Proceedings of the ASME 2011 International Mechanical Engineering Congress & Exposition

IMECE2011
November 11-17, 2011, Denver, Colorado, USA

IMECE2011-65159
APPLYING ADVANCED FMEA METHODS TO VEHICLE FIRE CAUSE DETERMINATIONS

Kerry D. Parrott, BSME, CFI, CFEI/CVFI

Sr. Engineer / Fire Investigation Analyst
Stahl Engineering & Failure Analysis, LLC
3201 Stellhorn Road, Suite C149
Fort Wayne, Indiana 46815
Kerry.Parrott@Stahl-Engineering.com

Pat J. Mattes
Quality Tools Consultant
9732 Lafayette Center Road
Yoder, Indiana 46464
patmatt247@gmail.com

Douglas R. Stahl, PE, CFEI/CVFI

Principal Engineer
Stahl Engineering & Failure Analysis, LLC
3201 Stellhorn Road, Suite C149
Fort Wayne, Indiana 46815
Doug.Stahl@Stahl-Engineering.com

ABSTRACT

following a systematic approach utilizing the scientific method

for fire origin and cause determinations. The rFMEA
methodology is proposed as a fire investigation tool that
assists in that process. This reverse FMEA methodology
will then be applied to a hypothetical, illustrative case study to
demonstrate its application.

This paper proposes that the advanced Failure Modes and

Effects Analysis (FMEA) techniques and methodology
currently used by the automotive industry for product and
process design can be reversed and used as an effective
failure/root cause analysis tool. This paper will review FMEA
methodologies, explain the newest advanced FMEA
methodologies that are now being used in the automotive
industry, and will then explain how this methodology can be
effectively reversed and used as a failure analysis and fire
cause determination tool referred to as a reverse FMEA
(rFMEA). This paper will address the application of these
techniques and methodology to vehicle fire cause
determination. This methodology is particularly suited to
situations where multiple potential fire causes are contained
within an established area of origin. NFPA 921 Guide for Fire
& Explosion Investigations [1] and NFPA 1033 Standard for
Professional Qualifications for Fire Investigator [2], often
referenced by the fire investigation community, prescribe

Key words: reverse FMEA, rFMEA, failure analysis, fire

investigation, origin, cause, effect, vehicle, failure modes and
effects analysis, FMEA, root cause analysis, RCA

BACKGROUND
Most modern motor vehicle and equipment manufacturers and
their suppliers use a Quality Management System such as QS9000 promulgated by the Automotive Industry Action Group
(AIAG) or ISO/TS 16949:2002, developed by the
International Automotive Taskforce (IATF), as part of their
product design processes. These systems provide a common

Copyright 2011 by ASME

Downloaded From: http://proceedings.asmedigitalcollection.asme.org/ on 03/05/2016 Terms of Use: http://www.asme.org/about-asme/terms-of-use

guideline for manufacturers to continually improve their

products; through emphasis on the prevention of errors and
defects in the design and manufacturing stages, through
reduction of design/manufacturing variation needed to
accomplish the required system/vehicle functionality, and
through consideration of potential customer or environmental
impact on the intended functionality [3].

automotive industry today have been designed and developed

using this methodology.
Root Cause Analysis (RCA) is a general engineering term
applied to a wide variety of reactive quality tools that have
historically been used to analyze failures after they have
occurred with the purpose of arriving at a root cause. The
root cause is the first, the base, the ultimate, or the initiating
cause that starts an unwanted sequence of events. It can also
be thought of as the single event which, had it been prevented,
would have stopped a causal sequence of events. It is the
answer to the why question. Common RCA techniques
used in the automotive industry include 8-Disciplines, Fault
Tree Analysis, Ishikawa Diagrams, Pareto Analysis, and the 5Why methodology favored by Toyota Motor Corp., and
originally developed by Sakichi Toyoda.

