Today, CFD modeling provides a powerful tool for engineers, researchers and other technical professionals to visualize what’s going on inside most types of chemical process equipment, such as mixers, reactors, pumps, combustion systems, material-handling equipment, pollution control systems, pipelines and other types of complex process vessels. Among the countless discussions at the ACHEMA-exhibition and at the international ACHEMA Congress this will be one of the topics. More than 4,000 exhibitors and 180,000 visitors from all over the world will meet from 11-15 May 2009 in Frankfurt/Germany at the leading international event for equipment suppliers to the chemical industry and all branches of the process industry. ACHEMA is once again set to spark off trend-setting impulses for technology developments, worldwide contacts and new business networks.

Computational fluid dynamics (CFD) software enables the numerical solution of the physics-based equations (otherwise known as the Navier-Stokes equations) that govern the conservation of mass, momentum and energy. By solving these equations at several thousand discrete points on a computational grid that is set up to approximate the geometry of the equipment component or system being modeled, CFD programs are able to effectively simulate such critical process phenomena as fluid flow, mass and heat transfer and chemical reactions. And, when CFD modeling is combined with other types of modeling software (discussed below), users are able to not only understand fluid behavior in the system, but to analyze and predict structural response, fatigue and vibration characteristics in the machinery.

CFD modeling has been available for several decades. However, in the early years, the modeling technique was mainly used by individuals in academia and industry who had the specific academic background and extensive training that was needed to work with such specialized software. Today, however, ongoing advances in all aspects of CFD modeling have made off-the-shelf programs faster, more powerful, more intuitive, and easier to use than ever before with only minimal training, and this has vastly increased the adoption of this powerful modeling technique.

For equipment designers, CFD modeling is now being used to streamline the equipment-design process, in order to reduce the design cycle, speed the time-to-market for the product, and trim manufacturing expenses by reducing the need for extensive prototype development and physical testing.
Meanwhile, for process operators, CFD modeling is being used to fine-tune operating parameters in order to troubleshoot and optimize operations, reduce emissions, address vibration problems, increase throughput rates and product yields, and more.

How CFD works

The software provides visual representations that make it easy for the user to analyze the behavior of virtually any flow field under real-world conditions, and to make predictions related to a range of operating phenomena such as reaction rates, mixing, flow separations, materials distribution, temperature or pressure distribution, and shear-rate distribution.

Users continue to benefit from ongoing advances in the upfront mesh-generation process, which is an integral part of the CFD modeling process, and the post-processing software that is used to fine-tune the analysis and visual display of the modeled results.

Many of today’s software programs also provide automatic error-checking functionality, which can spot inconsistencies during data input and model setup, and alert users before they move on to the next step. Similarly, “geometry healing” features help many of today’s CFD solvers to automatically reconcile “non-optimal” geometry characteristics, such as small gaps, non-matching edges and very small surfaces, in a way that improves both the consistency of the underlying geometry, and the accuracy of the modeled results.

Orders-of-magnitude improvements in the speed and memory capabilities of today’s computers have opened the door for larger and more complex equipment components and chemical process systems (which are typically characterized by inherently unsteady flow fields and involve large, unsteady datasets) to be reliably modeled using standard desktop and even laptop computers — not the supercomputers that were once required for CFD modeling.

Mesh generation

Prior to carrying out any CFD analysis, a virtual prototype of the modeled component must be created. To simplify the complex calculations that are carried out during CFD modeling, the geometry of the system or device to be modeled must first be divided into a mesh of tiny cells. When taken together, these discrete cells approximate the intricate geometry of the component being modelled. Then, the underlying equations are solved at each cell.

Meanwhile, to speed and automate the meshing process, most of today’s mesh-generation software programs are now able to accept computer-aided design (CAD) drawings of the component from virtually any commercial CAD program, and to use this as the starting input for the mesh-generation process. Once the user specifies some basic cell sizes and gradings (with guidance from the program), the software then does the bulk of the work to automatically create and smooth the mesh.

The number of mesh cells required to accurately convey the complex geometry that is being modeled can number from the hundreds of thousands of cells to millions of cells. The finer the mesh, the more accurate the calculations — although finer meshes traditionally have a larger number of cells, and thus, convergence of the modeled results take longer, and involves more data.

Today, automatic meshing capabilities are included in most commercial CFD modeling programs, and these continue to evolve. For instance, contemporary meshing programs now use a variety of predefined building blocks and standard mesh-style templates based on a variety of geometric elements, such as hexahedral, tetrahedral and polyhedral elements, pyramids, wedges, prisms (or any combination of these).

In recent years, the availability of so-called polyhedral meshes has improved mesh generation. Compared to the conventional tetrahedral meshes, the newer polyhedral meshes have fewer cells, so convergence during modeling is faster, thereby accelerating the simulation. These advanced meshing techniques also allow the geometry of virtually any complex component to be accurately represented, even those with complex features such as fuel nozzles and intricate holes and passageways. And, thanks to the availability of today’s user-friendly interfaces, users are guided through the mesh-generation process.

One CFD vendor, Ansys (Canonsburg, Pa.; www.ansys.com), has reported that in one recent application of its Fluent 6.3 CFD program, the mesh size requirements for a particular project were reduced from 800,000 cells to just 150,000 cells by using the latest meshing capabilities, and this allowed CFD simulations that used to take many days or even weeks to solve to now be carried out in a matter of hours.

