Author(s): Nicolas Hanousek; Reza Ahmadian

Linked Author(s): Nicolas Hanousek, Reza Ahmadian

Keywords: Tidal Range Energy; Error Analysis; Marine renewable energy; 0-D modelling; Energy optimisation

Abstract: With the increasing global desire for low-carbon electricity generation, tidal range schemes have been identified as a possible solution. Tidal Range Schemes (TRSs) are considered a reliable source of long-term renewable energy, which is both predictable and dispatchable. Due to their significant physical scale and the degree of complexity, various numerical modelling techniques are used to assess the potential outputs and impacts of the schemes. The preliminary design of TRSs includes a very large number of scenarios, due to variability of tides and a large number of options in terms of turbines and sluice gates. Simplified 0D modelling which has a limited computational requirement for each simulation has been widely used during the preliminary stages of design. 0D models assume that the lagoon has no impact on the external water level and as such the tidal time-series can be taken as a fixed input, allowing the impounded regime to be controlled by the head difference across the structure through turbines and sluice gates. The water levels at the location of a scheme are generally derived from hydro-environmental models that implement forms of the Navier-Stokes equations across a large domain. These models are typically calibrated against measured data available within their respective domain to minimize the general error across the domain. Errors however are ever-present in all models, to different levels, due to various reasons such as low grid resolution, inaccurate input data or measurements used for calibration, and more. At one site found in the literature, water levels from two different reports produced a difference of over 8% total energy. This work assesses the impact of different forms of error in the water levels generated and used in models on the performance of the TRSs. To do this, a set of baseline models was run at multiple TRS locations/configurations. Next synthetic errors, varying both in generation mechanic and magnitude, were applied to the water level time series and the models re-run using these modified series. The synthetic time series errors were tested against the original data using various statistical parameters, including R², Mean Average Error, Scatter Index, and measures of tidal range; and the final differences in total energy output and generation capacity of the lagoons compared to the baseline to ascertain the variations caused by the errors. This is to be used in the evaluation of data being used, and improve awareness of the likely biases present when assessing the outputs of models in the future.

DOI: https://doi.org/10.3850/IAHR-39WC252171192022369

Year: 2022