Christof Knüsel, Dürr Systems GmbH, Paint Systems Automotive, Stuttgart, Germany

A a full-range systems supplier, Dürr Systems GmbH offers turnkey paint shops for mass production paint finishing. The complete package contains buildings, plant and environmental engineering, conveyor equipment and control, automation, and material handling techniques. Dürr also offers a complementary range of manufacturing support services for all aspects of the paint finishing process. One important component of their work involves CFD simulations. Since 1998, Dürr has used FLUENT to model such things as air flow in spray booths and work stations, air flow and heat transfer in ovens, mist elimination in scrubbers, response to electric fields during cathode dip painting, and fluid flow in dip tanks.

Pretreatment is the first of many stages in the painting process. Here the automotive body is cleaned and prepared for subsequent coating processes using methods appropriate for the material involved (steel, aluminium, magnesium, etc). One phase of the pre-treatment process, called phosphating, is used to apply a zinc phosphate base coat. The process is normally carried out in dip tanks, where the flow is driven by 100 to 300 injection nozzles. This coat protects the body from corrosion and acts as a bonding base. A secondary reaction produces iron phosphate, which takes the form of sludge and is removed from the dip tank continuously.

When the process is applied to aluminium sections, it triggers a further secondary reaction, which produces cryolite. To counteract any reduction in surface quality arising from the presence of cryolite, the flow velocity should always be above 0.3-0.5 m/sec near the aluminium components. Since the current trend is toward bodies with more aluminium parts, phosphate tanks in existing plants often need to be upgraded for new car models in order to increase the flow velocity at critical points.

CFD simulation is an excellent tool for optimizing the flow in a phosphate tank. First, a simulation of one injection nozzle is carried out using a fine grid of approximately 150,000 cells. The results are used to generate velocity and turbulence profiles that are characteristic of the nozzle. Second, a simulation of the complete tank is performed using a larger (about 2 million cells), yet comparatively coarser grid. The velocity and turbulence profiles predicted in the first simulation are used as boundary conditions for the injection nozzles in the second. The profiles are modified slightly to ensure that the jet characteristics on the coarser grid are nearly identical to those on the fine grid of the first simulation. The second round of calculations usually requires several days to obtain a suitably converged solution, using a 1.0 GHz processor. Experiments using tanks filled with water show good agreement with the simulation.

CFD has enabled Dürr to develop new dip tanks with optimized flow conditions and offer customers individual solutions for optimizing existing tanks to suit new car models. The 3D simulation plots make it easy for customers to understand where the problem areas lie, and where modifications should be made to obtain a better surface quality. A decline in surface quality can result from poor flow in dip tanks, and can add expense to customers’ operating costs. In many cases it can be avoided with CFD.

Flow in a dip tank: the side-flooding system is illustrated by path lines

FluentNEWS