Computational fluid dynamics in filters - Essay Example

CFD FOR FILTERS COMPUTATIONAL FLUID DYNAMICS FOR FILTERS CFD FOR FILTERS 2 COMPUTATIONAL FLUID DYNAMICS (CFD) IN FILTERS Computational- Involving rigorous mathematics and intensive computational algorithms. Fluid- Anything that has a tendency to flow, i.e. all liquids and gases. Dynamics- Scientific discipline concerned with the study of particles in motion. Filters- A physical interface that removes particles/pollutants from polluted fluid flows. OPERATIONAL DESIGN OF THE SYSTEM [1] Essentially, a CFD simulation protocol is so computationally intensive and involves such esoteric mathematical skills (because it involves the solution of non-linear partial differential equations) that individual efforts at completion of these Herculean tasks is next to impossible. So, recently many software companies have sprung up that provide requisite software tools for the same (most notably FLUENT). The operational protocol of such companies (e.g. FLUENT) is briefly outlined below. Basically, every CFD simulation consists of three elementary stages: PRE-PROCESSING [1] A virtual prototype of the fluid model to be examined is built within the ambit of a Computer Aided Design (CAD) package which is characterized by a unique and suitable computational mesh which in turn is created after accounting the boundary conditions and specific fluid material properties for the operation. Standard preprocessing software tools such as GAMBIT, TGrid and G/Turbo are supplied to the customers by these companies. The rest of the protocol is efficiently conveyed in the following statement in the official fluent.com home page, "CAD geometries are easily imported and adapted for CFD solutions in GAMBIT, Fluent's own preprocessor. 3D solid modeling CFD FOR FILTERS 3 options in GAMBIT allow for straightforward geometry construction as well as high quality geometry translation. Among a wide range of geometry tools, Boolean operators provide a simple way of getting from a CAD solid to a fluid domain. A state-of-the-art set of cleanup and conditioning tools prepares the model for meshing. GAMBIT's unique curvature and proximity based "size function" produces a correct and smooth CFD-type mesh throughout the model. Together with our boundary layer technology, a number of volumetric meshing schemes produce the right mesh for your application. Parametric variations are also inherent to the process." For varied computational mesh requirements, other meshing tools as ANSA, Harpoon, Sculptor and YAMS are available. SOLVING [1] This step involves computer simulations of real world conditions by evaluating and assessing product functional efficiency in the specified boundary conditions. Several commercially available suites of softwares, most notable of them being FLUENT, FloWizard, FIDAP, and POLYFLOW (from FLUENT Corporation) boasts of intensive and flexible parallel computing capabilities that enables faster and accurate modeling by solving flow dynamics mathematics involving Navier-Stokes and Eulerian equations. An ideal software suite should have the following attributes- (a) An interactive platform that allows changes to be affected during analysis which saves time and enable more efficient refinement of designs thereby making the learning curve shorter and modeling process faster. (b) Physics and interface functions should be customizable according to design requirements. CFD FOR FILTERS 4 (c) Computational mesh capability should be dynamic and adaptive enough to be compatible with a wide range of physical prototypes which will enable modeling complex moving objects in relation to flow. POST-PROCESSING [1] It is the final step in CFD simulation where the data gathered in the previous computing step is harvested and analyzed in detail to provide a layman's interpretation of the same for broader comprehension and interpretation. Several layers of reporting of the same set of data can be done according to the audience, varying from sophisticated quantitative data analysis at one extreme to high resolution images and entertaining animations on the other. Commercially available post processing and visualization tools include Ensight, Fieldview and TechPlot. E.g. coupling FLUENT CFD solutions with structural codes such as ABAQUS, MSC and ANSYS, as well as to other engineering process simulation tools give dramatic interpretations of monotonous mathematical solutions. APPLICATION OF COMPUTATIONAL FLUID DYNAMICS 1. CFD can be utilized to model a utopian hospital environment (particularly operation theaters, oncology, maternity, ICU