Polarised Aerosol Characterisation
According to various statistics, over 40 % of the alarms of automatic fire alarm systems
installed in buildings are caused by nuisance aerosols, especially due to dust and water droplets
like steam or fog. Due to the higher sensitivity of aircraft fire detectors compared to detectors
in buildings, aviation is faced with a false alarm rate of even 100:1.
Against this background the Fire Detection Research Group at the Chair of Communication Systems
investigates since many years into new fire detection technologies with reduced false alarm ratio.
One outstanding research result is the polarimetric smoke detection technology "POLARISE".
POLARISE explores the possibility of distinguishing nuisance aerosols by their particular optical
light scattering characteristics, targeting a high specificity without reducing the sensitivity
to smoke. Rather than only evaluating the particle size, POLARISE focusses on special morphological
and chemical characteristics of the non-fire aerosols.
To investigate into the scattering characteristics of smoke, dust and water droplets like fog,
simulation tools based on the Mie-Theory and the Rayleigh-Debye-Ganz approximation were developed
and comprehensive simulations were carried out to explore the optimal scattering angles and polarisation
directions. With these results different measurement systems have been developed and integrated into
extensive measurement campaigns in the fire detection laboratory and in field experiments (e.g. stables).
First results are very promising. The discrimination between smoke and dust is based on the depolarisation
effect caused by the compact and irregular structure of dust particles, which cause a partial depolarisation
of the scattered light.
This effect can be observed by monitoring the light scattering of different aerosols illuminated by
a polarised light source, as depicted in figure 1. In figure 1 also results with smoke from an open
wood fire, a smouldering fire and dust is shown. From the left part of the camera pictures it can be
seen that dust partially depolarises the incident light.
Figure 1 Depolarisation measuring set-up and first results.
Figure 2 shows the depolarisation ratio sdepol, i.e. the ratio between the depolarised and the polarised
scattered light intensity of several different aerosols. TF1 – TF5 are different kinds of smoke arising
from standard test-fires according to the EN54. These test-fires include open fires (TF1, TF4 and TF5)
as well as smouldering fires (TF2 and TF3). Paraffin denotes the typical paraffin oil aerosol used
in the standard smoke detector tests. UltraFine, Medium, Cellulose and Dolomit 10 are test dusts with
different particle size distributions and particle shapes. These types of dust are used for the tests
of smoke detectors in non-fire scenarios or the test of automobile and vacuum cleaners air filters.
Figure 2 shows that the dusts show a distinctly higher depolarisation ratio than all other aerosols.
The depolarisation ratio can be therefore implemented as an indicator for dust.
Figure 2 Depolarisation ratio of different fire and non-fire aerosols.
The discrimination of water droplets (fog or steam) and smoke is performed by the analysis of the
polarisation ratio of the light scattered in a defined angle, similar to the rainbow effect.
Observations of the natural phenomenon of a rainbow show that they only occur at very pure water-droplets
at an angle of approximately 42 degrees from the direction opposite to the light source. Figure 3 shows
two images taken of a rainbow looking through a polarising filter. In the left image the transmitting
axis of the polarising filter was rotated around 55°, while in the right image this value is about 145°
(i.e. 55° + 90°). The comparison of both images show that the rainbow is highly polarised, as it can only
be seen in the left image. This effect is particular for pure rain droplets.
A similar effect called "fog bow" can also be observed at the scattering of smaller particles like
those in fog and water vapour. Smoke of smouldering fires also consists to a high degree of water, but
the mixture with soot changes the chemical composition of the particles, impacting on their optical
characteristics, so that no fog bow occurs.
Figure 3 Images of a rainbow through a polarising filter.
Figure 4 shows results of measurements of the polarisation ratio, i.e. the ratio of the scattered light
with parallel polarisation to the scattered light with orthogonal polarisation at the angle of a fog bow.
As described above, TF1 – TF5 are different kinds of smoke arising from standard test-fires according to
the EN54. These test-fires include open fires (TF1, TF4 and TF5) as well as smouldering fires (TF2 and TF3).
Paraffin denotes the typical paraffin oil aerosol used in the standard smoke detector tests.
Figure 4 shows that the polarisation ratio of fog and water vapour is much higher than the polarisation
ratio of different kinds of smoke. Thus the polarisation ratio can very well be applied for the discrimination
between fog, water vapour and smoke which is one of the major innovations in fire detection technologies in
the last years.
Figure 4 Polarisation ratio of different fire and non-fire aerosols.
Current works under the project POLARISE focus on the implementation of the investigated methods in
different applications and set-ups.