For all the advances in risk management and materials science in modern history, fire remains the leading cause of property loss at commercial and industrial facilities worldwide. Even more frustrating to owners and insurers than the cost of fires and collateral smoke and water damage is the expense of fire protection equipment that is often ineffective.
The challenge there is certainly not lack of experience, but the complications of empirical testing. Sophisticated burn rooms exist, but test fires require time-consuming preparation and forensics. Computer models are less expensive to run, but have not yet proven high fidelity.
Sergey Dorofeev, fire hazard and protection research area director at FM Global, has developed a computational fluid dynamics (CFD) computer model for fire suppression. It is designed to help develop fire-suppression systems that are closely tailored to specific facilities, as well as address hazards and scenarios that cannot be directly tested due to size or structural and safety considerations. It also provides quicker turnaround for clients and for FM Global's own property loss prevention data records. The new CFD model is expected to save millions of dollars in research costs. According to the company, live-fire tests at its research campus in Rhode Island can cost as much as $100,000 each, and the results are difficult to extrapolate to untested scenarios.
"Sergey's model is very different," said Louis Gritzo, FM Global's vice president of research. "It is a physically based model, arising out of natural laws for fire growth. That is very different from the standard Cat model that uses statistical analysis and stochastic calculations. The numerical and physical simulation of fire growth is a breakthrough."
Gritzo explained that up to this point, important data and knowledge has come either from actual losses, or from large-scale live-fire tests. He added that Dorofeev's CFD model will not replace full-scale testing but it will provide more data, more quickly and less expensively. In particular, the new approach incorporates fluid and heat-flow dynamics for both fire and water. While it is true that there are chaotic elements to every fire, much of fire growth and suppression is governed by physical laws.
"Sergey is a leader in this field because he has a strong background in both computational modeling and also empirical testing," said Gritzo. That means, he elaborated, that Dorofeev's model incorporates not just the physical model for fire and water, but also addresses the effectiveness of fire suppression systems. The idea is not just to knock down an inchoate fire, but also to develop fire-protection systems that minimize expense to property owners, operational downtime, supply chain interruption and actual losses for insureds.
That is proving to be an elusive goal, Gritzo said. "There are several challenges in fire protection today. For one, buildings are getting bigger, and ceilings are getting higher."
There are also more mundane and enduring challenges.
"Everyone hates sprinklers. They are often in the way, and they often get damaged," Gritzo added. Damaged sprinklers often mean collateral water damage. In a fire, sprinklers tend to be a blunt instrument, usually fairly effective at stopping the blaze, but even more effective at soaking an entire area.
"Through this project we are trying to pare down sprinklers," said Gritzo, "make them less intrusive and less expensive, but also more effective. Ultimately I think Sergey's model will lead to better standards for risk managers and for the entire industry, and better ways to get protection from fire."
Dorofeev's work on the CFD model started in 2008 and, like the live-fire testing it is meant to complement, the development has been challenging, time-consuming and expensive.
It has had to incorporate changes on the fly as insureds make their own changes in manufacturing processes, materials and storage and warehousing methods.
The development took three phases. The first called for an assessment of existing models. Focused experiments assessed material flammability properties, studies of sprinkler spray parameters and fire extinguishment tests.
Phase two began the collaboration between FM Global and research partners at the University of Maryland and University of Edinburgh, both leading academic institutions in the fire science field.
Advanced models for fire growth and validation protocols were developed that addressed turbulent combustion, finite-volume radiation heat transfer, soot radiation, material flammability, and fire extinguishment.
FM Global organized a first-ever open-source CFD fire-modeling workshop, held at its Center for Property Risk Solutions in Norwood, Mass. in 2009. It has since become an annual event.
Phase three began in 2011, using advanced models for protection system activation and fire extinguishment that were developed in the earlier phases. Updates and modifications were introduced, improved and validated. And that process continues through commercial operations.
--By Gregory DL Morris
September 15, 2013
Copyright 2013© LRP Publications