CN116933922A - WOA-CNN-LSTM-based photovoltaic power generation power prediction method - Google Patents
WOA-CNN-LSTM-based photovoltaic power generation power prediction method Download PDFInfo
- Publication number
- CN116933922A CN116933922A CN202310803517.3A CN202310803517A CN116933922A CN 116933922 A CN116933922 A CN 116933922A CN 202310803517 A CN202310803517 A CN 202310803517A CN 116933922 A CN116933922 A CN 116933922A
- Authority
- CN
- China
- Prior art keywords
- cnn
- woa
- model
- lstm
- lstm model
- 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 31
- 238000010248 power generation Methods 0.000 title claims abstract description 24
- 238000012549 training Methods 0.000 claims abstract description 30
- 238000012360 testing method Methods 0.000 claims abstract description 19
- 238000011156 evaluation Methods 0.000 claims abstract description 9
- 238000005457 optimization Methods 0.000 claims abstract description 7
- 239000010410 layer Substances 0.000 claims description 30
- 238000012545 processing Methods 0.000 claims description 12
- 238000010606 normalization Methods 0.000 claims description 8
- 241000283153 Cetacea Species 0.000 claims description 7
- 238000007781 pre-processing Methods 0.000 claims description 5
- 230000002159 abnormal effect Effects 0.000 claims description 4
- 239000002356 single layer Substances 0.000 claims description 4
- 238000011176 pooling Methods 0.000 claims description 2
- 238000013527 convolutional neural network Methods 0.000 description 13
- 238000002474 experimental method Methods 0.000 description 3
- 238000004458 analytical method Methods 0.000 description 1
- 238000013528 artificial neural network Methods 0.000 description 1
- 125000004122 cyclic group Chemical group 0.000 description 1
- 230000006870 function Effects 0.000 description 1
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06Q—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
- G06Q10/00—Administration; Management
- G06Q10/04—Forecasting or optimisation specially adapted for administrative or management purposes, e.g. linear programming or "cutting stock problem"
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/004—Artificial life, i.e. computing arrangements simulating life
- G06N3/006—Artificial life, i.e. computing arrangements simulating life based on simulated virtual individual or collective life forms, e.g. social simulations or particle swarm optimisation [PSO]
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/02—Neural networks
- G06N3/04—Architecture, e.g. interconnection topology
- G06N3/044—Recurrent networks, e.g. Hopfield networks
- G06N3/0442—Recurrent networks, e.g. Hopfield networks characterised by memory or gating, e.g. long short-term memory [LSTM] or gated recurrent units [GRU]
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/02—Neural networks
- G06N3/04—Architecture, e.g. interconnection topology
- G06N3/045—Combinations of networks
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/02—Neural networks
- G06N3/04—Architecture, e.g. interconnection topology
- G06N3/0464—Convolutional networks [CNN, ConvNet]
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/02—Neural networks
- G06N3/08—Learning methods
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06Q—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
- G06Q50/00—Systems or methods specially adapted for specific business sectors, e.g. utilities or tourism
- G06Q50/06—Electricity, gas or water supply
Abstract
The invention discloses a WOA-CNN-LSTM photovoltaic power generation power prediction method, which comprises the following steps: step 1, generating data and meteorological data of a photovoltaic power station to be controlled are calculated according to 4:1 is divided into training set data and test set data; step 2, constructing a WOA-CNN-LSTM model, and performing optimization training; and 3, inputting the test set data into a trained WOA-CNN-LSTM model, outputting a predicted result by the WOA-CNN-LSTM model, evaluating the WOA-CNN-LSTM model through a relevant evaluation index, comparing the result with the predicted result of the WOA-CNN-LSTM model, and performing model evaluation on the output result of the model. The method disclosed by the invention has the advantages of high convergence rate, good curve fitting degree and high prediction accuracy, and is suitable for predicting the photovoltaic power generation power.
Description
Technical Field
The invention belongs to the technical field of control systems, and relates to a WOA-CNN-LSTM-based photovoltaic power generation power prediction method.
