Industrial Fire Journal - Fire & Rescue - Hemming Group Ltd
Handling rejection
Published:  10 July, 2018

FEATURE Conclusive research tests show a reduction in false alarms when sophisticated multi-sensor detectors are used, write Bernard Laluvein and Robert Yates.

A study conducted by the UK’s Building Research Establishment in partnership with the Fire Industry Association and a number of detection manufacturers has concluded that multi-sensor detectors are better at rejecting false alarms.

The research was a continuation of prior research from three previous projects, which looked at the occurrence of false alarms in the UK.  Over the past 10 years, the FIA and BRE have partnered with King’s College, Buckinghamshire, Milton Keynes Fire Brigade, and Scotland Fire Brigade to investigate and find solutions to the false alarm problem that perpetuates across the UK.

Those previous projects had provided anecdotal indications that employing multi-sensor fire detectors would help reduce the occurrence of false alarms.  However, hard evidence was missing to establish how effective multi-sensors could be in reducing false alarms. 

Twelve manufacturers from the UK and Europe were involved in the research. These represented most of the detectors sold in the UK.  From those 12 manufacturers, 35 different detectors were tested against two standard optical smoke detectors, one commercial, and one domestic.

There are many different sorts of multi-sensors. The research focused on multi-sensor detectors which consisted of an optical smoke chamber together with a heat sensor. However, even within this detector type, performance varies greatly dependent upon the design and the settings used. As a result, the research graded them into standard, intermediate and advanced performance.

Standard multi-sensor detectors were those smoke/heat detectors which used the heat element to modify the response of the smoke sensor, meaning that the detector would respond more rapidly when heat was present. With no heat present the detector may then also delay its response.

The ‘intermediate’ category were the detectors that used a smoke chamber design that incorporated features of the ‘standard’ detectors, but with added design benefits, for example a more specifically designed chamber to eliminate potential for false alarms, such as an insect going into the chamber, or the effect of dust which may cause a false alarm.

The ‘advanced’ category, which the research predicted was the most likely to be the most effective at reducing false alarms, were those which had the features of the ‘standard’ and ‘intermediate’ categories, but also employed sophisticated targeted algorithms which would eliminate a false alarm. Following categorisation, the next stage of the research was to test the sensitivity of the different multi-sensor detectors compared to standard optical smoke detectors.

Ten different fire tests were conducted, looking specifically at the performance of the detector and their ability to detect fires. The fire tests included the standard fire tests – TF2, 3, 4, and 5 – which are in EN54-7, the European standard for smoke detectors.  TF1 and TF8 were then added, which are additional fire tests that have been included in EN 54-29, a new European standard for fire detectors combining smoke and heat sensors.  TF1 is specifically a hot wood-burning fire that produces heat, yet not much smoke; and TF8 is a fuel-burning fire that produces very little heat, but very heavy smoke. In this way, the testing represented the extreme ends of the spectrum for fires that may occur in real life.

A good fire detector needs to be used in almost every circumstance and must be able to detect a whole range of fire types that may be encountered depending on the fire risk.

To the six test fires listed above, four new fires were specifically designed for the research (these were not listed in EN54), to extend the range of test fire-types. These were based on fire-retardant material or MDF material, creating slow burning fires, again replicating real-life situations.

All the detectors – both multi-sensors and optical smoke detectors – were tested against all of the different test fires.

While there were some expectations that the multi-sensor detectors would perform better in terms of sensitivity and ability to detect fires, the research results were not so clear, however. It was difficult to differentiate very easily between the performance of the multi-sensors as a group, and the performance of the standard optical smoke detectors. Essentially, detector performance depended very much on their sensitivity – the high sensitivity detectors detecting quicker and faster than the low sensitivity detectors.  This held true for both multi-sensor detectors and standard optical smoke detectors.  There was no real way of differentiating performance between the ‘old-style’ optical smoke detectors and the ‘new’ multi-sensor detectors.

In knowing that there was not much difference in ability to detect fires, the second test was then to test all the detectors in their ability to reject typical false alarm situations. Would the multi-sensor detectors perform better?  The false alarm testing was developed in conjunction with the University of Duisburg in Germany; the team there had some expertise in developing false alarm testing.  In addition, bringing in a university meant that there was a further element of impartiality and independence from the industry.

Five different false alarm tests were carried out in total: burning toast, cooking, steam, dust, and aerosol spray.  Other tests such as cigarette smoke and long-term dust were also considered, but these were abandoned because repeatable results were not possible due to the amounts of variability in these two tests.

There were two outcomes from the tests – the first being the results of the tests themselves, and the second outcome was that the initial grading of the multi-sensor detectors into ‘standard’, ‘intermediate’, and ‘advanced’ categories was correct.

During the false alarm tests, all the multi-sensor detectors responded later than the optical smoke detectors – either the domestic or the commercial smoke alarms. For example, on the toast burning test, some of the multi-sensor detectors (those in the ‘advanced’ category) responded a good minute after the optical smoke detectors. By responding later, this demonstrates that multi-sensor detectors are not as prone to alarm as optical smoke detectors to events such as burning toast or steam, providing better overall resistance to false alarms.

Additionally, the fire tests reflected the way the researchers had categorised the different types of detectors. The standard category responded first (meaning they were not as good at rejecting false alarms); then the intermediate category alarmed; followed afterwards by the advanced category of detectors (with the benefit of sophisticated algorithms to filter out false alarms), which responded latest.

The conclusion? The more sophisticated the multi-sensor detector, the less likely it is to false alarm. And additionally, multi-sensor detectors are better at rejecting false alarms than optical smoke detectors of any category.

With the conclusion that multi-sensor detectors are more effective at false alarm rejection, and that effectiveness increasing depending on the categorisation of the specific detector, there is scope in the future for the FIA to be involved with developing a means of grading detectors within the British and European Standards. In addition, the FIA, in conjunction with other stakeholders, is also considering the introduction of new false alarm tests that were developed for the research, within the European or British Standards.  This means that detectors will come with a label identifying the level of resilience to specific types of false alarm.

This article was first published in Industrial Fire Journal, Q2 2018. Click here to read the online version.

Bernard Laluvein is director of ADT Fire & Security and chair of the FIA’s Fire Detection and Alarm Council.

Robert Yates is FIA technical manager and secretary of the FIA’s Fire Detection and Alarm Council

To find out more about the FIA, visit