Solar Radiation Prediction Using Satellite Cloud Cover Forecasts
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current methods for predicting solar power generation are limited by the inability to accurately account for rapid variations in renewable energy production due to natural factors, leading to unstable power supply and economic losses.
Innovation Solution
A method that uses a solar radiation amount prediction model incorporating weather information and cloud cover prediction information from satellite images for time series prediction, enhancing the accuracy of solar radiation forecasting.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If cloud cover forecast data from meteorological agency is used, then solar power generation prediction can be performed, but prediction accuracy is limited due to only 3 categories of cloud cover (clear, cloudy, overcast)
Solution Approach 1:
The patent changes the parameter of cloud cover classification from 3 categories (clear, cloudy, overcast) to 10 levels (0-10), providing more granular and accurate cloud cover information for solar radiation prediction
Solution Approach 2:
The patent introduces satellite image data as an additional dimension of information source, combining it with meteorological agency data to create a multi-source prediction system that improves accuracy without excessive complexity
2Measurement precision
If synoptic weather observation data is used, then total cloud cover values from 0 to 10 can be obtained, but forecast information is not included and observation points are limited
Solution Approach 1:
The patent merges synoptic weather observation data (providing precise 0-10 cloud cover values) with satellite image data and meteorological forecast data, creating a comprehensive prediction system that combines the advantages of multiple data sources
Solution Approach 2:
The patent uses satellite images to predict future cloud cover conditions in advance, enabling forecast information to be obtained before the actual event occurs, thus providing preliminary action capability
3Reliability
If traditional weather forecast data is used, then solar power generation prediction can be performed, but accurate prediction during rapid variations (ramping phenomena) is difficult
Solution Approach 1:
The patent incorporates cloud cover prediction information as feedback into the solar radiation prediction model, allowing the system to continuously adjust predictions based on changing cloud conditions, thereby improving reliability during rapid variations
Solution Approach 2:
The patent uses dynamic satellite image data to capture changing cloud cover conditions in real-time, enabling the prediction system to adapt to rapid variations in weather conditions rather than relying on static forecast data
Data Source
AI summary
According to an exemplary embodiment of the present disclosure, a method for predicting a solar radiation amount by using a solar radiation amount prediction model, which is performed by a computing device may include: inputting weather information into a solar radiation amount prediction model performing time series prediction; additionally inputting cloud cover prediction information into the solar radiation amount prediction model; and predicting a solar radiation amount based on the weather information and the cloud cover prediction information by using the solar radiation amount prediction model.