Background
In recent years, along with the continuous increase of the installed capacity of the photovoltaic and the influence of weather factors, the solar photovoltaic power generation has obvious intermittence and fluctuation, and has a certain influence on the stability of a power grid connected with a photovoltaic power generation output end.
Although there are many methods for predicting photovoltaic power generation power at present, most of the methods adopt a traditional cyclic neural network or a variant LSTM thereof, and have the problems of insufficient generalization capability, low prediction accuracy and the like.
Disclosure of Invention
The invention aims to provide a WOA-CNN-LSTM-based photovoltaic power generation power prediction method, which solves the problems of low prediction accuracy and low convergence rate in the prior art.
The technical scheme adopted by the invention is that the WOA-CNN-LSTM photovoltaic power generation power prediction method is implemented according to the following steps:
step 1, generating data and meteorological data of a photovoltaic power station to be controlled are calculated according to 4:1 is divided into training set data and test set data;
step 2, constructing a WOA-CNN-LSTM model, and performing optimization training;
and 3, inputting the test set data into a trained WOA-CNN-LSTM model, outputting a predicted result by the WOA-CNN-LSTM model, evaluating the WOA-CNN-LSTM model through a relevant evaluation index, comparing the result with the predicted result of the CNN-LSTM model, and performing model evaluation on the output result of the model.
The WOA-CNN-LSTM model has the advantages that the WOA-CNN-LSTM model improved by the WOA algorithm is high in convergence rate, good in curve fitting degree and high in prediction accuracy, and the result proves that the model is ideal for the photovoltaic power generation power prediction model.
Drawings
FIG. 1 is a general flow of the method of the present invention;
FIG. 2 is a flow chart of data preprocessing in the method of the present invention;
FIG. 3 is a graph showing the comparison result of the method of the present invention between example 1 and the CNN-LSTM algorithm;
FIG. 4 is a graph showing the comparison result of the method of example 2 of the present invention and the CNN-LSTM algorithm;
FIG. 5 is a graph showing the comparison result of the method of the present invention in example 3 and the CNN-LSTM algorithm.
Detailed Description
The invention will be described in detail below with reference to the drawings and the detailed description.
In the method, a WOA algorithm (namely a whale algorithm) is used for optimizing the initial learning rate of training parameters of an LSTM model and the number of super-parameter hidden layer nodes of the model, and the method comprises the following steps:
1) Setting parameters of WOA algorithm and initial parameters of LSTM model;
2) Starting searching: initializing the position of the belonged whale, performing iterative optimization, returning to a search agent beyond the boundary of the search space, calculating the objective function of each search agent, and updating;
3) Training a model, wherein the best global optimal parameters are found by using a whale algorithm;
according to the method, an LSTM model optimized by a WOA algorithm and a CNN model are combined into a WOA-CNN-LSTM model, and the WOA-CNN-LSTM model is used for experimental result analysis of photovoltaic power generation power prediction, and the method is implemented according to the following steps:
step 1, generating data and meteorological data of a photovoltaic power station to be controlled are calculated according to 4: the scale of 1 is divided into training set data and test set data,
the training set data are used for training a WOA-CNN-LSTM model built later, and the testing set data are used for evaluating the performance of the model;
step 2, constructing a WOA-CNN-LSTM model, performing optimization training,
the method comprises the steps of setting parameters of a WOA algorithm, selecting parameters of a CNN model and a LATM model, optimizing super parameters of the LSTM model through the WOA algorithm, wherein the CNN model comprises an input layer, a convolution layer, a pooling layer and an output layer, connecting the output layer of the CNN with the input of the LSTM model after WOA optimization, and outputting a prediction result of the WOA-CNN-LSTM model through a hiding layer, an output layer and a full connection layer;
and 3, inputting the test set data into a trained WOA-CNN-LSTM model, outputting a predicted result by the WOA-CNN-LSTM model, evaluating the WOA-CNN-LSTM model through related evaluation indexes (such as root mean square error and average absolute error), comparing the result with the predicted result of the CNN-LSTM model, and performing model evaluation on the output result of the model.
