The potential of Computational Fluid Dynamics (CFD) for efficient pulse-jet cleaning of fabric filters

Pulse-jet is a well-established technique for cleaning of fabric filters, which consists in exploiting the mechanical (shaking) action of a pressure wave produced by a com-pressed air jet and travelling across the bag filter to remove the dust cake from its surface. Considerable attention has been devoted to making the process more efficient and cost-effective, resulting in a collection of best practices and guidelines obtained from the professional experience of practitioners, often complemented by experimental testing on pilot plants at the laboratory scale. Here a different approach is adopted, namely, the computer simulation of the air flow through Computational Fluid Dynamics (CFD). Key advantages of CFD compared to the more traditional laboratory testing are the large amount of distributed information that a computer simulation can provide, the (relatively) low economic cost, as well as the (virtual) possibility to simulate any flow process regardless of its complexity and geometrical scale. At the same time, a serious issue of CFD resides in the high computational burden of the calculations. In order to make CFD a practical approach, the fluid flow equations obtained from the basic phys-ical principles are mathematically manipulated before being numerically approximated. This produces a number of coefficients, parameters, and closure equations which must be decided by the user and are, in the end, not well characterized and empirical in nature. So, the real challenge of CFD users is not only (and not so much) to solve the equations but, rather, to ensure that the combination of settings (sub-models, param-eters etc.) allows for reliable estimation of the quantities of interest. The present contribution aims at disclosing the potential of CFD in the application to pulse-jet cleaning of fabric filters. Particularly, the scope is to assess the capability of CFD as an effective engineering tool, satisfying the requirements of accuracy, robust-ness, and computational performance. Referring to a simplified benchmark case mim-icking the core area of the cleaning line (that is, the injector-bag system), an extensive analysis was carried out to quantify the influence of the CFD settings which can pro-duce uncertainty to the numerical results. These include both numerical settings, e.g. related to space-time discretization and solution strategies, and modelling settings, e.g. those related to the modelling of turbulence and to the modelling of the flow through the porous bag. Finally, the CFD setup was applied to investigate the influence of ge-ometrical and flow parameters on the numerical solution, so that possible directions are identified for a more energy efficient design of the pulse-jet cleaning process.