| |
Christof Knüsel, Dürr Systems GmbH, Germany
View the pdf of this article
DÜRR SYSTEMS GmbH designs and builds turnkey paint shops for painting cars and high-quality components used in automobile mass production. Process, application, conveyor, and control systems are combined to form integrated total solutions. Since 1998, Dürr has used FLUENT to model such things as air flow in spray booths, work stations, and ovens, mist elimination in scrubbers, electric fields during electro-coating, and fluid flow in dip tanks.
Dürr’s convective oven with air nozzles arranged along the sides
A CAD model of the car door (left) and CFD simulation model showing
surface temperature (right)
Simulated and measured heat-up curves for the car door
On the process line, ovens are placed after electro-coat, base coat, and clear coat spray booths. While convective ovens dominate in Europe, radiation zones are also common in the USA. In convective ovens, air at a temperature of 140° to 180° C is blown onto and inside the car body via 70 to 100 mm nozzles, which are placed at the side walls of the oven. The air velocity at the nozzles is typically in the range of 15 to 30 m/s. A typical curing process takes about 30 minutes, the first 15 minutes of which is used to heat up the car body or component. At the end of this 15 minute period, the temperature of oven with air nozzles arranged along the sides all parts of the body – from thin plates on the hood to thick plates on the rocker panel – must be within a defined temperature range, with a tolerance of about +5° C. It is therefore very important for Dürr to estimate the heat-up curves of an oven before it is built.
Toward this end, Dürr uses CFD to simulate the movement of the car body through the oven. The sliding mesh model is used and the turbulent flow with heat transfer includes both radiation and conduction. At this point, it is not possible for Dürr to simulate the heat up of complete car bodies in detail, since numerical resources are limited, especially considering the time required to accurately model the car body geometry. Instead, only critical components are simulated in detail, while a coarser model is used for the rest of the car body. In the detailed part, all metal sheets are modeled. For each sheet thickness and connection type, separate boundary zones are used with specific material properties. For each junction type, the heat conduction is extracted from a database, which is continuously upgraded by experiments.
After validating the CFD model using a simplified car shape, an actual car door was simulated as a single part – without the rest of the car body – and corresponding tests were made in the testing plant. The door had sheets with different thickness, a reinforcing sheet was placed inside the door, and there were several gaps and holes in the geometry. Experimental measurements of the heat-up behavior of the car door were found to be in very good agreement with the CFD predictions. Pilot simulations in automobiles with a main chassis beam and an A-pillar were also successful.
It is now possible for Dürr to simulate heat-up curves for different regions of a car body using CFD. Current efforts are aimed at reducing calculation time by taking into account the (approximate) periodicity of the flow and changing the updating interval of the radiation equations. These changes, along with enhancements to the database with total heat conductivity for different junction types, will allow Dürr to be able to offer oven simulations as a special service for car makers in late 2005.
|
|
|
