A Lagrangian Two-Particle Model for Dense Gas Dispersion Driven by LES Fields

A Lagrangian two-particle dispersion model (L2PDM) driven by large-eddy simulations (LESs) of a stable boundary layer (SBL) has been developed and used to make concentration estimates for a dense gas release. A relatively simple dense gas (DG) model with ''slumping", gravitational spreading, and upper boundary or plume-top entrainment as parameterized by Briggs et al. (2001) have been included. This model has been formulated in both integral form, which gives the correct plume spread versus distance (e.g., Britter, 1989), and a "particle" form for compatibility and coupling with the L2PDM. The buoyancy-generated velocity from the model pertains to "relative dispersion" of the DG plume about its meandering centerline and has been superposed on the ambient velocity from the L2PDM. The SBL flow fields were obtained using the NCAR LES model (e.g., Moeng and Sullivan, 1994; Sullivan and Patton, 2011) and were generated using 200^3 grid points over the model domain; the SBL had a height of 207 m, a mean wind of 7 m/s, and a surface friction velocity of 0.28 m/s.

A key advantage of the coupled DG-LPDM-LES model is that one can generate individual "realizations" of the concentration field from which the concentration statistics can be obtained. Here, 15 realizations of the concentration field have been computed with the maximum surface concentration in the crosswind direction found as a function of downwind distance. The calculations have been made for 1- and 2-ton releases of chlorine gas in the Jack Rabbit I experiments (Hanna et al., 2012). We have found good agreement between the predictions and observations from these experiments. Currently, the model is being extended to short-duration releases, where evolution from a continuous plume to an elongated "cloud" or "puff" occurs.