Rising demand for computational power necessitates implementation of efficient liquid cooling systems in servers. Liquid cooling, however, is leak-prone and can cause catastrophic damage.
Develop a quick, reliable, and durable system to detect leaks within server blades by triggering color-temperature changes from stimulating the flourscence of the coolent.
For my Engineering 101 project at Duke, I was selected to join the team working with NVIDIA. After meeting the engineers and learning more about the work at the Morrisville facility, we entered the lights-out test data-center that they had in the facility. Walking around, we noticed absorbant bags on the floor at the base of each server rack. After inquiring more, we learned that the current solution for leakages within the server was a late-stage prevention after the liquid had already traveled through the server blade and began to pool up on the ground. With damages to a single server blade being up to $40,000, and multiple of them at risk of being damaged on a rack and surrounding racks before detected, we knew there was a need for a system to detect leakages before they spread from the origin space.