🤖 AI Summary
This study systematically investigates key factors affecting forecasting accuracy for renewable energy generation and demand—namely sampling strategy, time-series stationarity, nonlinear modeling capacity, and hyperparameter optimization. Within a unified experimental framework, it conducts the first cross-dataset evaluation of seven deep learning architectures (LSTM, Stacked LSTM, CNN, CNN-LSTM, DNN, MLP, Encoder-Decoder) on dual-source (meteorological + generation) and multi-site photovoltaic data, assessing their generalizability and robustness. Overfitting is mitigated via early stopping, Dropout, and L2 regularization. Results show that lightweight models—MLP and LSTM—consistently outperform others on both real-world datasets, achieving state-of-the-art RMSE values and demonstrating superior efficiency and stability. The core contribution lies in revealing a non-monotonic relationship between model complexity and predictive performance, thereby establishing a reproducible benchmarking paradigm and practical modeling guidelines for energy time-series forecasting.
📝 Abstract
Unpredictability of renewable energy sources coupled with the complexity of those methods used for various purposes in this area calls for the development of robust methods such as DL models within the renewable energy domain. Given the nonlinear relationships among variables in renewable energy datasets, DL models are preferred over traditional machine learning (ML) models because they can effectively capture and model complex interactions between variables. This research aims to identify the factors responsible for the accuracy of DL techniques, such as sampling, stationarity, linearity, and hyperparameter optimization for different algorithms. The proposed DL framework compares various methods and alternative training/test ratios. Seven ML methods, such as Long-Short Term Memory (LSTM), Stacked LSTM, Convolutional Neural Network (CNN), CNN-LSTM, Deep Neural Network (DNN), Multilayer Perceptron (MLP), and Encoder-Decoder (ED), were evaluated on two different datasets. The first dataset contains the weather and power generation data. It encompasses two distinct datasets, hourly energy demand data and hourly weather data in Spain, while the second dataset includes power output generated by the photovoltaic panels at 12 locations. This study deploys regularization approaches, including early stopping, neuron dropping, and L2 regularization, to reduce the overfitting problem associated with DL models. The LSTM and MLP models show superior performance. Their validation data exhibit exceptionally low root mean square error values.