- February 26th, 2019
- Category: Article
AI is teaching us that generative design is key to optimal cooling
Artificial intelligence is changing the way engineers design thermal cooling solutions. But how? Imagine what would happen if you explained the laws of thermodynamics to a hyperintelligent machine. With the capacity to think, imagine how that machine would optimally design cooling fins and cooling channels to precisely fit your needs. Would it decide to use parallel cooling fins, S-shaped cooling channels, or something new? Would it try to maximize the contact surface area, or not?
Through artificial intelligence, the thermal engineers at Diabatix are discovering that parallel cooling fins and S-shaped cooling channels aren’t the only ways to efficiently dissipate heat. In fact, artificial intelligence has taught us that each heat sink deserves a generative design topology to reach the optimal cooling of a given technology. Wow - AI’s ability to conduct thousands of topology simulations in just a few days, and then uniquely adapt to the required material, shape, pressure drop, volume, and uniformity of a heat sink is changing the landscape of thermal design right before our eyes. There is no such thing as ‘cookie cutter’ cooling solution with AI – everything is customized to maximize potential.
For example, in the graphic below, you can visualize how the optimal topology of each heat sink changes based on the conductivity of the material used. This is AI at work.
AI is showing us that to achieve more powerful and efficient cooling, one crucial key is in the generative design process. With the incredible strides made today, whether it be in EV batteries and motors, lasers and LEDs, or enhanced medical imaging equipment, each technology deserves a unique and optimal cooling solution to operate at its best. Artificial intelligence is allowing thermal engineers around the world to keep up with these continual innovations in tech.
In the following video, you can visualize an air cooled heat sink, and the customized design process that our AI system goes through in coming up with the most optimal topology within the given constraints.