When the aftermath of a serious fire is being investigated, one of the
most common questions is: Why did the fire get so large? Until
relatively recently, the 'large' questions could only be answered qualitatively,
since means of quantifying a fire size in engineering units did not exist.
Eventually, it was recognized that since heat is the energy output of
the fire, and scientific means exist for measuring energy, the problem
may be soluble. The principles are clear. Heat is measured in units of
Joules. What is usually more of interest is the rate at which heat is
released, not the total amount. The heat release rate (HRR) can thus be
measured in Joules per second, which is termed Watts. Since a fire puts
out much more than 1 Watt, it is usually convenient to quantify the HRR
in kilowatts (1000 W) or megawatts (a million watts).
measurement of HRR
Prior to the 1970s, such ideas, while theoretically accessible, were not
usable, since actual means of measuring HRR from fires were not available.
The first instruments for HRR measurement started being available in the
1970s and they were bench-scale devices. (One specialized unit had been
already built in the 1950s in one lab.) Bench-scale means such instruments
can measure samples on the order of a few inches or a few centimeters
in size, but not real objects that could be man-sized (or even warehouse-sized).
The early HRR instruments (OSU apparatus, developed by Prof. Ed. Smith;
NBS-I calorimeter, developed by Alex Robertson and Bill Parker; etc.)
suffered from normal first-generation issues of usability and cost. The
NBS-II calorimeter, for instance, cost NIST $250,000 to build in 1977-78
(actual 1977 dollars). Shortly after joining NIST in 1977, I was tasked
to find a better way. Several years of exploration elapsed, and by 1982
I had invented the Cone
Calorimeter, in its first iteration. This has since become the world
standard, available at test laboratories around the globe.
calorimeters (large-scale products calorimeters)
Having a bench-scale HRR apparatus is not enough for comprehensive studies
of fires. In many cases, it is necessary to study the HRR of objects in
their full scale, or at least nearly full-scale. This development was
also started around 1979, and by 1982 two different apparatuses were independently
invented. The NIST furniture calorimeter was developed by myself, along
with Doug Walton, Randy Lawson, and Bill Twilley. The FMRC products collector
was developed by Gunnar Heskestad. These have also now become used around
the world and are the basis of numerous standards of ASTM, NFPA, and other
The final HRR measuring apparatus which was needed was a room calorimeter.
Furniture calorimeters can measure the HRR of discrete objects, able to
support themselves on the floor. This does not include such products as
ceiling tiles nor wallboard. Also, special measuring issues arise when
one wants to measure a whole burning room, fully furnished. For such studies,
room calorimeters were needed. Room calorimeters were developed in a parallel
effort between Fred Fisher and Prof. Brady Williamson at UC Berkeley and
by Billy Lee and Jin Fang at NIST. This effort was also largely completed
in 1982, meaning that instruments of all three needed scales became available
nearly simultaneously in 1982.
scale to use?
It is costlier and more difficult to test in larger-scale instruments,
thus it would seem that preference would always go towards running a bench-scale
test. This is not necessarily true, since to make intelligent use of the
bench-scale data one needs a predictive model. In other words,
it is not of much interest to know what a 10 cm size sample would do;
what is of interest is the full-scale behavior of a piece of furniture,
appliance, wall covering, or even a whole room. For some categories of
objects, such models have been developed. These include upholstered furniture,
wall linings, carpets, and some others. But the available categories are
few, while the types of objects which can potentially be of interest in
fire reconstructions are numerous. Thus, one of the things which must
first be determined is whether it is reasonable to run bench-scale tests
or whether full-scale testing is needed. We may note that for polymer
manufacturers and others developing new materials, it is often sufficient
to only use bench-scale testing. This is because they mainly wish to find
the relative differences in fire behavior, while actual product performance
may not be relevant to them since they do not even make the end product.
important role of HRR in fires
HRR is not just 'one of many' variables used to describe a fire. It is,
in fact, the single most important variable in describing fire hazard.
(The only notable exception is for explosions). There are three main reasons
1. HRR is
the driving force for fire.
The HRR can
be viewed as the engine driving the fire. This tends to occur in a positive-feedback
way: heat makes more heat. This does not occur, for instance, with carbon
monoxide. Carbon monoxide does not make more carbon monoxide.
2. Most other
variables are correlated to HRR
of most other undesirable fire products tends to increase with increasing
HRR. Smoke, toxic gases, room temperatures and other fire hazard variables
generally march step-in-step with HRR as HRR increases.
3. High HRR
indicates high threat to life.
hazard variables do not relate directly to threats to life. For instance,
if a product shows very easy ignitability or high flame spread rates,
this does not necessarily mean that fire conditions are expected to be
dangerous. Such behavior may merely suggest a propensity to nuisance fires.
High HRR fires, however, are intrinsically dangerous. This is because
high HRR causes high temperatures and high heat flux conditions, which
may prove lethal to occupants.
is so important, why are regulators not regulating it?
In the US, over the last decade, HRR has shown up in various regulations
and specifications, but this has been in specialized areas. Where it has
not yet shown up in is in the building codes. The US model building codes
still regulate products according to the Steiner Tunnel Test. This test
was developed during the late 1930s and early 1940s and, of course, predates
all of modern fire protection engineering knowledge. The test controls
flame spread which is not, as noted above, a primary factor in determining
human untenability. Over the years, a number of research projects documented
various shortcomings of this test. The basic reason why we have not yet
progressed beyond 1940s technology in the building codes has to do with
the inertia of the process and of the lack of funding resources necessary
to propel a building code change. In the US, there is no public-interest
organ with specific funding to conduct research leading to building code
improvements. Changes, instead, are usually originated by commercial entities.
As of now, no commercial group has decided that it would be advantageous
for them to sponsor a change, intended to introduce improved engineering
methods in this area. In fire litigation however, HRR testing is well
established, and eventually it is also certain to become utilized in building
It is certainly
wise to always control ignition sources and also to use less ignition-prone
materials, when possible. Such a strategy, however, can never be relied
upon to avoid an ignition. Neither HRR nor any other consequences of
fire will come into play as long as there is no ignition. However, when
an ignition does occur, limiting the HRR means that the fire has a chance
to be controllable and not disastrous.
have taken measures to control the ignitability, so we don't have to
worry about HRR
also realize that if the application is not in aircraft safety, military
or NASA areas, the affordable, commercial materials that are available
are not very ignition resistant. Studies have shown that even small ignition
sources normally apply about 35 kW m2 heat flux to their target. If one
then seeks materials able to resist an ignition flux of 35 kW m2,
one finds that these are rare and costly.
rests on the imprecise definition of the term 'toxicity.' Regulatory officials
sometimes presume that this means that 'toxic potency' is the root problem
and that this is what must be controlled. Toxic potency is the toxicologist's
term for defining how toxic is the substance when you inhale 1 gram of
it. But of course the victim will inhale something other than 1 g of
it. How much of the substance will be inhaled is governed by the
fire's mass loss rate. The mass loss rate is closely proportional
to the HRR of the fire. Now, what is important to realize is that studies
at NIST and elsewhere have shown that for commercial products, burned under
realistic fire conditions, toxic potencies vary only within a narrow band.
By contrast, mass loss rates (same as HRR) vary over an enormous range among
products of any given type. Since both toxic potency and mass loss rate
affect the total impact of the fire on the victim, it is clear that effective
control can be mounted by limiting mass loss rates, but there is little
that can be achieved by attempting to control toxic potencies.
tell us that inhalation of toxic fire gases is the main cause of fire
deaths, so we should control toxicity, not HRR
For further reading, see the textbook Heat
Release in Fires.