Academia.eduAcademia.edu

Figure 1 – uploaded by S. M. Tauseef

See full PDFdownloadDownload figure
Fig. 1. Computational domain with the cylindrical containment and cubical obstacle.

Figure 1 Computational domain with the cylindrical containment and cubical obstacle.

Related Figures (8)

Dense gas dispersion experiments in presence of obstacles. Table 1
Dense gas dispersion experiments in presence of obstacles. Table 1
Fig. 2. Wind profile at the beginning of the experiment showing the recirculation behind the obstacle. 2.1.1.2. Boundary B. Since the flow velocity and pressure are not known at this boundary and the simulation was performed with the assumption that freon-12/nitrogen mixture and air were incompressible, a zero diffusion flux was assumed for all flow variables at this boundary, which means a fully-developed flow is assumed at this boundary and the conditions are extrapolated from within the domain such that outflow velocity and pressure are consistent with the fully-developed flow assumption.
Fig. 2. Wind profile at the beginning of the experiment showing the recirculation behind the obstacle. 2.1.1.2. Boundary B. Since the flow velocity and pressure are not known at this boundary and the simulation was performed with the assumption that freon-12/nitrogen mixture and air were incompressible, a zero diffusion flux was assumed for all flow variables at this boundary, which means a fully-developed flow is assumed at this boundary and the conditions are extrapolated from within the domain such that outflow velocity and pressure are consistent with the fully-developed flow assumption.
Time from start of experiment/simulation and maximum recorded feron-12 concentration (% v/v). * Values obtained through digitization of reported graphs in literature (Sklavounos & Rigas, 2004). Table 2
Time from start of experiment/simulation and maximum recorded feron-12 concentration (% v/v). * Values obtained through digitization of reported graphs in literature (Sklavounos & Rigas, 2004). Table 2
Fig. 3. Turbulent field developed as a result of initial wind field, k (m/s).
Fig. 3. Turbulent field developed as a result of initial wind field, k (m/s).
Fig. 4. Gas cloud after 2 s of release. For the present case, the problem was initially solved in steady state to obtain initial values for the transient simulation (Figs. 2 and 3). The steady state runs converged after approximately 150—200 iter- ations. The convergence criterion was set as the residual RMS (residual of root mean square) becoming equal or less than 10~4 (Ferziger & Peri, 2002). After obtaining steady state wind and turbu- lence profiles at the start of the experiment the dispersion was simulated as a function of time. The transient runs required 10—20
Fig. 4. Gas cloud after 2 s of release. For the present case, the problem was initially solved in steady state to obtain initial values for the transient simulation (Figs. 2 and 3). The steady state runs converged after approximately 150—200 iter- ations. The convergence criterion was set as the residual RMS (residual of root mean square) becoming equal or less than 10~4 (Ferziger & Peri, 2002). After obtaining steady state wind and turbu- lence profiles at the start of the experiment the dispersion was simulated as a function of time. The transient runs required 10—20
Fig. 6. Comparison of results for experimental, standard and realizable k—¢ models at an elevation of 6.4m on the windward side of the obstacle. Figs. 4 and 5 show the pattern of dispersion at 2 s and 10 s after release, respectively. As is widely recognized and also observed during experiments (Deaves, 1992; Havens, 1992; Lees, 2005; Seeto & Bowen, 1983), atmospheric dispersion of gases denser than air may involve several fluid flow regimes: buoyancy-dominated, stably-stratified and passive dispersion. During the simulation done by us it was possible to observe all these fluid flow regimes. In the buoyancy-dominated regime, gravity-induced-slumping and
Fig. 6. Comparison of results for experimental, standard and realizable k—¢ models at an elevation of 6.4m on the windward side of the obstacle. Figs. 4 and 5 show the pattern of dispersion at 2 s and 10 s after release, respectively. As is widely recognized and also observed during experiments (Deaves, 1992; Havens, 1992; Lees, 2005; Seeto & Bowen, 1983), atmospheric dispersion of gases denser than air may involve several fluid flow regimes: buoyancy-dominated, stably-stratified and passive dispersion. During the simulation done by us it was possible to observe all these fluid flow regimes. In the buoyancy-dominated regime, gravity-induced-slumping and
Fig. 7. Comparison of results for experimental, standard and realizable k—¢ models at an elevation of 0.4 m on the leeward side of the obstacle.
Fig. 7. Comparison of results for experimental, standard and realizable k—¢ models at an elevation of 0.4 m on the leeward side of the obstacle.
Related topics:
EngineeringChemical EngineeringModeling and SimulationChemical Process SafetyComputation Fluid DynamicsComplex terrainDense Gas DispersionTurbulence ModelCFD simulation
580 California St., Suite 400
San Francisco, CA, 94104
© 2025 Academia. All rights reserved