and transplant wards which house immuno-compromised and/or critical patients) with minimal infection by opportunistic air-borne pathogens and hence a higher sterility in crucial areas/operations [2]. Air quality is affected by many crucial parameters, the most crucial of them being temperature, air-flow patterns, air velocity and overall air delivery and distribution. CFD allows for a detailed analysis of these key variables in isolation or in conjunction with each other and thereby CFD FOR FILTERS 5 optimize them to ensure sterile air delivery to critical surgical sites, air ventilation with optimum velocities and comfortable temperature conditions. Moreover, existing flaws in the air ventilation and delivery infrastructure can be identified by CFD and suitably rectified by strategic positioning of filters [2]. 2. Membrane filtration technologies lie at the heart of liquid-solid and emulsion purification in a wide range of industrial applications (notably industrial and household sewage treatment). But accumulation of particles near the membrane fluid interface leads to decline in the flux of the permeant filtrate and a concomitant increase in transmembrane pressure (called membrane fouling). [3]Increasing the shear-stress at the interface leads to reduced particle deposition which can be accomplished by applying a tangential fluid flow at the interface and/or rotating the membrane filter itself. In the former, shear stress and fluid flow rate are linked, while in the latter they are independent of each other and hence rotating membrane filters has the potential to achieve high filtration rates even in highly concentrated suspensions. Now shear stress can only be gained by energy expenditure which must be minimum in any membrane filter design for it to be commercially feasible. This is where CFD is invaluable as by taking into account the modalities of fluid flow at the membrane interface (during application of shear stress) can an optimum design be arrived at with minimum of energy consumption and little membrane fouling over extended periods of time [3] CFD FOR FILTERS 6 BENEFITS OF COMPUTATIONAL FLUID DYNAMICS 1. Some filter designs are too complicated (e.g. rotating membrane disk filters) and/or are installed in such environments (e.g. DPF, diesel particulate filters) that it is virtually impossible to make a prototype of it and/or experimentally evaluate its operational efficiency. In such cases, where models are inaccessible to experimental validation, simulation of their operation and efficiency over a requisite time scale can be easily performed using CFD. Thus, CFD provides a suitable interface to assess the efficiency and a deeper insight into the flaws as well as strengths of the design. [1] 2. Filters come in all shapes and sizes and operate under a plethora of environments. While diesel particulate filters (DPF) are used for filtering noxious pollutants from diesel engine exhausts, filters in sewage treatment plants have a completely different working environment as they encounter liquid rather than gaseous flow. So there is no universal filter prototype that operates under all environmental variables and filters all kinds of fluids. Moreover, as enumerated earlier, most of these prototypes are not amenable to experimental manipulation. CFD is basically a predictive software that evaluates the mode and efficiency of operation of a particular design within a given set of boundary conditions for the key variables. Thus by permutation and combinations of these variables and design modules available, a manufacturer can actually arrive at an optimal filter design for a given environment without actually manufacturing it thereby saving time and expenditure. 3. Various methods, both classical and contemporary, are available for predicting efficiency of fibrous filters. [4] A comparative analysis of these diverse methods by using CFD FOR FILTERS 7 CFD gives a clear insight into the efficiency of these predictive tools. The Most Penetrating Particle Size (MPPS) of any filter can be determined by challenging a filter to submicrometer aerosols and determine the size of the particles in the filtrate by Mass Spectrometry. Having determined the MPPS, filter efficiency can be predicted by a plethora of predictive tools as well as by CFD (theoretically). The efficiency of these tools in turn can be evaluated by a comparison with the ideal optimum efficiency as evaluated by CFD simulation. [4] EQUATIONS IN CFD [5] Equations that form the core of CFD is based a very fundamental set of equations whose solutions