparatus based on an integrated prediction model according to an embodiment of the present invention is shown.
As shown in fig. 10, the photovoltaic power generation prediction apparatus 200 includes a building module 210, an integrating module 220, and an outputting module 230.
The building module 210 is configured to build a neural network model in response to the acquired photovoltaic power station historical data, wherein the neural network model comprises an artificial neural network model, a deep neural network model and a convolutional neural network model; an integration module 220 configured to construct an integrated prediction model based on the artificial neural network model, the deep neural network model, and the convolutional neural network model as basis learners of a Bagging method, respectively; an output module 230 configured to output a prediction result of the photovoltaic power generation power based on the integrated prediction model in response to the acquired photovoltaic power plant real-time data.
It should be understood that the modules recited in fig. 10 correspond to various steps in the method described with reference to fig. 1. Thus, the operations and features described above for the method and the corresponding technical effects are also applicable to the modules in fig. 10, and are not described again here.
In other embodiments, the present invention further provides a non-volatile computer storage medium, where the computer storage medium stores computer-executable instructions, and the computer-executable instructions may execute the photovoltaic power generation prediction method in any of the above method embodiments;
as one embodiment, a non-volatile computer storage medium of the present invention stores computer-executable instructions configured to:
responding to the acquired historical data of the photovoltaic power station, and constructing a neural network model, wherein the neural network model comprises an artificial neural network model, a deep neural network model and a convolutional neural network model;
constructing an integrated prediction model based on the artificial neural network model, the deep neural network model and the convolutional neural network model which are respectively used as basis learners of a Bagging method;
and responding to the acquired real-time data of the photovoltaic power station, and outputting a prediction result of the photovoltaic power generation power based on the integrated prediction model.
The non-volatile computer-readable storage medium may include a storage program area and a storage data area, wherein the storage program area may store an operating system, an application program required for at least one function; the storage data area may store data created from use of the photovoltaic power generation prediction apparatus, and the like. Further, the non-volatile computer-readable storage medium may include high speed random access memory, and may also include non-volatile memory, such as at least one magnetic disk storage device, flash memory device, or other non-volatile solid state storage device. In some embodiments, the non-transitory computer readable storage medium optionally includes memory located remotely from the processor, which may be connected to the photovoltaic power generation prediction apparatus over a network. Examples of such networks include, but are not limited to, the internet, intranets, local area networks, mobile communication networks, and combinations thereof.
Embodiments of the present invention also provide a computer program product, which includes a computer program stored on a non-volatile computer-readable storage medium, where the computer program includes program instructions, and when the program instructions are executed by a computer, the computer executes any one of the above-mentioned photovoltaic power generation prediction methods.
Fig. 11 is a schematic structural diagram of an electronic device according to an embodiment of the present invention, and as shown in fig. 11, the electronic device includes: one processor 310 and memory 320, with one processor 310 being illustrated in fig. 11. The electronic device may further include: an input device 330 and an output device 340. The processor 310, the memory 320, the input device 330, and the output device 340 may be connected by a bus or other means, and fig. 11 illustrates an example of a connection by a bus. The memory 320 is a non-volatile computer-readable storage medium as described above. The processor 310 executes various functional applications and data processing of the server by running the nonvolatile software programs, instructions and modules stored in the memory 320, so as to implement the photovoltaic power generation prediction method of the above method embodiment. The input device 330 may receive input numeric or character information and generate key signal inputs related to user settings and function control of the photovoltaic power generation prediction apparatus. The output device 340 may include a display device such as a display screen.
The product can execute the method provided by the embodiment of the invention, and has corresponding functional modules and beneficial effects of the execution method. For technical details that are not described in detail in this embodiment, reference may be made to the method provided by the embodiment of the present invention.
As an embodiment, the electronic device is applied to a photovoltaic power generation prediction device, and is used for a client, and the electronic device includes: at least one processor; and a memory communicatively coupled to the at least one processor; wherein the memory stores instructions executable by the at least one processor to cause the at least one processor to:
responding to the acquired historical data of the photovoltaic power station, and constructing a neural network model, wherein the neural network model comprises an artificial neural network model, a deep neural network model and a convolutional neural network model;
constructing an integrated prediction model based on the artificial neural network model, the deep neural network model and the convolutional neural network model which are respectively used as basis learners of a Bagging method;
and responding to the acquired real-time data of the photovoltaic power station, and outputting a prediction result of the photovoltaic power generation power based on the integrated prediction model.
