SMART Sprinkler for Highly Challenging Fires


View the PDF here

SMART Sprinkler for Highly Challenging Fires

By: Yibing Xin, FM Global, Research Division, Norwood, MA 02062, USA


In the world of modern distribution, larger warehouse structures now combine high vertical storage and enormous enclosed volumes for the sake of efficiency-lifting and retrieval technologies, making it possible to stack products higher than ever. To minimize fire risks, some businesses have installed in-rack automatic sprinklers (IRAS): sprinkler systems within the rack configurations often used to store products on pallets or in containers. Compared to ceiling-only protection, the IRAS are often more expensive and inconvenient for warehouse operation. Moreover, in certain situations such as portable racks and forest product storage, combustible commodities could be stacked up to 80 ft (24 m) high, but there is no infrastructure to support the IRAS. In those storage scenarios, traditional ceiling-only sprinklers encounter tremendous challenges. Why? Traditional sprinklers descended from innovations of the 19th century rely on a sensible element within the sprinkler to activate water flow. If a fire begins at or near the base of a roll paper stack, the temperature at the ceiling may not be hot enough to activate until the fire has grown substantially and perhaps spread far beyond its origin. When activated, sprinklers farther away from the fire origin sometimes activate earlier than those closer (i.e. sprinkler skipping). Sprinkler skipping, caused by water droplet impingement, is detrimental to the effectiveness of protection. For the activated sprinklers, downward water transport is often impeded by a large or rapidly growing fire. However, efforts to make traditional sprinklers more sensitive, such as lower temperature ratings, could backfire due to erroneous activation in regions with very warm climates.

Figure 1. An example scenario of a highly challenging fire.

Highly Challenging Fires

Highly Challenging Fires (HCFs) are fire hazards that cannot be protected under existing protection recommendations. The combination of high storage and potential fire spread is a typical example of HCFs that demand protection beyond the guidelines for use of traditional sprinklers. FM Global Research explored the notion that the greatest strength of traditional sprinklers is their essential problem: the shared detection and actuation mechanisms. These mechanisms have yielded remarkably reliable and predictable responses to fire, but have proven insufficient when fires are far from the sprinkler. Moreover, the distance increases the possibility that sprinkler activation will come too late to contain the fire spread. Given the technologies now available, the option of activating sprinklers on the basis of sensors located at a distance, such as in a deluge system, where large groups of sprinklers are activated simultaneously, could be expanded to an intelligent protection system with limited and controlled sprinkler operation. This motivated the development of the SMART sprinkler protection system that uses Simultaneous Monitoring, Assessment, and Response Technology (SMART).

SMART Sprinkler Development

Although the SMART sprinkler may have begun as a concept, it soon took physical form in an effort to make a practical determination of how best to identify a hazard and respond effectively (Xin, Burchesky, de Vries, Magistrale, Zhou & D’Aniello, 2017) [[i]]. The system design starts with the objective of detecting, locating, and suppressing fire as early and as locally as possible. Avoiding the damage that can come with unnecessary sprinkler activation was also a concern and an important driver in requiring two types of detection signals. Laboratory experiments were conducted with a range of fire sizes and fire locations being used to test the detection capabilities and evaluate the wireless communication employed in the prototype system. This was followed by efforts to distinguish optimal sprinkler spacing and activation scenarios. For both detection accuracy and speed of response, the experiments showed the value of using multi-detector technology, specifically for rising temperature and the presence of smoke. Having two types of sensor reduces the likelihood of a false trigger and can help pinpoint a fire more than a single sensor.

Figure 2. Comparison of fire sizes upon SMART and traditional sprinkler activation.

Speed of actuation was a key design goal. In many applications, a ceiling-only approach using traditional sprinklers is inadequate due to its slow response relative to fire propagation and sprinkler skipping. The aim of SMART sprinkler technology is to connect rapid detection, intelligent fire location and suppression to create more cost-effective solutions for users. Rapid detection achieves early suppression when the fire is still small, which is the key either to reducing water demand or suppressing more challenging fires. In principle, detection is limited only by the sensor’s capabilities. The controlled and coordinated activation of a group of sprinklers around the fire origin allows for proactive response to get ahead of fire spread via pre-wetting, instead of the reactive response by traditional sprinklers. For example, in a fire test of 7-tier (35 ft or 11 m) rack storage of cartoned commodities, the traditional sprinkler would not activate until flames were approximately 40 ft (12 m) above the floor. The SMART system activated when the flames were only 10 ft (3 m) high, a 75% reduction (Xin, Burchesky, de Vries, Magistrale, Zhou & D’Aniello, 2017) [[i]].

By separating fire detection from sprinkler activation, the sprinklers can not only be activated earlier but also in a more intelligent manner, which is essential for highly challenging fire scenarios. One example is that building structures such as purlins and girders can channel ceiling flow to create highly distorted sprinkler activation patterns (Chatterjee, 2019) [[ii]]. Slope ceilings can generate similar effects. These practical problems can be addressed by including corrections in SMART sprinkler activation algorithm. The separation of fire detection from sprinkler activation also opens up the possibility of using different kinds of sensors or a network of sensors, including wireless sensors, infrared flame detectors and video image-based detection. In principle, optical sensors provide the fastest response and potential 3D locating of the fire without being affected by the ambient airflow, if the field of view allows direct transmission of flame radiation. In contrast, smoke and heat detectors are subject to the limits of convective flow and impact of ambient airflow, which can be mitigated by placing either wired or wireless detectors in the rack (i.e., ceiling-only SMART sprinkler with in-rack detection).

