Posters

Abstract

With the growing integration of renewable energy sources such as solar or wind energy, short term power system operations face increasing complexity primarily due to non-dispatchable and intermittent nature of these resources. Reserve procurement is the primary tool to manage renewable uncertainty, ensuring sufficient backup capacity to balance fluctuations. These reserves may include a generation unit operating below its rated capacity, a large curtailable load, or a battery energy storage system with available charging or discharging capacity. However conventional approaches – typically based on fixed percentage of renewable capacity – may often result in either excessive costs or inadequate reserves. This challenge presents an important trade-off between ensuring reliable operation of the power system and maintaining cost efficiency.  Conventionally, scenario-based optimization has been widely adopted as a tool for handling uncertainty in power systems. By enforcing feasibility across sampled realizations, it provides robustness against a range of possible outcomes. However, two important drawbacks limit its effectiveness. First, requiring full coverage of scenarios may lead to overly conservative schedules and significantly higher operating costs. Second, as the number of scenarios increases, the optimization program becomes computationally burdensome – requiring the use of parallel computing and decomposition-based strategies. To address these challenges, this study investigates the development of a stochastic unit commitment framework for efficient handling of uncertainty arising from renewable energy sources, thereby enabling reliable and cost-efficient operation of the power system. The proposed approach utilizes recent advancements in probabilistic forecasting models that provide point prediction together with an estimate of the associated uncertainty distribution. The information about uncertainty distribution is then used to optimize the scheduling of conventional generators and time dependent reserve requirements. In addition, the framework also incorporates probabilistic network, operational and security constraints ensuring that system reliability, grid limitations and contingency requirements are satisfied with high probability under uncertainty.