Failure Modes and Effects Analysis (FMEA) is one of the

most widely accepted and commonly used quality product
design (dFMEA) and process (pFMEA) tools in the
automotive industry today. FMEA was initially introduced in
the 1940s by the US military and quickly gained acceptance in
the growing aeronautical and aerospace fields [4]. The tools
and techniques gained favor in the automotive sector, and
subsequently the associated supplier sector, and to date are
gaining ground in Medical, Information Technology, and
Renewable Energy sectors. Since its early introduction into
the American automotive industry by Ford Motor Company,
this ever evolving tool has become a powerful technique
helping designers to recognize and evaluate potential product
failure modes early in the design and manufacturing processes
and, hence, eliminate or reduce potential failures and their
associated effects [5].
FMEA is integrated into the AIAG
Advanced Quality Planning Process (APQP). AIAG was
originally sponsored by GM, Ford, and Chrysler in 1982, and
now has expanded to include most all major vehicle OEMs
and OEM suppliers in the automotive and truck industries.
The AIAG also oversees the publication and distribution of
QS-9000 and ISO/TS 16949:2009, the newest standards for
automotive quality system and customer specific
requirements, quality manuals, and training [6].

The common and widely accepted base presumption for all

these different techniques is that for every effect, there was a
prior-occurring cause [8], hence the commonly referred to
cause and effect relationship.

DISCUSSION OF ADVANCED AND REVERSE

FMEA CONCEPTS AND METHODOLOGY
The fundamental premise of this paper is a two-fold proposal
on modifications to the traditional, predictive FMEA style of
analysis and also to the cause and effect relationship.
The first proposal introduces the concept of Advanced
FMEA which addresses the sequence linearity hierarchy, and
the relationship between function and failure.

More specifically, FMEA is an predictive engineering tool and

technique that strives to drive a deeper understanding of the
potential causes and effects in a design, and it helps define and
predict the possible effects of a potential failure while the
design team is still in the design phase of the product. In this
way, failure modes can be designed out of a product before
they are ever produced [7]. FMEAs also assign severity,
occurrence, and detection rankings that help to guide and
prioritize the design process. Finally, all of this information is
tabulated for future reference and updated as new information
becomes available. As such, it is a repository for product and
process design and performance knowledge, and also a
foundation for continuous product and process improvement.
Most products and processes involved in the modern

The second proposal is that the Cause - Failure - Effect

model introduced in Advanced FMEA can be conducted in the
reverse order, and enhanced with 5-Why root cause analysis
methods. The failure sequence then becomes Effect - Failure
- Cause. This sequence can be applied to the determination
of cause in vehicle fires (and more generally, to other failure
analysis investigations).
First, lets discuss and understand modern automotive FMEAs
in more detail. Advanced FMEA methodology goes beyond
the simple Cause and Effect relationship and teaches that
there is an intermediate step, namely the failure itself. A
Cause and Effect relationship really doesnt exist; the failure
has to be in the middle. Every effect that is noticed must have

Copyright 2011 by ASME

Downloaded From: http://proceedings.asmedigitalcollection.asme.org/ on 03/05/2016 Terms of Use: http://www.asme.org/about-asme/terms-of-use

in promoting the concept of Function Brainstorming, in lieu

of failure brainstorming, and in parallel with methodologies
first introduced by John L. Lindland [10], which offers
considerable alignment and focus on the tie-ins between
function analysis and failure mode identification. Once a
function or set of functions is established, all potential failure
modes will then fall into one or more of just seven different
categories listed below.

a preceding failure that occurred and for every failure that

occurs there must have been a preceding cause. Indeed, there
are multiple effects and multiple causes of failure but, for sake
of simplicity in this paper, only singular levels are considered
here. The Cause - Failure - Effect model is illustrated in
Figure 1 below.

1.
2.
3.
4.
5.
6.
7.