are exceedingly complex when applied to real life scenarios and hence the requirement for intensive computational algorithms. NAVIER-STOKES EQUATION [5] (Claude-Louis Navier and George Gabriel Stokes) are a set of partial differentiation equation characterized by the following attributes: (a) Describe the motion of fluids, i.e. liquids and gases. (b) An extension of Newton's second law. It states, "Changes in momentum in infinitesimal volumes of fluid are simply the sum of dissipative viscous forces, changes in pressure, gravity, and other forces acting inside the fluid." (c) Can't specify the explicit relationship between the variables of interest, instead they can establish the relation among their rates of change. E.g. acceleration (rate of change of velocity) produced in an inviscous ideal fluid is directly proportional to the pressure gradient (rate of change of pressure with distance). CFD FOR FILTERS 8 (d) Solutions of these equations constitute a unique velocity/flow field- A representation of the velocity of the fluid at a given point in space and time. (e) Are non-linear in real situations. Non-linearity is due to convective (flow) acceleration- Acceleration associated with change in velocity over position- which makes problems infinitely difficult to model and solve without huge computations. FORM OF THE EQUATIONS [5]: Navier-Stokes equations basically espouse the conservation of mass, momentum and energy for an infinitesimally small volume of liquid. In its most general form, it can be written as: Or Where (a) - Density of the fluid. (b)Dv/Dt- Rate of change of velocity. (c) -Pressure gradient that is present in any non-ideal fluid flow. (d) is a measure of shear forces in the liquid, e.g. viscous effects. (e) - represents other forces acting on the fluid, as gravitational forces. term has too many unknowns, so the general form of the equation is not directly applicable to real life problems. So, it is quantified in terms of familiar variables like CFD FOR FILTERS 9 velocity for applicability. So, under assumptions that the fluid flow is incompressible and Newtonian, =. Navier-Stokes equations are strictly a statement of conservation of momentum. To make it more comprehensive and informative, statutes concerned with conservation of mass and energy is incorporated into it to form the continuity equation. Under assumptions of negligible temperature effects and incompressible fluid flow, density remains constant, and continuity equation reduces to,. This is more accurately a statement of the conservation of volume. The most applicable form of Navier-Stokes equation is obtained by assuming an incompressible flow and constant viscosity (Newtonian Fluids): Where, (a) f- other body forces, gravity/centrifugal force. (b) Only convective terms are non-linear. (c) Convective acceleration is an acceleration caused by the change in velocity over position, e.g. increase in velocity of a fluid when it enters a constricted zone in its path. (d) Viscosity is diffusion of momentum and is the Laplacian of the velocity field. CFD FOR FILTERS 10 REPRESENTATION [5]: (1) Cartesian coordinate system: The continuity equation becomes, gx,gy,gz- gravitational forces as body forces. Can also be represented in cylindrical and spherical coordinates. APPLICABILITY OF NAVIER-STOKES EQUATIONS AND BOUNDARY CONDITIONS [5]: Navier-Stokes equations are so generic in nature that their application to real life problems can be as varied and complicated as the number of problem models available. To attenuate the complexity of the task, certain flow assumptions and initial/boundary condition specification is followed by scale analysis. The Boundary Element Method (BEM) attempts to use the given boundary conditions to fit boundary values into the integral equation, rather than values throughout the space defined by a partial differential equation which makes the computational task much easier to perform. An illustrative example follows: CFD FOR FILTERS 11 Assuming steady, parallel, one dimensional, nonconvective pressure driven flow between parallel plates, the resulting scaled (dimensionless) boundary value problem is: The boundary condition is the no slip condition. This problem is easily solved for the flow field: Now, more quantities of interest can be determined, such as viscous drag force or net flow rate. EXAMPLES FROM SIMULATED CASES AND POTENCY OF CFD [6]: The following is a beautiful illustration of a real life application of CFD in enhancing filter efficiency in air treatment processes. Efficiency of CFD application in filters is also very nicely evident in the following excerpt from online peer-reviewed journal- SCIENCE DIRECT- doi:10.1016/S0009-2509(99)00441-8. 