Example 1
Referring to fig. 1, the WOA-CNN-LSTM photovoltaic power generation power prediction-based method of the present invention optimizes super parameters of an LSTM model by using a WOA algorithm on the basis of a CNN module-LSTM module, and specifically includes the following steps:
1) Data preprocessing, referring to fig. 2, the original data is subjected to missing value processing and abnormal value processing to obtain a better data set, and then data normalization processing is performed.
2) Setting network parameters, wherein a CNN model is of a single-layer structure through a comparison experiment, and the size of a convolution kernel in the CNN model is 3 multiplied by 3; the LSTM model has 3 nodes, 180 nodes in the hidden layer and 1 node in the output layer. The number of times of network training is 200, the initial learning rate is 0.005, the learning rate adjustment factor is 0.2, and the learning rate is adjusted after training is set for 100 times.
3) Optimizing parameters, and optimizing the LSTM model by using a WOA algorithm, wherein the number of WOA module populations is set to be 30, and the iteration number is 20. After the LSTM model parameters are optimized by using whale algorithm, the number of hidden layer nodes is 152, and the initial learning rate is 0.0076.
4) Training and testing, wherein the training and testing are performed by using the preprocessed data and the built WOA-CNN-LSTM model.
5) And (5) carrying out inverse normalization on the output result to obtain the photovoltaic power generation power of the predicted object.
Referring to FIG. 3, the prediction accuracy and fitness of the WOA-CNN-LSTM model and the CNN-LSTM model are compared by using the same test set. As can be seen from FIG. 3, the WOA-CNN-LSTM model of the present invention has better accuracy and better fitting degree than the CNN-LSTM model, which indicates that the model has higher accuracy and better prediction capability.
Example 2
According to the procedure consistent with the embodiment 1, the implementation and comparison are carried out, and referring to fig. 1, the WOA-CNN-LSTM photovoltaic power generation power prediction-based method of the present invention optimizes super parameters of an LSTM model based on a CNN module-LSTM module by using a WOA algorithm, specifically comprising the following steps:
1) Data preprocessing, referring to fig. 2, the original data is subjected to missing value processing and abnormal value processing to obtain a better data set, and then data normalization processing is performed.
2) Setting network parameters, wherein a CNN model is of a single-layer structure through a comparison experiment, and the size of a convolution kernel in the CNN model is 3 multiplied by 3; the LSTM model has 3 nodes, 180 nodes in the hidden layer and 1 node in the output layer. The number of training times of the network is 200, the initial learning rate is 0.0055, the learning rate adjustment factor is 0.21, and the learning rate starts to be adjusted after training is set for 100 times.
3) Optimizing parameters, and optimizing the LSTM model by using a WOA algorithm, wherein the number of WOA module populations is set to be 30, and the iteration number is 25. After the LSTM model parameters are optimized by using whale algorithm, the number of hidden layer nodes is 152, and the initial learning rate is 0.0077.
4) Training and testing, wherein the training and testing are performed by using the preprocessed data and the built WOA-CNN-LSTM model.
5) And (5) carrying out inverse normalization on the output result to obtain the photovoltaic power generation power of the predicted object.
The difference from example 1 is that the final result is referred to fig. 4. As can be seen from FIG. 4, the WOA-CNN-LSTM model of the present invention has better accuracy and better fitting degree than the CNN-LSTM model, which indicates that the model has higher accuracy and better prediction capability.
Example 3
According to the procedure consistent with the embodiment 1, the implementation and comparison are carried out, and referring to fig. 1, the WOA-CNN-LSTM photovoltaic power generation power prediction-based method of the present invention optimizes super parameters of an LSTM model based on a CNN module-LSTM module by using a WOA algorithm, specifically comprising the following steps:
1) Data preprocessing, referring to fig. 2, the original data is subjected to missing value processing and abnormal value processing to obtain a better data set, and then data normalization processing is performed.
2) Setting network parameters, wherein a CNN model is of a single-layer structure through a comparison experiment, and the size of a convolution kernel in the CNN model is 3 multiplied by 3; the LSTM model has 3 nodes, 180 nodes in the hidden layer and 1 node in the output layer. The number of training times of the network is 200, the initial learning rate is 0.0045, the learning rate adjustment factor is 0.20, and the learning rate starts to be adjusted after training is set for 100 times.
