Aspiring for Better

Early, highly sensitive smoke detection is critical for many industries. This system provides the intelligence necessary.

• By Marc Barrera
• Jun 13, 2007

PEOPLE aren't perfect and are sometimes prone to making errors; it's the reason why pencils have erasers and keyboards have a backspace key. Of course, one can never completely remove all chance of error, but actions can be taken to minimize its occurrence. Requiring extensive training and certification of safety personnel in order to provide the safest work environment possible is one way. Steven Joseph, CEO, North American Operations for AirSense Technology USA Ltd., hopes his company's device will help further reduce that chance of error in one particular area: fire protection.

The Stratos® high-sensitivity, aspirating relative sensitivity smoke detection system has something no other detector on the market has--ClassiFire® perceptive artificial intelligence (AI) technology, he said. Unlike most aspirating smoke detection systems that use fixed sensitivity threshold settings--set by the manufacturer or a technician during installation--that regulate the levels of smoke required to produce an alarm, this technology allows the system to learn about its environment over time and adjust its sensitivity accordingly.

"To avoid having technicians coming in and arbitrarily setting a threshold, we're taking the burden off their shoulders," said Joseph, adding that during installation, you give the system a reference threshold based on its application that is meant only to serve as a starting point. "For the first 15 minutes of operation, when the detector first gets fired up, it goes through what is called a 'fast learn.' During that fast learn period, it analyzes the ambient background and builds a fast growing histogram. Then over the next 24 hours, it fine tunes that with a slow-growing histogram. And then from there on after, for the life of the detector, it continues to learn that ambient background threshold and adjusts the sensitivity accordingly."
This auto-adjusting feature, coupled with its high sensitivity, makes the system ideal for a wide range of applications where several factors can be momentarily introduced and must be adjusted for. These areas can include  warehouses with diesel-powered forklifts, rental storage facilities, production facilities, welding shops, waste recycling plants, etc.

The Filter Factor
One issue present with all aspirating detectors using filter cartridges is the fact that over time, the filter can become overloaded with particulates. The danger this poses with a fixed-sensitivity detector is that it lessens the device's sensing capabilities as the clouded filter begins to increasingly block more and more combustible particulates from reaching the sensor for detection.
With Stratos' AI capabilities, Joseph said, the device will note the changes in the sampling environment and increase its sensitivity accordingly. "Our system performs a function where ClassiFire will note over time that the environment has gotten cleaner and will adjust the sensitivity to follow more closely, and eventually you'll get a sensor fault to replace the filter," he said.

The detector not only continuously monitors and adapts itself to suit the environment, but also can be programmed to distinguish between changes in its setting, such as day and night, occupied and unoccupied, and operational and non-operational levels, and during holiday shutdowns.

A factor that can cause detection systems to falsely alarm, Joseph said, is the influx of particulates from the outside surrounding environment through a building's ventilation system. "In a lot of applications where high sensitivity is desirable, such as a telecom communications facility, the makeup air handlers will bring in a certain percentage of fresh air, and often a lot of these telecom sites might be in residential areas. It's not unusual when someone is out in a patio having a barbecue for the makeup air handlers to bring in a particulate and introduce that into the space where high-sensitivity sensors are deployed and actually cause it to go into alarm," he said.

The Stratos system combats this occurrence through its ability to use any of its monitoring devices as a reference detector with an attached probe that is inserted into a building's duct work. By network, that detector will alert the system of the outside pollutant and the system will adjust itself to offset the influence.

The Human Element
Relying solely on a machine, regardless how advanced it may be, may not be a wise decision. Joseph said steps have been taken to ensure the human supervisory role is not eliminated. "It also has real-time viewer software we provide for free that anyone can load into their PC," he said. "You can then go in and look at the event logs, print the event logs. You can also look at a real-time view of the sensitivity threshold, where historically it has been, and also look at the actual live output of the detector in terms of obscuration. You can also pull open graphical smoke trends and look over historical periods and actually see graphically where the smoke levels have been and where your sensitivity levels have been."
Joseph said the device's versatility makes it very economical by eliminating the need for superfluous devices and components. "Installing one detector that replaces a whole slew of them decreases the costs of installation and the ongoing life-cycle costs," he said, noting the detectors can have anywhere from 14 to 100 sampling ports, depending on the model. "The advantage there is that you're not running conduit, pulling wire, mounting conjunction boxes, and mounting several detectors."

This article originally appeared in the June 2007 issue of Occupational Health & Safety.