'Numerical simulation of flows in air treatment devices using activated carbon cloths filters.' Chemical Engineering Science. Volume 55, Issue 10, May 2000, Pages 1807-1816. Jean-Nol Balo, and Albert Subrenat, Pierre Le Cloirec cole des Mines de Nantes, Dpartement Systmes nergtiques et Environnement, La CFD FOR FILTERS 12 Chantrerie, 4 rue Alfred Kastler, BP 20722, 44307 Nantes Cedex 3, France Received 3 February 1999; accepted 10 July 1999. Available online since 12 January 2000. [6]. "The determination of pressure drops is one of the most important considerations in the design and the sizing of air treatment processes, which often use activated carbon granules, beads in packing columns, or more recently, carbon cloths, for the removal of volatile organic compounds (VOC) or odorous molecules charged in the air (see for instance Suzuki, 1990; Crittenden & Thomas, 1998; Dabrowski, 1999). The basic engineering parameters describing air flows through filters in air treatment processes are the pressure drop across the filter and the penetration (see for instance Payet, Boulaud, Madelaine & Renoux, 1992). These parameters depend upon the filter structure (packing density, fiber radius), the operating conditions (mainly the velocity) and the characteristics of the aerosols filtered ( Thomas, Contal, Renaudin, Penicot, Leclerc & Vendel, 1999). Basically, the behavior of a filter over time encompasses a first steady phase in which the deposition of particles in the filter is sufficiently negligible so that the filter efficiency is unaffected (i.e. the pressure drop does not vary with time), and a second unsteady phase, in which the formation of aggregates and the loading of the filter affect the pressure drop. A model for the prediction of the pressure drop during this second phase can be found in Thomas, Contal, Renaudin, Penicot, Leclerc & Vendel, 1999. The maximum allowable pressure drop controls the minimum filter area, for a given thickness of porous filtration media, and at a given flow rate. Hence, the cost of CFD FOR FILTERS 13 filtration in a process may be accurately minimized by predicting the pressure losses, since the quantity of adsorbent material in treatment systems is a compromise solution between the surface and thickness of filtration material needed to remove VOCs, and the sizing of the aerodynamic system. Solving conservation (mass and momentum) and coupled transport equations linked to the turbulent nature of the flow, using classical computational fluid dynamics (CFD) techniques such as the finite volume method, provides the spatial distribution of relevant variables describing the flow. The CFD analysis has been thoroughly used for the simulation of fluid flow through filter media represented by parallel or staggered arrays of rectangular and cylindrical arrangements of fibres (see for instance Fardi & Liu, 1992; Liu & Wang, 1996) in two dimensions, or more recently in complex three-dimensional fibre arrangements ( Dhaniyala & Liu, 1999). In these studies, a periodic spatial structure of the porous media is considered and the evolution of the pressure drop or of the drag coefficient is essentially analysed as a function of the packing density of the media and as a function of the Reynolds number. Lu, Tung & Hwang, 1996 studied the effect of woven structures on fluid flow through basic weaves of monofilament filter cloths in three dimensions using CFD. However, in practice, relatively little data is available concerning the hydrodynamic analysis of air flows through industrial filters (and especially ACC filters) from the engineering point of view. CFD FOR FILTERS 14 The CFD simulations have nevertheless been previously used, as for instance, for the simulation of air flow through heavy-duty filter elements in automotive engine air intake systems (Brown, Cheng, Borgia, Rosenthal & Blaylock, 1988) or in the case of a laminar air flow through perpendicularly pleated filters, for the optimization of the filter design ( Chen, Pui & Liu, 1995). In this paper, focus is made on the possibility of predicting accurately quantitative pressure losses in different modular configurations and on the analysis of the associated flow patterns in the global air treatment device, rather than on the analysis of the flow microstructure in the media or on the geometrical design of particular filters. The latter are composed of activated carbon cloths, which show interesting adsorption properties for the