Abstract:Forecasting solar photovoltaic (PV) power generation has become important due to the chaotic and intermittent nature of solar power. With the development of artificial intelligence (AI) techniques, various deep learning algorithms have been shown to be very powerful for renewable power forecasts. This study aims to develop a probabilistic forecasting approach that leverages a hybrid model, named Transformer-Lower Upper Bound Estimation (Transformer-LUBE) with data processing techniques, for one-day-ahead PV power generation forecasts. Especially, this study presents a novel methodology that integrates an effective imputation technique, a modified Transformer architecture, and a combined post-processing technique. During the pre-processing stage, missing data were imputed using XGBoost with the combination of predictive mean matching (PMM) and bootstrapping, then Pearson Correlation Coefficient was utilized to compute the correlation between the input features. In contrast to the original Transformer architecture, the proposed Transformer model incorporates additional normalization and dropout layers between blocks, and also incorporates multi-head attention mechanisms. Finally, a post-processing strategy was implemented based on the utilization of daily numerical fitting curves along with Gated Recurrent Unit (GRU) to correct the bias of the residual errors. The proposed framework was trained with historical measurement data obtained from ten solar farms located in Taiwan, as well as Numerical Weather Predictions (NWPs) data sourced from Taiwan Central Weather Bureau (CWB). The accuracy of the proposed model has been evaluated by comparing other AI models: Artificial Neural Network (ANN), Long Short-Term Memory (LSTM), GRU, and XGBoost. Through a comprehensive experimentation, the proposed framework demonstrates its superiority in accuracy and reliability.

A Hierarchical Parametric and Non-Parametric Forecasting Source Models with Uncertainties: 10 Years Ahead Prediction of Sources for Electric Energy Production

Nonparametric Stochastic Differential Equations for Ultra-Short-Term Probabilistic Forecasting of Wind Power Generation

Probabilistic Wind Power Forecasting with Hybrid Artificial Neural Networks

A Machine Learning Approach to Forecasting Hydropower Generation

A Hybrid Approach for Probabilistic Forecasting of Electricity Price.

Power System Parameters Forecasting Using Hilbert-Huang Transform and Machine Learning

Short-term photovoltaic power production forecasting based on novel hybrid data-driven models

One-day ahead forecasting of energy production from run-of-river hydroelectric power plants with a deep learning approach

Short and long-term wind turbine power output prediction

Forecasting the Evolution of Hydropower Generation

A Hybrid Method for the Run-Of-The-River Hydroelectric Power Plant Energy Forecast: HYPE Hydrological Model and Neural Network

Day-Ahead Wind Power Forecasting Using a Two-Stage Hybrid Modeling Approach Based on SCADA and Meteorological Information, and Evaluating the Impact of Input-Data Dependency on Forecasting Accuracy

Forecasting Annual Power Generation Using a Harmony Search Algorithm-Based Joint Parameters Optimization Combination Model

A New Hybrid Model for Power Forecasting of a Wind Farm Using Spatial–temporal Correlations

An optimized deep learning approach for forecasting day-ahead electricity prices

Multi-step ahead forecasting for electric power load using an ensemble model

A hybrid approach based on autoregressive integrated moving average and least-square support vector machine for long-term forecasting of net electricity consumption

Adaptive machine learning for forecasting in wind energy: A dynamic, multi-algorithmic approach for short and long-term predictions

A hybrid approach based machine learning models in electricity markets

Enhancing One-Day-Ahead Probabilistic Solar Power Forecast With a Hybrid Transformer-LUBE Model and Missing Data Imputation

Probabilistic forecasting of photovoltaic power supply — A hybrid approach using D-vine copulas to model spatial dependencies