Finally, it should be noted that: the above examples are only intended to illustrate the technical solution of the present invention, but not to limit it; although the present invention has been described in detail with reference to the foregoing embodiments, it will be understood by those of ordinary skill in the art that: the technical solutions described in the foregoing embodiments may still be modified, or some technical features may be equivalently replaced; and such modifications or substitutions do not depart from the spirit and scope of the corresponding technical solutions of the embodiments of the present invention.
Claims (7)
1. A photovoltaic power generation prediction method based on an integrated prediction model is characterized by comprising the following steps:
step 1, responding to acquired photovoltaic power station historical data, and constructing a neural network model, wherein the neural network model comprises an artificial neural network model, a deep neural network model and a convolutional neural network model, and the constructing of the neural network model specifically comprises the following steps:
step 1.1, constructing an artificial neural network model, searching and performing multiple experiments by a grid search method to obtain an optimal hyper-parameter required by modeling of the artificial neural network model, finally determining the number of nodes of a hidden layer, inputting the number of nodes and outputting the optimal hyper-parameter combination of the number of nodes, selecting an Adam optimization function, loading a first sequential model into the artificial neural network model, and inputting a layer into the artificial neural network model one by an add () method;
step 1.2, constructing a deep neural network model, searching and performing multiple experiments by a grid search method to obtain an optimal hyper-parameter required by model modeling, finally determining the optimal hyper-parameter combination of the number of hidden layer nodes, the number of input nodes and the number of output nodes, selecting an Adam optimization function, loading a second sequential model in the deep neural network model, and inputting layer into the deep neural network model one by an add () method;
step 1.3, constructing a convolutional neural network model, setting the number of layers of a convolutional layer, a pooling layer and a full-link layer, simultaneously setting related parameters of corresponding convolutional cores, pooling windows and the number of neurons, setting a flatten layer for a connection part of the pooling layer and the full-link layer for feature extension, finally setting the number of nodes of a hidden layer, the number of input nodes and the number of output nodes when the convolutional neural network model feeds data, selecting an Adam optimization function, loading the Adam optimization function into a third sequential model, and inputting layer into the convolutional neural network model one by one through an add () method;
step 2, constructing an integrated prediction model based on the artificial neural network model, the deep neural network model and the convolutional neural network model which are respectively used as basis learners of a Bagging method;
and 3, responding to the acquired real-time data of the photovoltaic power station, and outputting a prediction result of the photovoltaic power generation power based on the integrated prediction model.
2. The integrated prediction model-based photovoltaic power generation prediction method according to claim 1, wherein before step 1, the photovoltaic power plant historical data is preprocessed, and the preprocessing comprises:
filling missing variable values of characteristic data in the historical data of the photovoltaic power station based on a random forest;
calculating and visualizing the importance of the feature data based on the XGBOOST method, selecting the required feature data, and performing data normalization processing on the related feature data, wherein the data normalization processing is a Z-score standardization method:
3. The integrated prediction model-based photovoltaic power generation prediction method according to claim 1, wherein in step 1, the photovoltaic power plant historical data comprises power generation historical data and meteorological historical data, wherein the power generation historical data comprises instantaneous power historical data, and the meteorological historical data comprises irradiance historical data, ambient temperature historical data, humidity historical data and wind speed historical data.
4. The integrated prediction model-based photovoltaic power generation prediction method according to claim 1, wherein in step 3, outputting the prediction result of the photovoltaic power generation power based on the integrated prediction model in response to the acquired real-time data of the photovoltaic power station comprises:
respectively inputting the real-time data of the photovoltaic power station into the artificial neural network model, the deep neural network model and the convolutional neural network model in the integrated prediction model, and respectively outputting a first prediction result, a second prediction result and a third prediction result;
and calculating the first prediction result, the second prediction result and the third prediction result based on an averaging method to obtain a prediction result of the photovoltaic power generation power.