Post-processing software, to manage the modeled results

Today, sophisticated post-processing software (including both software that is embedded in many off-the-shelf CFD programs, and software that is available as standalone programs) capabilities are helping investigators to more effectively interrogate large, unsteady datasets, and to visualize critical aspects of complex simulations more meaningfully. Post-processing tools help the user to produce high-resolution color graphics and animations that represent such things as velocity vectors, contours of pressures, and lines of constant flow-field properties.

During CFD analyses, users are usually interested in working with a much smaller volume of targeted data, related to one particular aspect of the model. For instance, if the subject of a particular CFD study is surface-pressure behavior, then only the surface data are needed for analysis during post-processing. Typical data extracts include graphical elements such as iso-surfaces, cutting planes, computational surfaces, and/or particle paths, or the data can be integrated along specific lines and surfaces to obtain relevant length-, area- and mass-weighted averages.

Partnering CFD with other modeling techniques

These days, CFD modeling is increasingly being coupled with other types of modeling software, to integrate fluid-flow modeling with other types of physics-based modeling and simulation. For instance, when CFD models are linked with other stress-analysis models in a fully coupled fashion, users are able to carry out more complex fluid-structure interaction (FSI) analyses.

Using FSI analysis, users can perform detailed thermal and stress analyses of the solid components, not just the flowing streams, and model, for instance, how the flow of high-temperature gases impacts the structural components. When such insight is available, users can then implement strategic design or operational changes to improve overall equipment performance and minimize the risk of problems related to such things as erosion, vibration and fatigue. For instance, users are able to model not just the fluid flow inside a mixer, but the potential structural deformation that may result on mixer components, turbine blades and so on, from the flow of fluids and entrained particles in the system and other aerodynamic loads.

In recent years, the engineering community has also been working to integrate CFD modeling with a newer modeling technique called discrete element modeling (DEM). DEM is used to simulate the behavior of granular solids in a wide variety of flow situations, and it does this by modeling the actual behavior of a statistically relevant number of particles in a solid-liquid mixture.

As with CFD, DEM capabilities have also improved significantly in recent years, thanks to ongoing software advances and the availability of faster, more-powerful computing capabilities. For instance, recent advances in DEM modeling programs now allow the behavior of more than a million particles in a system to be modeled simultaneously.

Industry examples

Mixers:

Engineers throughout the chemical process industries are routinely challenged by their efforts to select the best mixer geometry and mixing conditions to ensure the desired results. In mixing applications, CFD is helping both designers and users to get a better idea of flow-distribution patterns and energy-dissipation rates inside the vessel. During industrial mixing, the ability to minimize the formation of stagnant zones and hot spots helps to not only optimize blending times and mixing conditions, and ensure desired product uniformity, but to reduce wear and tear on mixing equipment, as well.
In recent years, it has become standard practice in the mixing community to use CFD to model mixers across the entire spectrum, including impeller-based mixers that are used for relatively low-shear mixing applications, more complex rotor-stator mixers that are used for high-shear mixing (to produce emulsions and dispersions and to carry out wet milling or particle size reduction), and static mixers, as well.

Gas turbines and combustion systems:

CFD has also proven to be a useful tool for modeling all types of burners and combustion systems, particularly those used in turbines to produce power and generate electricity. For instance, when CFD is used to characterize fluid flow, combustion-related reactions and temperature distributions at any given cross-section, users are able to evaluate the potential impact of competing burner designs and arrays inside the combustor, and the impact of competing operating scenarios.

Such insight lets them make the changes needed to optimize inlet air temperatures, maximize combustion efficiency, minimize the overheating of turbine blades, vibration and fatigue-related failures, and improve overall turbine performance and efficiency.

Here again, users of CFD have been able to identify relatively minor adjustments to the design or operating parameters of a gas turbine and see positive results, in the form of increased power output, reduced emissions and vibration, and reduced unscheduled downtime and manufacturing costs.
Most often, the cost of implementing simulation software is justified based on the direct payback that such modeling provides, in terms of reductions in expenses and improvements in operational efficiency. With engineering simulation becoming so widespread, the big question now for equipment designers and process operators is not whether to use the technology, but how best to take advantage of the many opportunities it provides.

Team work the criterion for successful CFD applications

The development and application of CFD in chemical process technology calls for team work between specialists from theory and practice – that means physicists, IT specialists, process engineers, plant engineers and software developers. This aim also constitutes a key objective of the interdisciplinary work of the ProcessNet Technical Committee „Computational Fluid Dynamics“ (www.processnet.org/FDTT), an initiative of GVC (VDI Gesellschaft Verfahrenstechnik und Chemieingenieurwesen) and DECHEMA (Gesellschaft für Chemische Technik und Biotechnologie, Frankfurt am Main, Germany).

The work of the Computational Fluid Dynamics Technical Committee focuses primarily on modeling transport phenomena in chemical processes. To this end it provides a forum for the discussion of research and development findings, the exchange of insights and experience, for continuing education and the advancement of next-generation scientists. This interdisciplinary work generates many interfaces to other ProcessNet Technical Committees, such as Multiphase Flow, Heat and Mass Transport, Mixing, Rheology, High Temperature Technology, Gas Purification.

The wide range of current themes from practice, which the Congress Programme of ACHEMA 2009 will also highlight, covers hybrid separation processes, separation and flow dynamics of complex mixtures, advanced fluids, separation of biological products and even micro-scale separation technology, and, last but not least, processes and methods that make a substantial contribution to product and process optimization and thus to competitiveness.