Figure 1

Another key concept in Advanced FMEA is that function

precedes failure. Functions are to be thought of as a set of
intended outcomes, such as provide heat, provide light,
transfer energy, etc., and can even be thought of as a set of
outcomes to be avoided, such as avoid transmission of
sound, or tank should not leak. A failure mode can then be
thought of as unintended outcomes for those functions [9].
Using the above examples, a failure mode would be the
inability to provide heat or light, or transfer energy. Similarly,
a failure mode for outcomes to be avoided would be that they
were not avoided, and therefore, failures would be identified
as sound transmitted, or tank leaked. This is in contrast
with the traditional FMEA practices that involve the analysis
team establishing the intended functions of the item being
studied, the potential failure mode, the potential effects of that
failure, and then the potential causes of failure (controls and
action-planning intentionally left out of the immediate
discussion). While the traditional style is a practical and
effective mentality, the methodology misses a few critical
points especially in terms of the linearity and sequencing of
the analysis steps.

Does not function

Functions insufficiently (too little)
Functions excessively (too much)
Functions too soon (too quickly)
Functions too late (too slowly)
Functions unevenly
Functions erratically

This methodology first ensures that the design team fully

understands the functions the product or process needs to
accomplish; then offers a clean, concise pattern for the
determination of potential failure modes against those
functions. Using the previous example of provide heat as an
established function, the analysis team can then use the failure
categorization described above to rationalize the validity of
the potential failure modes:

During the analysis of functions to determine failure modes, a

common mistake that analysis teams make, and one which is
typically promoted throughout the automotive industry, is the
concept of Failure Brainstorming. In such activity, the team
will typically spend their energies brainstorming all of the
potential ways a particular product could fail. The downfall of
this activity is two-fold; it focuses on failure in lieu of
intended functions, and it tends to focus on failure modes that
are extremely unlikely to ever occur even if numerous noncoincidental conditions spontaneously coincided.

Primary failure
1. Does not provide heat (valid)
Range failure
2. Insufficiently provides heat (valid)
3. Excessively provides heat (valid)
Time failure
4. Provides heat too slowly (valid)
5. Provides heat too quickly (perhaps
invalid)
Stability failure
6. Unevenly provides heat (valid)
7. Erratically provides heat (valid)

Once functions are determined, followed by failure mode

identification, the establishment of potential causes of failure
then becomes the focus. Causes of failure can be numerous,
and are typically viewed as the immediate reasons for failure
modes. However, failures can also be precipitated by other
failures and this is known as an error sequence or failure
sequence. Since every failure has a cause, then it stands to
reason that a singular failure can have a cause or a cause of a
cause. This is known as a cause sequence and is the

The focus should be on brainstorming the functionality of the

component, subsystem, or system under development or
review. One of the authors, Pat Mattes, has been instrumental

Copyright 2011 by ASME

Downloaded From: http://proceedings.asmedigitalcollection.asme.org/ on 03/05/2016 Terms of Use: http://www.asme.org/about-asme/terms-of-use

foundation for the 5-Why style of Root Cause Analysis

(RCA) technique.
Generally, most all potential motor vehicle failure mode
causes can be generally classified into five categories, based
on the authors and industry experience, as illustrated in
Figure 2.
Variation
Customer
use

Adjacent
Systems
Cause
Categories

Changes
over time

Figure 3

This figure also illustrates how the FMEA and RCA

techniques can be procedural opposites of each other, enabling
the same methodology to work in either direction. The
fundamental insight is that in traditional FMEA, Causes lead
to Failures and Failures lead to Effects; this relationship is also
true in the reverse direction for RCA, Effects are analyzed to
determine Failures, and Failures are analyzed to determine
Causes. This allows us to use a reverse FMEA (rFMEA)
methodology to determine fire causation in motor vehicles and
their systems. During a vehicle fire investigation, the fire
investigation analyst starts by documenting the observable
effects noted in the area of origin, analyzes these effects to
determine the functional failure or failures that precipitated the
observed effects, and then deduces causes of the failures to
determine cause of the fire.

