1 Risø National Laboratory for Sustainable Energy, Technical University of Denmark2 Department of Wind Energy, Technical University of Denmark3 unknown
The COFIN project studied existing remote-sensing Lidar data on concentration fluctuations in atmospheric dispersion from continuous sources at ground level. Fluctuations are described by stochastic models developed by a combination of statisticalanalyses and surface-layer scaling. The statistical moments and probability density distribution of the fluctuations are most accurately determined in a frame of reference following the instantaneous plume centreline. The spatial distribution of thesemoments is universal with a gaussian core and exponential tails. The instantaneous plume width is fluctuating with a log-normal distribution. The position of the instantaneous plume centreline is modelled by a normal distribution and a Langevin equation,by which the meander effect on the time-averaged plume width is predicted. Fixed-frame statistics are modelled by convolution of moving-frame statistics and the probability distribution for the plume centreline. The distance-neighbour function generalizedfor higher-order statistics has a universal exponential shape. Simulation tools for concentration fluctuations have been developed for either multiple correlated time series or multi-dimensional fields. These tools are based on Karhunen- Lo`eve expansionand Fourier transformations using iterative or correlation-distortion techniques. The input to the simulation is the probability distribution of the individual processes, assumed stationary, and the cross-correlations of all signal combinations. The usein practical risk assessment is illustrated by implementation of a typical heavy-gas dispersion model, enhanced for prediction and simulation of concentration fluctuations.