Air-cooling optimization at system level

A computational model and a method is presented for the optimization of air cooling for power electronics in terms of thermal performance, cooling-system size and acoustic noise, which are key parameters for power-electronic applications. A number of case studies is performed, showing the relationship and the trade-offs between these quantities. In general increasing thermal performance leads to bigger size and/or higher noise, while lowering noise for constant thermal performance increases size. Within the basic physical limits, optimization is possible, which must be carried out at system level, as the results cannot be found by separately optimizing individual cooling-system components. Examples are the enlargement of the heat sink or the addition of volume for porous duct lining, which both can lead to a smaller overall system for fixed thermal performance and noise. Furthermore it can be said that any measure to reduce pressure drop, such as the introduction of guiding vanes in a bend, reduces size and/or noise at fixed thermal performance. The results are shown to be in accordance with theoretical scaling laws, relating thermal resistance, size and sound pressure. The trade-off between thermal performance, size and noise worked out for air-cooling systems is expected to be of general nature, applying to any kind of cooling system. In future work, the present method could be used to compare air cooling with other types of cooling systems, in particular water cooling and two-phase cooling.