The reality is however that multi-sensors are not new. The first analogue addressable systems were launched into the market in the mid 1980s along with optical heat multi-sensors. Most manufacturers now offer an optical heat multi-sensor in their range and it is commonplace to design a system using these.

Most of the multi-sensors produced today still use the same technology that was developed in the 1980's. There have been advances in the processing of the signals from the detector and some sophisticated software filters are used in some systems to reduce susceptibility to false triggers. More advanced fire detection control panels use statistical analysis over time, looking at trends and patterns in day-to-day non-fire situations. By logging this data, trend averages can be calculated, and a technique known as pattern matching used to determine whether certain types of fire (or false alarms for that matter) are present. From analysing this data, algorithms can be developed to control the various channels of the sensor, enabling some, and desensitising others, to achieve optimum performance.

A large amount of effort has also been invested in the mechanical design of detectors to keep foreign bodies out of the sensor chambers. Insects, especially thrip flies, have traditionally been a real problem during harvest time in triggering false alarms. As it is not desirable to make the insect mesh of an optical chamber any finer, "safe zones" have been created in some sensors so that if thrips do get in, there is a reduced chance that they cause a false alarm.

Once any system is commissioned and the building is occupied, the fire detection system then faces challenges that arise from both legitimate use of the building as well as possible misuse and abuse.

There are a high number of trouble-some applications with inherent risks that are common causes of false alarms. Steam, aerosols, vehicle exhausts or excessive cigarette smoke are all contributors to nuisance alarms. To a regular optical sensor, steam from a hotel bathroom is identical to grey smoke. Heat sensors in a kitchen cannot discern between heat from a fire and heat from an oven under a poorly sited sensor.

Dirty environments in process industries or manufacturing locations can also render optical sensors useless in a matter of weeks by coating the chamber with particles. This may even make it impossible for a fire condition to be recognised.

In summary, traditional multi-sensors such as optical scatter combined with heat are successful in detecting both smouldering and fast burning fires. They have also allowed us to largely design-out environmentally unfriendly ionisation detectors. However, the increasing adoption of optical heat multi-sensors has driven little improvement in false alarm rates.

Advances in Sensor technology

Recently, optical sensor performance has been significantly advanced with the introduction of dual optical technology. In these detectors, the optical chamber typically either contains two optical paths at different angles or two different wavelengths of light. In both cases, the two signals are used to provide information about the nature and size of the particles being sensed. This allows more accurate discrimination between particles of combustion and non-fire particles that would normally trigger false alarms.

With a dual angle chamber, the most practical arrangement is to have two Infra-Red (IR) LEDs with a single photodiode receiver. One IR LED will typically be at an angle of 120-140 degrees from the receiver, (termed the forward scatter angle). The second will be at ½ this angle, (termed the backward scatter angle). By dividing the intensity of the backward scatter signal by the forward signal level, a ratio can be obtained that is very low for steam, but progressively increases for smouldering fires that produce grey smoke (such as Standard Test Fires TF2 and TF3), up to flaming fires producing black smoke (such as standard test Fires TF1, TF4, TF5 and TF8).

There are various ways this ratio information can be used by the processing algorithms in fire alarm panels. The calculated ratio can be used to adjust the alarm threshold level down for ratios that correspond to flaming fires, thereby increasing the sensor?s dark smoke sensitivity and giving a flatter response to different fire types. The ratio is also used to adjust the alarm threshold up for very low and very high ratios, corresponding to steam and dust particles, thereby decreasing the system?s sensitivity to non-combustion aerosols. In all cases, the forward scatter signal level still goes through sophisticated signal processing to measure the particle density over time. In addition, as in standard optical heat multi-sensors, the dual optical heat multi-sensor benefits from the inclusion of a heat element to help in the early detection of hot, clean burning fires (like TF6), which produce invisible particles that are not easily detected by an optical-only sensor.

In addition to advances in optical technology, carbon monoxide gas sensors have recently become available as an option for fire detection systems. CO is normally a product of incomplete combustion, so its presence is a very good indicator of a slow smouldering fire but a poor indicator of flaming fires. In addition, it is easily diluted by, for example ventilation systems. So, as a single technology, CO detection cannot be relied on for safe fire detection. But, if it is integrated with other sensing technologies, it delivers a very powerful fire detector, especially if the CO is combined with a dual-optical sensor able to determine the type of smoke in a fire.

If these sensors are connected to a powerful fire control panel, sophisticated algorithms can be used to match patterns of activity with hundreds of hours of test data. Pre-defined rules can be used that optimise the performance of the sensor for real world applications. For example, in a hotel bedroom full of steam, the readings from the dual optical elements and the lack of CO would pattern match as a false alarm, so increasing the fire trigger threshold and avoiding an unwanted alarm.

Once the patterns and rules do see a match for a real fire, the control panel can be programmed to operate the alarm or better still, verify the pattern match with other sensors to see whether they have also seen any products of combustion.

In addition, to the patterns and rules, On some fire systems, operating sensitivity states can be defined that match specific applications and risks. Selection of these states in the control panel allows each and every sensor in the system to be individually set to an optimum state that will balance early detection of real events and rejection of false alarms. In more advanced systems, these states can be automatically controlled by the time clock and calendar. This allows sensitivity to be increased during the night and at weekends for example when buildings are unoccupied.

The greatest benefits are realised when a set of products has been designed to achieve optimum performance working together as a system. The Vigilon product family from Gent is one system that can deliver all of the above benefits to significantly improve performance in any fire detection application. Vigilon has the flexibility and performance to allow this improvement to be a reduction of false alarms, an increase in sensitivity to real fires, or a combination of both. The world-leading Gent S-Quad multi-criteria sensors, working with sophisticated patterns and rules built into the Vigilon fire panel combined with the ability within Vigilon to allow user-set, application specific states can provide a sensitivity to real fires with a degree of false alarm rejection we could only have dreamed about twenty years ago.

Go back »