Figure 3. Fire development in a 7-tier rack storage test of cartoned unexpanded plastic (CUP) commodity.

Validation Testing and Reliability

After confirming its fast activation and accurate location, the prototype SMART sprinkler was tested in a series of large-scale fires using cartoned unexpanded plastic (CUP) commodities up to seven tiers of rack storage. In all fire tests, the sprinkler protection was triggered based on two conditions: (1) trigger of at least one smoke detector; and (2) temperature rise > 5°C.  When these conditions were met in the 7-tier test, the flames barely touched the second tiers, in contrast to the traditional sprinkler where flames touched the 40-ft (12-m) ceiling. The flames never exceeded the rack storage height and the target array across a 4-ft (1.2-m) aisle was never ignited. Within 20 minutes of ignition, the fire was nearly extinguished. The SMART sprinkler not only suppressed the fire with less water, but also did so with much less fire and water damage. Recent large-scale fires tests conducted at FM Global demonstrated that water usage can be reduced by > 50% for a variety of protection scenarios including low-pile storage with high ceiling clearance and high rack storage using in-rack detection.

Figure 4. Comparison of water demand between SMART and traditional sprinklers.

Based on the rack storage fire tests completed so far, the SMART sprinkler has several advantages over the traditional sprinkler:

  • Detects fire very early in its development.
  • Determines the fire location accurately.
  • Simultaneously activates a group of sprinklers upon fire detection and location.
  • Controls or suppresses the fire using less water.
  • Results in less fire and water damage.

In addition to protecting HCFs, this technology is useful when the available water supply is limited. The key to using fewer sprinklers and less water is targeted water delivery. This advantage can also be vital from a cost perspective. If fire protection can be achieved with less water, users may be able to avoid the construction of large storage tanks or the installation of expensive pumps. Building owners may even be pleased with the “green” aspects of using less water. From an economic standpoint, locational accuracy and early detection combine to reduce the number of sprinklers that may be needed to achieve the same effect. Where a traditional design might require 12 or even 30 sprinklers, the same results might now be achieved using fewer than 10 sprinklers with even lower loss expectancy.

For reliability, FM Global Research has analyzed the availability of the prototype SMART sprinkler system (Chatterjee, 2016) [[1]]. The reliability analysis starts with an understanding of the system working mechanism so that a failure model and effect (unavailability) tree can be developed. This tree outlines the system components and their logic relationships to reflect relevant failure modes and effects, which, with component reliability data, feed a statistical model to compute the system availability subject to a given inspection, testing and maintenance (ITM) frequency.  The results show that with sufficient ITM frequencies, the SMART sprinkler can achieve comparable reliability levels to those of traditional sprinklers. The availability aspect is also important to the certification of the SMART sprinkler system, which should also include component testing, algorithm assessment and large-scale fire testing. Because the SMART sprinkler is new technology, one should be cautious about where and how it will be used. For instance, the exposure of the electronic components could damage the system affecting fire detection and sprinkler activation, ambient air flows could cause significant deviation to locate the fire origin using ceiling smoke and heat detectors, and obstructions could interfere with optical or image-based detectors. Furthermore, the working mechanisms of any new system will be carefully studied to rectify any unforeseen vulnerabilities.

Future of SMART sprinkler

The future of the SMART sprinkler is promising not merely because it delivers where a traditional sprinkler cannot, but because of its ability to suppress fires faster with less damage and less water. The potential reduction of water demand by at least half while decreasing fire and smoke damage could be a huge advantage, especially for occupancies storing high-value products, or for regions with limited water resources. Furthermore, SMART sprinkler fundamentally changes the way of detecting and locating a fire, as compared to a slow heat detector relying on buoyant flow as the choice for a traditional sprinkler. These advantages, together with the intelligent activation of multiple sprinklers at the same time will allow fire protection to reach its ultimate goal of extinguishing a fire as early and as locally as possible.


[[1]] Xin, Y., Burchesky, K., de Vries, J., Magistrale, H., Zhou, X. & D’Aniello S. (2017). SMART Sprinkler Protection for Highly Challenging Fires—Part 1: System Design and Function Evaluation. Fire Technology.  53: 1847–1884.

[[2]] Xin, Y., Burchesky, K., de Vries, J., Magistrale, H., Zhou, X. & D’Aniello S. (2017). SMART Sprinkler Protection for Highly Challenging Fires—Part 2: Full-Scale Fire Tests in Rack Storage. Fire Technology. 53: 1885–1906.

[[3]] Chatterjee, P. (2019). Sprinkler Performance under Sloped and Obstructed Ceilings. FM Global Technical Report: Project ID 0003059743.

[[4]] Chatterjee , K. (2016). Evaluation of the Availability of the SMART Sprinkler System. FM Global Technical Report: Project ID 0003053857.