5. A photovoltaic power generation prediction device based on an integrated prediction model is characterized by comprising:
the building module is configured to build a neural network model in response to the acquired photovoltaic power station historical data, wherein the neural network model comprises an artificial neural network model, a deep neural network model and a convolutional neural network model, and the building of the neural network model specifically comprises the following steps:
constructing an artificial neural network model, searching and performing multiple experiments by a grid search method to obtain an optimal hyper-parameter required by modeling of the artificial neural network model, finally determining the optimal hyper-parameter combination of the number of hidden layer nodes, the number of input nodes and the number of output nodes, selecting an Adam optimization function, loading a first sequential model into the artificial neural network model, and inputting layer into the artificial neural network model one by an add () method;
constructing a deep neural network model, searching and performing multiple experiments by a grid search method to obtain an optimal hyper-parameter required by model modeling, finally determining an optimal hyper-parameter combination of a hidden layer node number, an input node number and an output node number, selecting an Adam optimization function, loading a second sequential model in the deep neural network model, and inputting a layer into the deep neural network model one by an add () method;
constructing a convolutional neural network model, setting the number of layers of a convolutional layer, a pooling layer and a full-link layer, setting related parameters of the number of corresponding convolutional cores, pooling windows and neurons, setting a flat layer for a connection part of the pooling layer and the full-link layer for feature extension, finally setting the number of nodes of a hidden layer, the number of input nodes and the number of output nodes when the convolutional neural network model feeds data, selecting an Adam optimization function, loading the Adam optimization function into a third sequential model, and inputting layer into the convolutional neural network model one by one through an add () method;
an integration module configured to construct an integrated prediction model based on the artificial neural network model, the deep neural network model, and the convolutional neural network model as basis learners of a Bagging method, respectively;
and the output module is configured to respond to the acquired real-time data of the photovoltaic power station and output a prediction result of the photovoltaic power generation power based on the integrated prediction model.
6. An electronic device, comprising: at least one processor, and a memory communicatively coupled to the at least one processor, wherein the memory stores instructions executable by the at least one processor to enable the at least one processor to perform the method of any of claims 1 to 4.
7. A computer-readable storage medium, on which a computer program is stored, which, when being executed by a processor, carries out the method of any one of claims 1 to 4.
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN202110784293.7A CN113240217A (en) | 2021-07-12 | 2021-07-12 | Photovoltaic power generation prediction method and device based on integrated prediction model |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN202110784293.7A CN113240217A (en) | 2021-07-12 | 2021-07-12 | Photovoltaic power generation prediction method and device based on integrated prediction model |
Publications (1)
Publication Number | Publication Date |
---|---|
CN113240217A true CN113240217A (en) | 2021-08-10 |
Family
ID=77135420
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN202110784293.7A Pending CN113240217A (en) | 2021-07-12 | 2021-07-12 | Photovoltaic power generation prediction method and device based on integrated prediction model |
Country Status (1)
Country | Link |
---|---|
CN (1) | CN113240217A (en) |
Cited By (2)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN113780644A (en) * | 2021-08-31 | 2021-12-10 | 济南大学 | Photovoltaic output prediction method based on online learning |
CN117318024A (en) * | 2023-09-19 | 2023-12-29 | 华能澜沧江水电股份有限公司 | CNN neural network-based photovoltaic power generation power prediction management method and system |
Citations (6)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN105184678A (en) * | 2015-09-18 | 2015-12-23 | 齐齐哈尔大学 | Method for constructing photovoltaic power station generation capacity short-term prediction model based on multiple neural network combinational algorithms |
CN108345856A (en) * | 2018-02-09 | 2018-07-31 | 电子科技大学 | The SAR automatic target recognition methods integrated based on isomery convolutional neural networks |
CN109146162A (en) * | 2018-08-07 | 2019-01-04 | 河海大学 | A kind of probability wind speed forecasting method based on integrated Recognition with Recurrent Neural Network |
CN110909919A (en) * | 2019-11-07 | 2020-03-24 | 哈尔滨工程大学 | Photovoltaic power prediction method of depth neural network model with attention mechanism fused |
CN111161207A (en) * | 2019-11-14 | 2020-05-15 | 西安工程大学 | Integrated convolutional neural network fabric defect classification method |
CN112906987A (en) * | 2021-03-29 | 2021-06-04 | 福州大学 | Photovoltaic power prediction method based on convolutional neural network and two-dimensional meteorological matrix |
-
2021
- 2021-07-12 CN CN202110784293.7A patent/CN113240217A/en active Pending
Patent Citations (6)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN105184678A (en) * | 2015-09-18 | 2015-12-23 | 齐齐哈尔大学 | Method for constructing photovoltaic power station generation capacity short-term prediction model based on multiple neural network combinational algorithms |
CN108345856A (en) * | 2018-02-09 | 2018-07-31 | 电子科技大学 | The SAR automatic target recognition methods integrated based on isomery convolutional neural networks |
CN109146162A (en) * | 2018-08-07 | 2019-01-04 | 河海大学 | A kind of probability wind speed forecasting method based on integrated Recognition with Recurrent Neural Network |
CN110909919A (en) * | 2019-11-07 | 2020-03-24 | 哈尔滨工程大学 | Photovoltaic power prediction method of depth neural network model with attention mechanism fused |
CN111161207A (en) * | 2019-11-14 | 2020-05-15 | 西安工程大学 | Integrated convolutional neural network fabric defect classification method |
CN112906987A (en) * | 2021-03-29 | 2021-06-04 | 福州大学 | Photovoltaic power prediction method based on convolutional neural network and two-dimensional meteorological matrix |
Non-Patent Citations (8)
Title |
---|
MUHAMMAD QAMAR RAZA等: ""An ensemble framework for day-ahead forecast of PV output power in smart grids"", 《IEEE TRANSACTIONS ON INDUSTRIAL INFORMATICS》 * |
SAMEER AL-DAHIDI等: ""Ensemble Approach of Optimized Artificial Neural Networks for Solar Photovoltaic Power Prediction"", 《IEEE ACCESS》 * |
朱晓飞: ""超短期光伏发电功率预测方法的研究"", 《中国优秀硕士学位论文全文数据库 工程科技Ⅱ辑》 * |
浙江省统计局: "《数字的魅力:基于统计视野的浙江经济社会发展研究》", 31 July 2019, 浙江工商大学出版社 * |
王继民等: ""基于集成学习的中小河流洪水预报"", 《计算机与现代化》 * |
程鑫: "《互联网环境下农户诚信评价研究 基于支持向量机方法》", 30 April 2018, 武汉大学出版社 * |
蒋俊梅: ""基于深度学习的光伏板温度和光伏发电功率预测"", 《中国优秀硕士学位论文全文数据库 工程科技Ⅱ辑》 * |
高志强等: "《深度学习 从入门到实战》", 30 June 2018, 中国铁道出版社 * |
Cited By (3)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN113780644A (en) * | 2021-08-31 | 2021-12-10 | 济南大学 | Photovoltaic output prediction method based on online learning |
CN113780644B (en) * | 2021-08-31 | 2023-09-01 | 济南大学 | Photovoltaic output prediction method based on online learning |
CN117318024A (en) * | 2023-09-19 | 2023-12-29 | 华能澜沧江水电股份有限公司 | CNN neural network-based photovoltaic power generation power prediction management method and system |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
CN110909919A (en) | Photovoltaic power prediction method of depth neural network model with attention mechanism fused | |
CN113240217A (en) | Photovoltaic power generation prediction method and device based on integrated prediction model | |
CN112381673B (en) | Park electricity utilization information analysis method and device based on digital twin | |
CN115425680B (en) | Power prediction model construction and prediction method of multi-energy combined power generation system | |
CN113705922A (en) | Improved ultra-short-term wind power prediction algorithm and model establishment method | |
CN112183877A (en) | Photovoltaic power station fault intelligent diagnosis method based on transfer learning | |
CN109921462B (en) | New energy consumption capability assessment method and system based on LSTM | |
CN115859099A (en) | Sample generation method and device, electronic equipment and storage medium | |
CN116599050A (en) | Photovoltaic prediction method and related device based on self-attention mechanism | |
CN112215478B (en) | Power coordination control method and device for optical storage station and storage medium | |
CN112836876A (en) | Power distribution network line load prediction method based on deep learning | |
CN113033898A (en) | Electrical load prediction method and system based on K-means clustering and BI-LSTM neural network | |
CN117458453A (en) | Method, system, equipment and storage medium for predicting transient stability of power system | |
CN117390550A (en) | Low-carbon park carbon emission dynamic prediction method and system considering emission training set | |
KR102583178B1 (en) | Method and apparatus for predicting power generation | |
CN113723670B (en) | Photovoltaic power generation power short-term prediction method with variable time window | |
CN112598186B (en) | Improved LSTM-MLP-based small generator fault prediction method | |
Yuan et al. | A novel hybrid short-term wind power prediction framework based on singular spectrum analysis and deep belief network utilized improved adaptive genetic algorithm | |
CN111967689A (en) | Model and method for wind power generation prediction by combining multivariate and stepwise linear regression and artificial neural network | |
CN112215383A (en) | Distributed photovoltaic power generation power prediction method and system | |
Amreen et al. | Solar power forecasting using hybrid model | |
CN113486601B (en) | Feeder voltage calculation method and device based on CPSO-BP optimization model | |
Luo et al. | Prediction of Photovoltaic Power Generation Based on PSO-RNN and SVR Model | |
Zhao et al. | An interpretable ultra-short-term wind power prediction model based on the feature matrix reconstruction through regression trees | |
CN116957223B (en) | Photovoltaic power prediction method and device based on illumination image |
Legal Events
Date | Code | Title | Description |
---|---|---|---|
PB01 | Publication | ||
PB01 | Publication | ||
SE01 | Entry into force of request for substantive examination | ||
SE01 | Entry into force of request for substantive examination | ||
RJ01 | Rejection of invention patent application after publication | ||
RJ01 | Rejection of invention patent application after publication |
Application publication date: 20210810 |