Please enable JavaScript.
Coggle requires JavaScript to display documents.
Seamless photovoltaic power generation forecasting, Figure 1 : Seamless…
Seamless photovoltaic power generation
forecasting
Hybrid solar
radiation forecasting models
General Ensemble Learning Approach (GELA)
Residual based Ensemble Learning Approach(RELA)
Decomposition-Clustering based Ensemble Learning Approach (CDELA)
Evolutionary based Ensemble Learning
Approach (EELA)
Cluster Based Ensemble Learning Approach (CELA)
ecomposition Based Ensemble Learning Approach (DELA)
Post-processing models in solar forecasting
Probabilistic-to-deterministic
(P2D) post-processing
Deterministic-to-probabilistic (D2P) post-processing
Deterministic-to-deterministic (D2D) post-processing
Probabilistic-to-probabilistic (P2P)
Post-processing forecasting combining deterministic solar irradiance forecasts
Linear combinations
Variance–covariance methods
Ex : Combination through mean square error (MSE)
Variance–covariance methods
Ex : the linear regression
Penalized regressions
Lasso regression
Ridge regression
Simple average
Trimmed averaging
Non-linear combinations
Deep learning
Artificial Neural Networks (ANNs)
Classification
Support vector machine (SVM)
Random forest
Current forecasting techniques
Hybrid artificial intelligence
Statistical methods
Physical methods
Seamless forecasting
ANAKLIM++ : combining methods
from data assimilation with Gaussian weights
Blend NWP and nowcasting for horizons from 1 to up to 5 hours
Analog ensemble model combining NWP and satellite imagery inputs
Predictions are made from 5 min to 36 hours with a probabilistic approach
Figure 1 : Seamless PV power generation forecasting litterature review mindmap
