The influence of open boundary location on tidal lagoon modelling

Abstract

The Bristol Channel and Severn Estuary comprise the area most thoroughly investigated for tidal lagoons development, due to its second largest tidal range in the world and the high demand for clean electricity in the surrounding area. Accurate hydrodynamic modelling of tidal lagoons is a solid foundation for predicting potential electricity generation and environmental impact assessment. However, it is reported that in correct selection of an open boundary may amplify any disturbance associated with the tidal lagoons by affecting the resonant modes. Thus a model that simply held the identical open boundary condition for pre- and post-lagoons conditions may contain inaccuracies in the electricity generation and the impacts on the hydrodynamics of the region. To investigate the influence of open boundary location on tidal lagoon modelling, the West Somerset Lagoon (WSL) was simulated using different hydrodynamic models with different open boundary locations. Two hydrodynamic models were established using the TELEMAC system, one of which covers the whole Bristol Channel and Severn Estuary (SEBC) as the most prior research used. Another one is a Continental Shelf (CS) model, which was centred on the Bristol Channel, and has its open boundary extended beyond the Continental Shelf. Both SEBC and CS models were run for pre- and post-WSL, to achieve the power output of WSL and the hydrodynamic impact in each model. The WSL was introduced into both hydrodynamic models using the domain decomposition method, and full momentum conservation was achieved by refining the momentum source terms at the turbine locations. Although the hydrodynamic influences were generally similar between CS and SEBC models, results showed the influence of WSL on water level extended to the outer Bristol Chanel in the CS model, with over 10 cm decrease of tidal range on the location of the open boundary of SEBC model. However, there was a minor difference in far-field velocities prediction between the two models. The annual energy generation of WSL using the different models showed slight differences, i.e. less than 6%. However, this could also be exacerbated by the fact that similar operation was used in both scenarios.. This study concludes that  SEBC could be considered as a suitable model for early-stage studies and preliminary environmental impact modelling due to lower computational and set up time requirements. However, for later stages of the TRS design, such as power prediction for accurate revenue assessment and business case development, then a more precise open boundary condition is expected to be needed, either by extending the model domain to the Continental Shelf or theoretically modifying the open boundary characteristics.