treatment of water (Le Cloirec, Brasquet & Subrenat, 1997) and air ( Brasquet & Le Cloirec, 1997). The analysis of spatial distributions of flow variables provides quantitative information that can be used in order to improve such treatment systems from the aerodynamic behavior point of view. However, to be precise, the simulations must previously be validated. Presently, this validation is performed by comparing calculated and measured pressure drops." FUTURE RESEARCH INITIATIVES IN DEVELOPMENTS OF FILTERS UTILISING THE TECHNOLOGY OF COMPUTATIONAL FLUID DYNAMICS The following is an interesting account of a collaboration between a CFD application corporation and an automobile giant with a view to process and design more efficient and environmentally compatible Diesel Particulate Filters (DPF) for purification CFD FOR FILTERS 15 of diesel engine exhausts [7]. If such healthy trends become a way of life, then filters with previously unheard of potential will soon flood the markets. The account as appeared in Filtration Industry Analyst, Volume 2006, Issue 5, May 2006, Page 2 is reproduced below: "Fluent Inc, a developer of computational fluid dynamics (CFD) software and services, and J Eberspacher GmbH, a German automotive parts manufacturer, have teamed up to develop a diesel particulate filter simulation tool. This new tool will allow companies to computationally design and optimize particulate filters in passenger car and truck applications with regard to soot loading and regeneration behavior for a filter's projected full lifecycle, before it is even manufactured, the companies says. Fluent's German office, based in Darmstadt, and Eberspacher have created the combined 3D/1D modelling tool to calculate diesel filter soot distribution, pressure, and regeneration under any operating condition. "Our collaboration with Eberspacher has produced a unique design methodology that I believe will transform the design process for complex automotive diesel filters," claims Werner Seibert, global automotive segment manager for Fluent Inc. "The net effect will be a new tool to improve filter design and thereby reduce particulate emissions at the early design stage which should lead to an improved environment for everyone." [7] REFERENCES [1] http://www.fluent.com/software/index.htm. Fluent is a market leader in Computational Fluid Dynamics Services and Applications. [2] Indoor and Built Environment, Vol. 12, No. 1-2, 81-88 (2003) DOI: 10.1177/1420326X03012001013 2003 International Society of the Built Environment Computational Fluid Dynamics Applications in Hospital Ventilation Design John Colquhoun ,Bassett Applied Research, Australia. Lester Partridge, Bassett Applied Research, Australia. Source- Online SAGE Journals, http://www.sagepub.co.uk/ [3] Chemical Engineering Journal Volume 72, Issue 1, 29 January 1999, Pages 1-17. Rotating membrane disk filters: design evaluation using computational fluid dynamics Christophe A. Serraa, Mark R. Wiesnera, * and Jean-Michel Lanb a Department of Environmental Science and Engineering, Rice University, 6100 Main Street Houston, TX 77005 USA b CIRSEE, Suez - Lyonnaise des Eaux Le Pecq France Received 8 January 1998; accepted 29 July 1998. Available online 2 February 1999. [4] Journal of Aerosol Science Volume 27, Supplement 1, September 1996, Pages S625-S626. Comparison of classical and contemporary methods for prediction of fibrous filter efficiency using Computational Fluid Dynamics (CFD) D. P. Mortimera, I. Pottsa and T. H. Frosta a Department of Mechanical, Materials and Manufacturing Engineering, University of Newcastle-upon-Tyne Newcastle-upon-Tyne, NE1 7RU England Available online 12 June 2002. [5] Navier-Stokes equations From Wikipedia, the free encyclopedia. http://en.wikipedia.org/wiki/Navier-Stokes_equations. [6] Online peer-reviewed journal- SCIENCE DIRECT- doi:10.1016/S0009-2509(99)00441-8. 'Numerical simulation of flows in air treatment devices using activated carbon cloths filters.' Chemical Engineering Science. Volume 55, Issue 10, May 2000, Pages 1807-1816.Jean-Nol Balo, and Albert Subrenat, Pierre Le Cloirec cole des Mines de Nantes, Dpartement Systmes nergtiques et Environnement, La Chantrerie, 4 rue Alfred Kastler, BP 20722, 44307 Nantes Cedex 3, France Received 3 February 1999; accepted 10 July 1999. Available online since 12 January 2000. [7] Filtration Industry Analyst Volume 2006, Issue 5, May 2006, Page 2. Online peer-reviewed journal- SCIENCE DIRECT: doi:10.1016/S1365-6937(06)71122-4. News :Collaboration formed to simulate diesel filters Available online 20 May 2006. Read More