Combining Tidal Energy Yield Uncertainties

Abstract

A robust understanding of the uncertainty in a yield estimate for a tidal energy project is a key investor requirement and a common barrier to the commercialisation of the nascent sector. The Root Sum Squared (RSS) method is commonly used to combine the uncertainty in site resource (i.e. velocity, m/s) with the uncertainty in plant performance and losses (i.e. energy, GWh). The validity of the assumptions underlying RSS has been questioned in literature, particularly for early stage projects. RSS assumes that all uncertainties are independent and normally distributed, that the relation between yield and velocity is linear for small variations and that the combined yield uncertainty is also normally distributed. Monte Carlo Analysis (MCA) is a competing method for uncertainty analysis which is not limited by the same assumptions. This study quantitatively compares the combined yield uncertainty for 4 realistic test cases derived using the two different methods with the same input uncertainty distributions. An excellent agreement is found for cases where the uncertainties are relatively small and where the site resource is low relative to the turbine rated velocity. Some divergence in results is shown for projects with higher uncertainties but it is noted that these projects are likely to be early stage with a higher tolerance for inaccuracy in the uncertainty estimate. RSS predicts a higher P90 yield than MCA but it is prudent to adopt the more conservative view. The point at which the divergence occurs is hard to define as it is a complex function of site resource, turbine rated velocity and project uncertainties. As such, the confidence in RSS results is somewhat compromised, particularly for early stage projects.