Machine Learning and Weather Model Combination for PV Production Forecasting

Amedeo Buonanno,Giampaolo Caputo,Irena Balog,Salvatore Fabozzi,Giovanna Adinolfi,Francesco Pascarella,Gianni Leanza,Giorgio Graditi,Maria Valenti

DOI: https://doi.org/10.3390/en17092203

IF: 3.2

2024-05-04

Energies

Abstract:Accurate predictions of photovoltaic generation are essential for effectively managing power system resources, particularly in the face of high variability in solar radiation. This is especially crucial in microgrids and grids, where the proper operation of generation, load, and storage resources is necessary to avoid grid imbalance conditions. Therefore, the availability of reliable prediction models is of utmost importance. Authors address this issue investigating the potential benefits of a machine learning approach in combination with photovoltaic power forecasts generated using weather models. Several machine learning methods have been tested for the combined approach (linear model, Long Short-Term Memory, eXtreme Gradient Boosting, and the Light Gradient Boosting Machine). Among them, the linear models were demonstrated to be the most effective with at least an RMSE improvement of 3.7% in photovoltaic production forecasting, with respect to two numerical weather prediction based baseline methods. The conducted analysis shows how machine learning models can be used to refine the prediction of an already established PV generation forecast model and highlights the efficacy of linear models, even in a low-data regime as in the case of recently established plants.

energy & fuels

What problem does this paper attempt to address?

The paper is primarily dedicated to addressing the issue of photovoltaic (PV) power generation forecasting, especially in the context of significant variations in solar radiation, which necessitates accurate predictions for effective management of power system resources. The proposed method in the paper aims to improve forecasting accuracy by combining machine learning techniques with weather model-based PV power predictions. Specifically, the researchers tested various machine learning methods (including linear models, Long Short-Term Memory networks (LSTM), Extreme Gradient Boosting (XGB), and Light Gradient Boosting Machine (LGBM)) and compared these methods with two baseline methods based on numerical weather prediction. The results showed that the linear model performed the best in PV production forecasting, improving the root mean square error (RMSE) by at least 3.7% compared to the baseline methods. Additionally, the paper emphasizes the role of machine learning models in enhancing existing PV power generation forecasting models, and even in cases with limited data (such as newly established PV plants), the linear model still demonstrates good performance. In summary, the goal of this paper is to further improve the accuracy of PV power generation forecasting by integrating machine learning algorithms with existing techniques, thereby supporting the effective operation of microgrids and larger-scale power grids to address the challenges posed by the integration of renewable energy sources.

Machine Learning and Weather Model Combination for PV Production Forecasting

Machine Learning Models for Regional Photovoltaic Power Generation Forecasting with Limited Plant-Specific Data

Machine Learning Approaches for Short-Term Photovoltaic Power Forecasting

Comparative study of machine learning approaches for predicting short-term photovoltaic power output based on weather type classification

Machine Learning-Based Forecasting of Temperature and Solar Irradiance for Photovoltaic Systems

Machine learning methods in forecasting solar photovoltaic energy production

Photovoltaic Power Generation Forecasting with Hidden Markov Model and Long Short-Term Memory in MISO and SISO Configurations

Comparison of machine learning methods for photovoltaic power forecasting based on numerical weather prediction

Deep and Machine Learning Models to Forecast Photovoltaic Power Generation

On-grid and off-grid photovoltaic systems forecasting using a hybrid meta-learning method

Research on Photovoltaic power prediction technology Based on Machine Learning

Short-term predictability of photovoltaic production over Italy

Effect of Climate on Photovoltaic Yield Prediction Using Machine Learning Models

Analysis Of Solar Power Generation Forecasting Using Machine Learning Techniques

Improving Photovoltaic Power Prediction: Insights through Computational Modeling and Feature Selection

Combined Machine Learning and Physics-Based Forecaster for Intra-day and 1-Week Ahead Solar Irradiance Forecasting Under Variable Weather Conditions

A Comparison of the Performance of Supervised Learning Algorithms for Solar Power Prediction

A Machine Learning Model for Long-Term Power Generation Forecasting at Bidding Zone Level

Weather Impact on Solar Farm Performance: A Comparative Analysis of Machine Learning Techniques

A hybrid machine learning forecasting model for photovoltaic power

Short-Term Prediction of PV Power Based on Combined Modal Decomposition and NARX-LSTM-LightGBM