Stochastic Time Series Forecasting Approach Using Markov Chain for Robust Energy System Planning

Krishnan, Rohith Krishnan Bala, Voswinckel, Sebastian, Reinhardt, Theresa, Wesselak, Viktor, Schmidt, Christoph

ISES Solar World Congress 2025 · Fortaleza, Brazil · 2025-11-03
Published by International Solar Energy Society (ISES)
DOI: 10.18086/swc.2025.05.20

Abstract

Time series forecasting is essential for accurately modeling renewable energy system, particularly when dealing with the variability and uncertainty of generation from sources like Wind or Photovoltaics (PV). This paper addresses uncertainties in energy system modelling by proposing a Markov Chain integrated Machine Learning method for time series generation. Using photovoltaic feed-in profile as a case study, the stochasticapproach is validated against the historical data. Applied to a household PV system in Germany and optimized using the open energy modelling framework (oemof), the method generates multiple future scenarios. A comparative analysis reveals that this probabilistic approach provides a superior representation of uncertainty, resulting in more robust energy system outcomes compared to traditional methods.

Keywords

Time Series Forecasting, Renewable Energy, Photovoltaic, Markov Chain, Machine Learning, Uncertainty, Energy System Modelling, oemof, Germany, Stochastic Forecasting

← All proceedings papers