3) Optimizing parameters, and optimizing the LSTM model by using a WOA algorithm, wherein the number of WOA module populations is set to be 30, and the iteration number is 20. After the LSTM model parameters are optimized by using whale algorithm, the number of hidden layer nodes is 152, and the initial learning rate is 0.0075.
4) Training and testing, wherein the training and testing are performed by using the preprocessed data and the built WOA-CNN-LSTM model.
5) And (5) carrying out inverse normalization on the output result to obtain the photovoltaic power generation power of the predicted object.
The difference from example 1 is that the final result is referred to in fig. 5. As can be seen from FIG. 5, the WOA-CNN-LSTM model of the present invention has better accuracy and better fitting degree than the CNN-LSTM model, which indicates that the model has higher accuracy and better prediction capability.
Claims (5)
1. The WOA-CNN-LSTM photovoltaic power generation power prediction method is characterized by comprising the following steps of:
step 1, generating data and meteorological data of a photovoltaic power station to be controlled are calculated according to 4:1 is divided into training set data and test set data;
step 2, constructing a WOA-CNN-LSTM model, and performing optimization training;
and 3, inputting the test set data into a trained WOA-CNN-LSTM model, outputting a predicted result by the WOA-CNN-LSTM model, evaluating the WOA-CNN-LSTM model through a relevant evaluation index, comparing the result with the predicted result of the CNN-LSTM model, and performing model evaluation on the output result of the model.
2. The WOA-CNN-LSTM photovoltaic power generation power prediction based method of claim 1, wherein: in step 1, the training set data are used for training a WOA-CNN-LSTM model built later, and the testing set data are used for evaluating the performance of the model.
3. The WOA-CNN-LSTM photovoltaic power generation power prediction based method of claim 1, wherein: in the step 2, the specific process is as follows:
the method comprises the steps of setting parameters of a WOA algorithm, selecting parameters of a CNN model and a LATM model, optimizing super parameters of the LSTM model through the WOA algorithm, wherein the CNN model comprises an input layer, a convolution layer, a pooling layer and an output layer, connecting the output layer of the CNN with the input of the LSTM model after WOA optimization, and outputting a prediction result of the WOA-CNN-LSTM model through a hiding layer, an output layer and a full connection layer.
4. A method of WOA-CNN-LSTM photovoltaic power generation power prediction based on claim 3, characterized by: on the basis of a CNN module-LSTM module, optimizing the super parameters of the LSTM model by using a WOA algorithm, wherein the specific steps are as follows:
1) Data preprocessing, namely performing missing value processing and abnormal value processing on original data to obtain a better data set, and performing data normalization processing;
2) Setting network parameters, wherein a CNN model is of a single-layer structure, and the size of a convolution kernel in the CNN model is 3 multiplied by 3; the number of nodes of the LSTM model input layer is 3, the number of nodes of the hidden layer is 180, and the number of nodes of the output layer is 1; the number of times of network training is 200, the initial learning rate is 0.005, the learning rate adjustment factor is 0.2, and the learning rate is adjusted after training is set for 100 times;
3) Optimizing parameters, and optimizing the LSTM model by using a WOA algorithm, wherein the number of WOA module populations is set to be 30, and the iteration times are 20; after the LSTM model parameters are optimized through a whale algorithm, the number of hidden layer nodes is 152, and the initial learning rate is 0.0076;
4) Training and testing, wherein the training and testing are carried out by utilizing the preprocessed data and the built WOA-CNN-LSTM model;
5) And (5) carrying out inverse normalization on the output result to obtain the photovoltaic power generation power of the predicted object.
5. The WOA-CNN-LSTM photovoltaic power generation power prediction based method of claim 1, wherein: in step 3, the correlation evaluation index refers to a root mean square error and an average absolute error.
