Simulating Wind Turbine Ice Throw: QBlade and Statistical Analysis

Ice throw and fall analysis using a semi-empirical Monte Carlo approach seems to be a good trade off between empirical formulation and complex full physics simulations. This paper looks at some of the challenges of running these simulations particularly at the uncertainty of certain input data. The field testing so far has helped provide data on a few of the key properties of the ice chunks. However, the output probability fields still suffer from a relatively high level of uncertainty due to the difficulty in collecting large sample sizes. A convergence study was able to establish that 60000 particle samples appear to be enough to converge the output probability field for the number of input variables selected. In a demonstration case using the new QBlade ice module, it was possible to demonstrate that it may be possible to avoid heating the inner sections of the wind turbine blades without increasing the risk compared to a standstill wind turbine.