printer friendly version

Where there is fire, there is not always smoke

Microprocessor-based intelligence has knocked the perennial bugs out of today's life safety systems. But can any fire alarm be truly fail-safe under all conditions?

The airport terminal was a hive of activity when the fire alarm evacuation signal sounded. Thousands of anxious travellers were herded to the exits as staff and security personnel tried their best to empty the terminal. No one was injured in the fire, and only minimal property damage resulted. However, the costs were both real - and avoidable.

The fire started in a maintenance room on the departures level, but because the initial flames were fed by cleaning fluid, they produced no smoke and the detector just outside the room failed to respond until the fire had spread into the walls.

Even then, airport staff, harassed by a series of false alarms in the weeks leading up to the incident, were slow to react. It was not until emergency response personnel arrived on the scene that the fire was discovered, but by then the terminal had already filled with smoke.

This story, though hypothetical, illustrates just how fragile the integrity of a fire alarm system can be and just how serious the outcome can get. False alarms and the unpredictable nature of fires make a dangerous combination.

Advances in technology have cleared the way for new products and approaches that would make it virtually impossible for an airport's system to fail.

Self-monitoring detectors improve reliability

In recent years, designers and manufacturers of fire alarm systems have pushed beyond conventional technologies into the realm of distributed intelligence.

Like most revolutionary ideas, distributed intelligence rests on a deceptively simple premise: spread the computing power of a life safety system among its devices, free the control panel from mundane processing tasks; and, decentralise the system's core processing functions. With the lightning speed we have come to expect of the technological age, the approach is quickly becoming an industry standard.

In itself, distributed intelligence is nothing new. The Internet, for example, was originally conceived, decades ago, as a means of providing a kind of life safety system that would support military communications in the event of war. On a micro-scale, modern fire alarm systems provide much the same fail-safe backup that enable them to continue to provide basic life safety functions, even if the control panel or network node is knocked out of action.

What is new is the capacity of today's intelligent life safety devices to do much more than simply send information to the control panel. The result is something that is both subtle and complex. Intelligent systems monitor their surroundings and adjust themselves to compensate for naturally-occurring environmental conditions. In other words, they know the difference between smoke and something that may look like smoke.

The driving force behind this development has been the need for a design that is more reliable and less susceptible to nuisance alarms. This has been accomplished through modifications to the way information is processed, rather than to the way it is gathered: even though tremendous gains in detector reliability have been made over the past few years, the basic principles of detection have remained virtually unchanged.

Ion-, photo- and heat sensors - the mainstays of any fire alarm system - still have their own specific applications for which they are best suited.

Trade-offs characterise past approaches

Because of the unpredictable nature of fire, manufacturers have found it necessary to modify detectors so they perform reasonably well under a whole range of conditions.

For example, a photoelectric detector must also be able to respond to a smouldering type fire. The result is a device that operates reasonably well, but not optimally. The trade-off has come at a price and false alarms have become the nature of the beast. The problem stems from the fact that detectors that are sensitive to smoke are also sensitive to dust; and those sensitive to heat can also be affected by humidity.

Intelligent systems have furnished the means of overcoming this problem. With the introduction of the analog detector a few years ago, and of addressable devices a few years before that, manufacturers were able to assign sliding alarm thresholds to devices.

This enabled the device to monitor its own sensitivity and 'understand' its environment. If dust or humidity levels increase the chance of a false alarm, the device itself is able to compensate automatically by raising its own alarm threshold. There is no danger, however, that the threshold will be pushed so far as to compromise the device's ability to detect fire. Before that point is reached, the device sends out a message that it is time for a cleaning.

On-board microprocessors have also provided a means of addressing another concern: the perennial problem of choosing the best type of detector for a particular application.

Choosing the best detector for a given application

With the advent of multisensor detectors, photo-, ion- and heat sensors have been incorporated into a single unit. Independently, these different types of sensors can come up with conflicting conclusions concerning the same environmental conditions. However, when these sensors are combined in a single smart detector they can be monitored over time, thus reducing the chance of the device reacting to the wrong set of circumstances.

And that is where the sophistication of the system comes into play. True multisensor detectors compare values received from the on-board independent sensors to a pre-set algorithm. The device's microprocessor can then determine whether there is an actual danger, or whether one of the sensors is reacting to a non-threatening environmental condition such as dust or humidity.

This data-filtering process means the detector will only initiate an alarm when conditions exactly match the characteristics of a fire.

Refinements simplify installation and maintenance

While all of this marks a great leap in terms of the dependability of fire alarm systems, such complex refinements raise serious questions about how far the technology has brought the industry, and where it will eventually lead. As the technology becomes more sophisticated, so too must its application.

Installing and maintaining these systems requires a higher level of skill than ever before.

Maturity of the life safety industry

The life safety industry is already addressing this issue. Now that the technology is established, there is less experimentation with radically new processes. Developers have begun to turn their attention to refining what they know works, and customers themselves are exhibiting more savvy when it comes to assessing new equipment. Today's customers are less likely to be dazzled by the technology and more inclined to question its place if it makes the system unduly complicated. In short, they are demanding that system sophistication be amply tempered with ease of use.

And manufacturers are listening. New software streamlines system set-up and verification, mimicking the familiar Windows interface. Control panels are becoming easier to operate. Touch screens offering graphical displays of building plans are rapidly replacing rows of buttons and switches.

Life safety in the South African context

In South Africa Fire alarm systems have traditionally been operated as standalone systems. However the end-user is becoming more sophisticated as a result of the wealth of information available to them through the Internet.

We are now finding that most of our customers are asking for the fire alarm system to form part of an integrated life safety platform that is also able to communicate to their building management system (BMS) through industry standard protocols such as LonWorks, BACnet, Modbus to name a few.

Without a doubt, the future of life safety is inextricably entwined with the technology that drives it.

Share this article:

Further reading: