It’s no secret that carmakers are under a lot of pressure to reduce the number, size, and weight of engine components for better fuel economy. In one case, we were tasked with designing a smaller and more-efficient heat sink to thermally regulate the electronic components in Toyota hybrid vehicles. Instead of using typical analytical design methods and trial-and-error physical prototyping, we first used multiphysics software to design and test possible prototypes.
Hot under the hood
Toyota hybrid vehicles have sophisticated electrical systems in which many power diodes and power semiconductors — such as insulated- gate bipolar transistors — handle power conversion and other applications. These components are standard planar silicon devices measuring a few centimeters on each side. The devices mount on aluminum heat sinks, or cold plates, with channels in which a water-and-glycol-coolant mixture flow.
In earlier model years, the cold plate featured a fluid inlet on one side and an outlet on the other side. The long channels in between were mostly straight. They provided adequate heat transfer but at the cost of a significant pressure drop across the plate. We needed a much smaller and more-efficient design, but this meant that thermal management would be more difficult.
It might seem reasonable to simply redesign the cold plates so more coolant could flow through them. But that would require more pumping power. The limited space in the engine compartment meant that using a larger, more-powerful pump or an additional pump was unacceptable. Therefore, we had to reengineer the cold plate to simultaneously get optimum heat transfer and negligible additional pressure drop.
Jets not enough
Interestingly, many researchers have identified jet impingement as an attractive way to cool surfaces. But while jet impingement performs well with respect to heat dissipation close to the jet, it doesn’t work well further away from the orifice. That’s because most heat transfer happens close to the jet entrance where the fluid is the coolest and flows fast. Much of the heat-transfer capability is lost by the time the coolant exits. One answer was to combine jet impingement with a peripheral channel to boost the area-average heat transfer.
Comsol Multiphysics CFD and Heat Transfer Modules were essential to the numerical simulations at the heart of our work. And Comsol’s LiveLink for Matlab also let us work with the multiphysics simulations in a high-level scripting language to automate the process of improving the cold plate’s topology.
We examined how the topology influenced variables such as steady-state convection-diffusion heat transfer and channel fluid flow. We used well-established materialinterpolation techniques and a special optimizer, iterating between Comsol and Matlab.
The aspect ratio of the channels (i.e., ratio of height to width) was important, but to simplify the numerical simulations we assumed a thin 3D structure and then further “flattened” it. Once we derived an initial channel topology, we investigated the height of the fins that separate the cooling channels and performed a separate parametric sizing study. We had already done similar studies, so our assumptions were well informed.
Our numeric simulations generated an optimal cooling channel topology with fluid streamlines in branching channels. Because the channels efficiently distributed coolant throughout the plate, we used the fractal-like topology to guide the design of a cold-plate prototype. We set the size of the plate to 60 × 45 mm, with a middle cooling zone of 25 × 15 mm to match a specific heat source. We assumed the plate’s base substrate thickness to be 1 mm.
After using Comsol and Matlab to improve optimize the channel topology, we then used the final channel concept to design and evaluate a prototype using Comsol’s LiveLink for SolidWorks. LiveLink lets users actively link to CAD, and it was easy to import various designs from SolidWorks back into Comsol to verify pressure drop and heat transfer. This capability lets users quickly establish a reasonable starting point and progress from there quickly.
Last, we fabricated two physical prototypes based on the SolidWorks designs using standard micromachining techniques. One prototype was a dual jet/hierarchical microchannel version and the other used jet impingement of a simple flat plate.
We tested the prototypes to see whether the dual configuration would indeed provide higher-performance cooling in an ultracompact package. On average, it did dissipate 12.8% more heat than the flat-plater version. With regard to pressure drop, both cold plates demonstrated similar results, but the dual version performed slightly better at higher flow rates.