Optimization of Heat Sinks using a 3-Level DOE Methodology 2006-01-1453
This paper evaluates the interaction between three design variables of a conventional extruded heat sink used for automotive audio systems in a passive convection environment.
A 3-level Design of Experimentation (DOE) methodology, with a center point design, is utilized to quantify the non-linear behavior and the interaction effects between three design variables of the heat sink. A full factorial analysis of 27 CAE runs is employed to begin the DOE analysis.
A second order non-linear response function is developed, using an estimated regression coefficient, that's used to optimize the heat sink design.
The second-order quadratic model that includes the 2-way linear interactions between design variables is proven to be the model of choice that can characterizes the heat sink temperature more effectively and accurately.
This model is simplified, to include the most significant effects and their associated interactions, and presented in a coded and non-coded format with a 95% confidence level.