Modeling and Predicting Uncertainty in Tidal Turbine Power Output: A Data-Driven Time-Series Approach

The integration of sustainable energy sources into modern power grids is crucial for achieving a more resilient, equitable, and environmentally sustainable electricity infrastructure. As the demand for renewable energy increases, tidal turbines have emerged as a promising solution for expanding power generation and contributing to a more diverse mix of sustainable energy sources. However, tidal energy is influenced by variations in environmental and hydrodynamic conditions, which affect tidal turbine power output. As a result, the amount of power a tidal turbine generates at any given moment remains uncertain. This uncertainty poses significant challenges for grid energy management and maintaining a supply-demand balance. To address this issue, this study introduces a systematic approach to quantifying, modeling, and predicting uncertainty in tidal turbine power output. Unlike conventional methods that focus on variability, this framework defines uncertainty as unpredictability and applies time-series modeling to characterize and forecast uncertainty in power generation. The methodology employs an Autoregressive Integrated Moving Average (ARIMA) model to capture predictable patterns, while a Generalized Autoregressive Conditional Heteroskedasticity (GARCH) model quantifies and predicts uncertainty in tidal turbine power output. The proposed method was implemented using high-resolution experimental data from controlled tidal turbine tests conducted at the FloWave facility at the University of Edinburgh. The results confirm that the GARCH model effectively predicts future uncertainty levels. The findings establish a foundation for advancing tidal energy deployment and enhancing the integration of renewable resources into the power grid.