Work

Improving Renewable Energy Forecasting With a Grid of Numerical Weather Predictions

Year: 2017

Type: article

Abstract: In the last two decades, renewable energy forecasting progressed toward the development of advanced physical and statistical algorithms aiming at improving point and probabilistic forecast skill. This paper describes a forecasting framework to explore information from a grid of numerical weather predictions (NWP) applied to both wind and solar energy. The methodology combines the gradient boosting trees algorithm with feature engineering techniques that extract the maximum information from the NWP grid. Compared to a model that only considers one NWP point for a specific location, the results show an average point forecast improvement (in terms of mean absolute error) of 16.09% and 12.85% for solar and wind power, respectively. The probabilistic forecast improvement, in terms of continuous ranked probabilistic score, was 13.11% and 12.06%, respectively.

Source: IEEE Transactions on Sustainable Energy

DOI: 10.1109/tste.2017.2694340

Authors: Jose R. Andrade, Ricardo J. Bessa

Institution: INESC TEC

Language: English

Topic: Energy Load and Power Forecasting

Subfield: Electrical and Electronic Engineering

Field: Engineering

Domain: Physical Sciences

SDGs: Climate action, Affordable and clean energy, Industry, innovation and infrastructure

Open Access status: green

Funder: European Regional Development Fund