⚡ News

AI Growth Elevates Data Center Fire Risks: Research Pinpoints Causes and Solutions

AI Growth Elevates Data Center Fire Risks: Research Pinpoints Causes and Solutions

Over the past decade, artificial intelligence (AI)'s capacity to accelerate complex processes has cemented its role as a pivotal tool in engineering research. While AI is often associated with cloud computing, its expansion necessitates a growing infrastructure, prominently featuring data centers.

Researchers from Texas A&M University, led by Dr. Qingsheng Wang and chemical engineering Ph.D. students Tylee Kareck and Chi-Yang Li, are investigating an unexpected but critical threat to these data centers: elevated fire risk. In a recent collaborative publication with George Washington University and the University of California, Berkeley, the team meticulously analyzed common causes of data center fires and identified strategies to mitigate these hazards. Their findings have been published in the *Journal of Loss Prevention in the Process Industries*.

“Our work provides critical insights to assess fire risk, enabling engineers to design more secure and resilient data centers,” stated Wang, a professor in the Artie McFerrin Department of Chemical Engineering.

As data centers proliferate and become more powerful, they demand increasing energy from battery systems and backup generators to ensure uninterrupted operation—a fundamental requirement for data storage. This escalation in energy utilization directly correlates with an increased fire risk.

Kareck highlighted, “Modern data centers house enormous quantities of electrical energy, meaning failures can escalate rapidly if not adequately controlled.”

The research identified multiple ignition sources for data center fires, including battery failures, electrical faults such as arc flashes, equipment malfunctions, and human error. A significant concern is the presence of large quantities of batteries and high power density, which heightens the risk of thermal runaway—an uncontrolled chemical reaction within a battery that generates substantial heat. This phenomenon can lead to battery explosions, ignition, and potentially cascade into neighboring batteries.

Notably, researchers were surprised to discover that the causes of data center fires frequently overlap. For instance, human error during battery installation can precipitate arc flashes, which then ignite other components and escalate into a fire. However, arc flashes can also occur independently due to short circuits, even without human error.

↗ Read original source