The next performance limit is thermal
A system with 95 percent efficiency delivers 100 kW of useful power. This results in approximately 5.3 kW of power loss, which must be dissipated as heat. If the power output increases to 120 kW while maintaining the same efficiency, the power loss rises to about 6.3 kW—often within the same or even a smaller installation space. What sounds like technological progress thus quickly becomes a physical challenge: Where to dissipate the heat?
Whether it’s an AI processing unit, a vehicle inverter, a fast-charging battery, or a satellite: the applications could hardly be more different. Yet developers are grappling with variations of the same problem. Greater computing power, shorter charging times, more compact electronics, and increasing power densities mean that more and more heat must be managed within a limited space. In high-performance electronics, local heat flux densities can already reach several hundred watts per square centimeter. The next performance limit of modern technology could therefore increasingly be determined by thermal considerations.
When Cooling Becomes an Architectural Issue
“If cooling isn’t considered until the end of the development process, the actual problem has often already been built in,” says Dipl.-Ing. (TU) Stefan Merkle, managing partner of Merkle CAE Solutions GmbH. “With high power densities, thermal management becomes an architectural issue. Installation space, materials, flow patterns, contact surfaces, and operating strategies must all be considered together.”
This is because a maximum permissible temperature alone says little about a system’s thermal quality. In power electronics, an unfavorable interface can cause a local hotspot. In a battery system, significant temperature differences can arise between cells despite an acceptable maximum temperature. In high-precision structures, temperature gradients can, in turn, cause thermal expansion and thus deformations that are critical to function.
A thermal problem becomes a multiphysics problem
As a result, it is increasingly insufficient to simply calculate how hot a component becomes. What matters most is how heat is generated, transported through materials and contact surfaces, absorbed by cooling media, and dissipated into the environment. Equally important is the question of what effects the temperature subsequently has on other physical quantities.
This is precisely where coupled simulations come into play. CFD, for example, models cooling flows and heat transfer; thermal models show temperature distributions; and FEM calculates the resulting stresses and deformations. Parametric studies also make it possible to virtually modify cooling channels, flow rates, materials, or component arrangements before the design is finalized.
No More Performance at Any Cost
The benefit, therefore, lies not only in ensuring a viable cooling solution. Simulation can reveal, even in early development phases, where the thermal limits of an entire system lie and what design changes can push those limits further.
This also changes the role of thermal management. It no longer necessarily follows the development of more powerful components but becomes, in itself, a prerequisite for further performance gains.
For developers, the crucial question in the future may therefore no longer be simply, “How much power can we generate?” but rather, “How much of that can we thermally manage?”
