A Probabilistic Scenario Generation Framework for Optimal Decision Making in Turkish Renewable Energy Market

Turkey is one of the richest countries in terms of renewable energy resources. At the same time, the largest portion of the account deficit of Turkey is due to energy import. Optimization studies for design, integration and management of renewable energy is therefore crucial in terms of increasing overall energy efficiency. In addition, energy sources and demands in Turkey have significant uncertainty and meeting the market conditions in a profitable manner is a challenging task. Stochastic programming is an efficient approach to deal with the aforementioned challenge. It requires introducing representative and comprehensive scenarios for the optimal design and scheduling. In this study, the aim is to propose a systematic and generic scenario generation method which is compatible with historical data, dependable for forecasts, and easily tunable for the scenario likelihood. Main contributions of this work are as follows: The uncertainty in the model parameters are propagated to the forecasts to obtain prediction intervals under a desired confidence, which provides probabilities of each scenario over the whole time horizon with predetermined likelihood the generated scenario set (e.g., likely, rare or statistically very low probability). Thus, scenario reduction is not needed. We implemented our method for Yalova, a developing city in Turkey, for the scenario generation of wind speed, population, air temperature, electricity consumption and solar irradiance, with a prediction horizon of 20 years. We also developed a preliminary mixed-integer linear programming (MILP) decision making model which computes both the optimal equipment investments and the optimal sub-hourly scheduling sequences of the equipment based on these scenarios and economic considerations.