Work

Sparse Variational Gaussian Processes model for prediction energy parameters in solar power plant and wind farm

Year: 2025

Type: conference-paper

Abstract: This study applies Sparse Variational Gaussian Processes (SVGP) for probabilistic forecasting of energy output from a solar plant and a wind farm with eight turbines.Trained on two years of hourly SCADA data, the SVGP model-combining a neural mean function and kernel-based covariance-produces accurate forecasts with confidence intervals over 24-120 hour horizons.Results confirm its scalability and precision thanks to evaluation metrics such as MAE, RMSE, CRPS, and R 2 across renewable energy types.

DOI: 10.51408/csit2025_63

Authors: Konstantin Koshelev, Sergei Strijhak, Ilia Stulov

Institution: Institute for System Programming

Language: English

Topic: Energy Load and Power Forecasting

Subfield: Electrical and Electronic Engineering

Field: Engineering

Domain: Physical Sciences

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

Open Access status: gold

OpenAlex ID: W4414454197

FWCI:

Cites:

Cited by:

Related to: