Chemical Engineering May 2016 - 72

Table 1. Common IndusTrIal
FIre sourCes
Input layer
Alcohols
Diesel fuels
Gasoline
Kerosene
Jet fuels
Ethylene
Hydrogen
Liquefied natural gas (LNG)
Liquefied petroleum gas (LPG)
Paper
Textiles
Solvents
Sulfur
Wood
UV, visible and IR spectral bands.
Process, safety and plant engineers
must choose from among
these technologies to find the device
that is best suited to their individual
plant's requirements for
flame monitoring by deciding upon
the importance of the detection
range, field of view, response time
and immunity against certain false
alarm sources.
Ultraviolet/infrared (UV/IR). By integrating
a UV optical sensor with
an IR sensor, a dual-band flame detector
is created that is sensitive to
the UV and IR radiation emitted by
a flame. The resulting UV/IR flame
detector offers increased immunity
over a UV-only detector, operates at
moderate speeds of response, and
is suited for both indoor and outdoor
use.
Multispectral infrared (MSIR). Advanced
MSIR flame detectors combine
multiple IR detector arrays with
neural network intelligence (NNT).
They provide pattern-recognition capabilities
that are based on training
to differentiate between real threats
and normal events, thus reducing
false alarms. MSIR technology allows
area coverage up to six times greater
than that of more conventional UV/IR
flame detectors.
NNT is based on the concept of artificial
neural networks (ANNs), which
are mathematical models based on
the study of biological neural networks.
A group of artificial neurons
in an ANN process information and
actually change structure during a
learning phase. This learning phase
allows ANNs to model complex relationships
in the data delivered by
sensors in a quick search for patterns
that results in pattern recognition
(Figure 3).
Flame detectors with NNT operate
similarly to the human brain;
they have thousands of pieces of
data stored in their memories from
hundreds of flame and non-flame
Sensor 1
Sensor 2
Output
Sensor 3
Sensor 4
FIGURE 3. Many flame detectors employ technology based on artificial neural networks (ANNs) to more
accurately analyze flames
events observed in the past. These
detectors have been trained through
NNT intelligence to recognize flames
based upon those data, and determine
if they are real events or potential
false alarm sources.
Visual flame-imaging flame
detectors. The design of visual
flame detectors relies on standard
charge-couple-device (CCD)
image sensors, commonly used in
closed-circuit television cameras,
and flame-detection algorithms to
establish the presence of fires. The
imaging algorithms process the live
video image from the CCD array
and analyze the shape and progression
of would-be fires to discriminate
between flame and nonflame
sources.
Visual flame detectors with CCD
arrays do not depend on emissions
from carbon dioxide, water
and other products of combustion
to detect fires, nor are they influenced
by fire's radiant intensity. As
a result, they are commonly found
in installations where flame detectors
are required to discriminate
between process fires and fires resulting
from an accidental release
of combustible material.
Visual flame detectors, despite
their many advantages, cannot
detect flames that are invisible to
the naked eye, such as hydrogen
flames. Heavy smoke also impairs
the detector's capacity to detect
fire, since visible radiation from the
fire is one of the technology's fundamental
parameters.
Flame detection requirements
When configuring a flame-detection
system and evaluating the available
ChemiCal engineering www.Chemengonline.Com may 2016
technology alternatives, there are
many performance criteria that must
be considered. The following sections
outline some of these important
detector criteria.
False alarm immunity. False alarm
rejection is one of the most important
considerations for the selection
of flame detectors. False alarms are
more than a nuisance - they are
both productivity and cost issues. It
is therefore essential that flame detectors
discriminate between actual
flames and benign radiation sources,
such as sunlight, lighting fixtures, arc
welding, hot objects and other nonflame
sources.
Detection range and response
time. A flame detector's most
basic performance criteria are detection
range and response time.
Depending on a specific plant-application
environment, each of the
alternative flame-detection technologies
recognizes a flame within
a certain distance and a distribution
of response times. Typically,
the greater the distance and the
shorter the time that a given flamesensing
technology requires to detect
a flame, the more effective it is
at supplying early warning against
fires and detonations.
Field of view (FOV). Detection
range and FOV define area coverage
per device. Like a wide-angle
lens, a flame detector with a large
field of view can take in a broader
scene, which may help reduce the
number of flame detectors required
for certain installations. Most of today's
flame detector models offer
fields of view of about 90 to 120
deg (Figure 4).
Self diagnostics. To meet the high71
Hidden
layer
Output layer

http://www.Chemengonline.Com

Chemical Engineering May 2016

Table of Contents for the Digital Edition of Chemical Engineering May 2016