An optimization method for the design of extrusion wire coating dies

Author(s): 
N. Lebaal (Universite´ de Technologie de Belfort-Montbe´ liard), S. Puissant (Brugg Cables Industry AG), F.M. Schmidt (Universite´ de Toulouse), D. Schläfli (Maillefer Extrusion SA)
The coat-hanger melt distributor is commonly used in the wire coating process. Its task is to distribute the melt around the conductor uniformly. Balancing the distribution of flow through a die to achieve a uniform velocity distribution across the die exit is one of the most difficult tasks of extrusion die design. For the polymer extrusion industry, the most challenging and demanding work is to explore how to reduce or even
eliminate die correction.

 
 
 
 
 
 
 
 
 
 
 
The objective of this article is to determine a wire coating-hanger melt distributor geometry to ensure a homogenous exit velocity distribution that will best accommodate a wide material range and multiple operating conditions (i.e., die wall temperature and flow rate change). modeFRONTIER is coupled with a FE simulation to ensure the dimensional precision of the final products. For this purpose the velocity relative difference and the swell ratio are taken as objectives functions. The computational approach thus incorporates FE analysis to evaluate the performance of a die design and includes a nonlinear constrained optimization algorithm based on the Kriging interpolation and sequential quadratic programming algorithm to update the die geometry. Two optimization problems are then solved, and the best solution is taken into account to manufacture the optimal distributor. The results show an improvement of the velocity distribution for two polymers with different rheological behaviours. The Taguchi method is then used to investigate the effect of the operating conditions, i.e., melt and die wall temperature, flow rate and material change, on the velocity distribution for the optimal die. In the example chosen, the wire coating die geometry is optimized by taking into account the geometrical limitations imposed by the tool geometry. The optimal die was manufactured and an experimental comparison allowed the validation of the overall simulation procedure and the robustness of the optimization strategy. The experimental measurements and the flow dynamic results were in excellent agreement with the experimental data.