Computational fluid dynamics

An increasing general interest in computational fluid dynamics (CFD) in relation to finned, air-cooled heat exchangers is an indication that CFD will become an indispensable tool for developing new units in the foreseeable future. Dr. Andres Zürner of Guentner, explains.


ost of the work with CFD at present is still coming from the university environment – generally projects running over several months or even years, which look at a specific application in detail in order to answer fundamental questions comprehensively. Rapid progress in the area of chip manufacturing, however, is now also allowing CFD to be used efficiently for simpler tasks with little expense and effort. Numerical flow simulations of consistently high quality can, meanwhile, even be performed with normal desktop PCs within an acceptable time period. If larger projects with an increased need for computing capacity ever have to be processed, then cloud computing offers the possibility of cost-effective leasing of the required resources on the web. Numerical flow simulations have already been used

successfully for many years now at Güntner to handle the most varied tasks. The article to hand briefly outlines the method involved and uses a few examples to illustrate the extent to which insights and findings have been gained as a result, which would be scarcely feasible using standard measurement methods, or only with a great degree of difficulty.

What is CFD? In simple terms, computational fluid dynamics sets out to describe fluid motions with the aid of computers. The concept of fluid dynamics is based largely on the Navier- Stokes equations and their variations. In general terms, only a few very specific problems can be resolved analytically with these partial differential equations. This means that to solve the actual problems that confront us on an everyday basis, iterative approaches have to be used almost without exception. By nature, however, these can then ‘only’ deliver approximate solutions. The deviations in this respect should be understood relative to an analytically precise solution and are still smaller in general by orders of magnitude than would

26 July 2018

normally be interpreted as a deviation. Because iterative approaches generally involve significant computational effort, computers are usually used for these tasks.

Basic approach

A corresponding (generally 3D) drawing is first created for a given problem using a CAD program. In our example, consisting of cold room, heat exchanger, casing and fan, it is split into a number of small, individual cells. This step, which is termed ‘meshing’, determines the points at which the calculation software calculates the individual simulation variables like pressure, velocity, and temperature. Meshing therefore contributes significantly to the precision of the simulation results and has to be matched to the problem at hand – a task that is still not fully automated and therefore demands a certain level of experience and basic expertise in CFD on the part of the user. Following on from the meshing step, a range of additional parameters that are important for simulation are also specified, such as definition of the interfaces, selection of solvers and the turbulence model, but also substance-specific fluid properties. The points just mentioned are generally summarised under the term ‘preprocessing’. We use our own internal CFD workstation for this purpose. The actual simulation – the iterative solving of the equation systems – is performed on the basis of cloud computing using servers leased on the web. Thanks to this extremely flexible solution, we are virtually unrestricted in terms of computing and memory capacity and reduce running costs to a minimum. Once the simulation is complete, the results are transferred back onto the company server, where they are evaluated accordingly.

Existing measurement data is used where possible as a basis for validating the simulation results. The convergence of specific parameters (generally pressure or temperature), the independence of the results from the level of discretisation,

Page 1  |  Page 2  |  Page 3  |  Page 4  |  Page 5  |  Page 6  |  Page 7  |  Page 8  |  Page 9  |  Page 10  |  Page 11  |  Page 12  |  Page 13  |  Page 14  |  Page 15  |  Page 16  |  Page 17  |  Page 18  |  Page 19  |  Page 20  |  Page 21  |  Page 22  |  Page 23  |  Page 24  |  Page 25  |  Page 26  |  Page 27  |  Page 28  |  Page 29  |  Page 30  |  Page 31  |  Page 32  |  Page 33  |  Page 34  |  Page 35  |  Page 36  |  Page 37  |  Page 38  |  Page 39  |  Page 40  |  Page 41  |  Page 42  |  Page 43  |  Page 44  |  Page 45  |  Page 46  |  Page 47  |  Page 48  |  Page 49  |  Page 50  |  Page 51  |  Page 52  |  Page 53  |  Page 54  |  Page 55  |  Page 56  |  Page 57  |  Page 58  |  Page 59  |  Page 60  |  Page 61  |  Page 62  |  Page 63  |  Page 64