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN202310803517.3A CN116933922A (en) | 2023-06-30 | 2023-06-30 | WOA-CNN-LSTM-based photovoltaic power generation power prediction method |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN202310803517.3A CN116933922A (en) | 2023-06-30 | 2023-06-30 | WOA-CNN-LSTM-based photovoltaic power generation power prediction method |
Publications (1)
Publication Number | Publication Date |
---|---|
CN116933922A true CN116933922A (en) | 2023-10-24 |
Family
ID=88388715
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN202310803517.3A Pending CN116933922A (en) | 2023-06-30 | 2023-06-30 | WOA-CNN-LSTM-based photovoltaic power generation power prediction method |
Country Status (1)
Country | Link |
---|---|
CN (1) | CN116933922A (en) |
Cited By (2)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN117113267A (en) * | 2023-10-25 | 2023-11-24 | 杭州海兴泽科信息技术有限公司 | Prediction model training method based on big data and photovoltaic power generation performance detection method |
CN117633494A (en) * | 2023-11-20 | 2024-03-01 | 中国矿业大学 | Coal mine earth surface deformation prediction method based on AWC-LSTM model |
-
2023
- 2023-06-30 CN CN202310803517.3A patent/CN116933922A/en active Pending
Cited By (3)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN117113267A (en) * | 2023-10-25 | 2023-11-24 | 杭州海兴泽科信息技术有限公司 | Prediction model training method based on big data and photovoltaic power generation performance detection method |
CN117113267B (en) * | 2023-10-25 | 2024-02-09 | 杭州海兴泽科信息技术有限公司 | Prediction model training method based on big data and photovoltaic power generation performance detection method |
CN117633494A (en) * | 2023-11-20 | 2024-03-01 | 中国矿业大学 | Coal mine earth surface deformation prediction method based on AWC-LSTM model |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
CN108900346B (en) | Wireless network flow prediction method based on LSTM network | |
CN107330514B (en) | Air quality prediction method based on integrated extreme learning machine | |
CN116933922A (en) | WOA-CNN-LSTM-based photovoltaic power generation power prediction method | |
CN111860982A (en) | Wind power plant short-term wind power prediction method based on VMD-FCM-GRU | |
CN110942194A (en) | Wind power prediction error interval evaluation method based on TCN | |
CN112149879B (en) | New energy medium-and-long-term electric quantity prediction method considering macroscopic volatility classification | |
CN109002948B (en) | CDA-BP-based microgrid short-term photovoltaic power generation power prediction method | |
CN110826791A (en) | Hybrid wind power prediction method based on long-time and short-time memory neural network | |
CN110766200A (en) | Method for predicting generating power of wind turbine generator based on K-means mean clustering | |
CN113222289B (en) | Prediction method of energy power based on data processing | |
CN112434891A (en) | Method for predicting solar irradiance time sequence based on WCNN-ALSTM | |
CN115374995A (en) | Distributed photovoltaic and small wind power station power prediction method | |
CN112149883A (en) | Photovoltaic power prediction method based on FWA-BP neural network | |
CN112949895A (en) | Wind speed prediction method based on dynamic extensible neural network model | |
CN117117859B (en) | Photovoltaic power generation power prediction method and system based on neural network | |
CN115423145A (en) | Photovoltaic power prediction method based on Kmeans-VMD-WT-LSTM method | |
CN111832839A (en) | Energy consumption prediction method based on sufficient incremental learning | |
CN116629416A (en) | Photovoltaic power station power prediction method and device | |
CN112052913B (en) | Distributed photovoltaic power station power data virtual acquisition method | |
CN112001537A (en) | Short-term wind power prediction method based on gray model and support vector machine | |
CN116681154A (en) | Photovoltaic power calculation method based on EMD-AO-DELM | |
CN116628488A (en) | Training method of wind power prediction model, wind power prediction method and device | |
CN116345555A (en) | CNN-ISCA-LSTM model-based short-term photovoltaic power generation power prediction method | |
CN116341728A (en) | Ultra-short-term photovoltaic output power prediction method based on data driving | |
CN113610665B (en) | Wind power generation power prediction method based on multi-delay output echo state network |
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 |