Method for constructing photovoltaic power station generation capacity short-term prediction model based on multiple neural network combinational algorithms
A neural network algorithm and neural network technology, applied in the forecast of solar photovoltaic power generation, photovoltaic power generation and grid-connected technology, can solve problems such as single algorithm, poor applicability in different weather, large measurement error of prediction model, etc., to improve prediction accuracy, Practicality guarantee, effect of reducing prediction error
- Summary
- Abstract
- Description
- Claims
- Application Information
AI Technical Summary
Problems solved by technology
Method used
Image
Examples
Example Embodiment
[0039] Specific embodiment one: see figure 1 To explain this embodiment, the method for constructing a short-term prediction model of photovoltaic power generation capacity based on multiple neural network combination algorithms described in this embodiment is as follows:
[0040] Step 1: Selection of neural network algorithm;
[0041] The four neural network algorithms of BP, Elman, RBF and GRNN are used to construct neural network prediction sub-models A, B, C and D respectively, and the ambient temperature T at different time points during the daily power station working period i , Daily average solar radiation intensity Average daily wind speed As the input data of the short-term prediction model of photovoltaic power generation, to predict the photovoltaic output power P at the corresponding time point of the day i As the output data of each neural network prediction sub-model A, B, C and D;
[0042] Step 2: Selection of sample data;
[0043] Through the photovoltaic data collec...
Example Embodiment
[0048] Embodiment 2: The difference between this embodiment and the method for constructing a short-term prediction model of photovoltaic power generation capacity based on multiple neural network combination algorithms described in the first embodiment is that the method also includes step four: short-term photovoltaic power generation Revision of the forecast model;
[0049] a) First, normalize the sample data in step two,
[0050] b) Secondly, genetic algorithm and particle swarm algorithm are used.
[0051] In this embodiment, first, the sample data in step 2 is normalized to further reduce the prediction error of the prediction sub-model. Secondly, using genetic algorithm and particle swarm algorithm, two methods are used to optimize all neural network predictive sub-models, which can avoid the problem of neural network algorithm easily falling into local optimality, and further reduce the prediction error of neural network predictive sub-model.
[0052] In this embodiment, the ...
Example Embodiment
[0059] Embodiment 3: The difference between this embodiment and the method for constructing a short-term prediction model of photovoltaic power generation capacity based on multiple neural network combination algorithms described in the first or second embodiment is that the method also includes step five: Quantitative short-term forecasting model evaluation;
[0060] Two error evaluation methods of average absolute percentage error MAPE and root mean square error RMSE are used to evaluate the error of the short-term prediction model of photovoltaic power generation.
[0061] M A P E = 1 N X i = 1 N | P p i - P a i P a i | X 100 % ,
[0062] R M S E = X i = 1 N ( P p i - P a i ) N 2 ,
[0063] Among them, N represents the total amount of data, i is a positive integer, Represents the predicted value of the i-th data point, Represents the actual...
PUM
Abstract
Description
Claims
Application Information
- R&D Engineer
- R&D Manager
- IP Professional
- Industry Leading Data Capabilities
- Powerful AI technology
- Patent DNA Extraction
Browse by: Latest US Patents, China's latest patents, Technical Efficacy Thesaurus, Application Domain, Technology Topic.
© 2024 PatSnap. All rights reserved.Legal|Privacy policy|Modern Slavery Act Transparency Statement|Sitemap