|
Abstract:
|
Automotive wiring harnesses have gained weight and
complexity through the last decades due to the increasing number
of electrical components, which has raised the interest on its
weight optimization. For this purpose, it is essential to know
at least the maximum amount of steady current that either
single wires or bundles can carry. However, the large amount
of combinations of the customer-speci¿c wire harnesses makes
it impossible to exhaustively simulate all of the combinations
that would allow for a reliable analysis and optimization of the
network. The proposed approach consists of achieving accurate
predictions of the wire thermal behaviour using fast on-line
polynomial functions, which have been created as regression
models using data from off-line worst-case ¿nite element sim-
ulations. These regression models provide good accuracy for
the critical dimensions of wire bundles in a much shorter time
than simulations, so that they can be used on-line in optimiza-
tion algorithms. Two different approaches of optimization are
presented here in order to assign discrete values of available
wire cross-section to the wire bundles: The ¿rst one uses integer
linear programming, and the second one consists of a recently
created custom algorithm whose objective is to reduce the
computation time of the integer linear programming approach.
This latter objective is satisfactorily accomplished. Results of both
optimization approaches are validated by means of ¿nal ¿nite
element simulations, and they promisingly ful¿ll the objectives
of this study. |