Environment

Figure 2

Variation can include variations from design, manufacturing,

or assembly specifications. Some causes can also be some
combination of these categories. With knowledge and
experience gained in various vehicle application segments,
these generalized categories can be customized to tailor an
analysis to specialized markets.
FMEA techniques are also currently in use and the basis for
qualitative Fire Risk Analysis within the Fire Protection
Engineering community [11].
RCA techniques strive to determine how every effect can be
traced back to its original cause. So if we apply the new
FMEA sequencing to RCA, we can say that every effect, such
as an effect observed during a vehicle fire investigation, is
associated with a failure, and that every failure has a cause,
and because of the cause sequence, every cause also has a
cause. So from a practical application approach, FMEA and
RCA can be thought of as mirror opposites in time, for
product performance. The FMEA looks forward into the
future to predict how a design will perform or possibly fail.
The RCA looks back into the past to see what went wrong
during the design, manufacturing, or maintenance process
[12]. Therefore, since these sequences of events are mirror
opposites, we can reverse the FMEA methodology to arrive at
a root cause methodology. This chronology in the life cycle
of a vehicle system or component through design, a fire event,
and the causal investigation can be symbolically expressed as
shown in Figure 3 below.

Application of the reverse FMEA technique also opens the

possibility of referencing the original design and process
FMEA(s) to determine if the cause was evaluated as an
original design or process potential failure mode. New causal
information can then be added to previous FMEAs, providing
real world feedback to the FMEA process for continuous
product improvement at the design level.

rFMEA APPLICATION TO FIRE INVESTIGATION

The National Fire Protection Association has estimated using
National Incident Fire Reporting System data and their own
fire department surveys that between 1999 and 2003 there
were an average of 325,100 motor vehicle fires per year in this
country with cars, trucks, and other highway vehicles
accounting for 95% of reported vehicle fires [13].

Copyright 2011 by ASME

Downloaded From: http://proceedings.asmedigitalcollection.asme.org/ on 03/05/2016 Terms of Use: http://www.asme.org/about-asme/terms-of-use

The usefulness of this reverse FMEA approach is that once

an area or point of origin is determined, this methodology
provides a tool to help sort out the complexities of each
system and interactions between the systems involved in a
systematic and well documented approach that scientifically
drives towards determining the root cause. It is especially
useful in situations where multiple potential causes may exist
and the true root cause is therefore difficult to establish. In the
proposed methodology, the standard method of establishing
the origin of the fire is retained due to the maturity and
stability and acceptance of the practice, but the improved
current methods within Advanced FMEA are reversed to
assist the fire investigation analyst in establishing a root cause
in a more robust and better documented fashion and in
accordance to the scientific method.

Determining the cause of a vehicle fire is generally considered

by the fire investigation community to be a challenging
endeavor at best. Vehicle fires can be very difficult to
investigate because of the multitude, complexity, and
proximity of the many mechanical, hydraulic, and electrical
systems found on modern vehicles. Each of these systems has
its own failure modes inherent to the design of a particular
system, to the interaction between different systems, and to the
continual changes that occur to the systems over time as they
operate, wear, and are maintained over time in a continually
varying environment. This complexity is further complicated
by the variations in systems from different manufacturers,
repairs that may or may not have been properly done,
modifications made to incomplete vehicles by up-fitters, final
stage manufacturers, and by modifications or abuse by
owners, operators, or maintainers.

The Scientific Method is generally well established, and

prescribed for use in the determination of the cause of a fire,
as referenced for the fire investigation community in both
NFPA 921 (Guide for Fire & Explosion Investigations) and
NFPA 1033 (Standard for Professional Qualifications for Fire
Investigator). This reference denotes it as the seven step
process noted below.

There are several other main factors that make motor vehicle
investigations so challenging. First is the large number and
variety of ignitable materials contained in a motor vehicle. In
2004, it was estimated the typical passenger car contained 258
pounds of plastic materials [14, 15]. Additionally, most fluids
typically found in motor vehicles are ignitable by hot surfaces
[1, 16, 17]. The second factor is the large number of potential
ignition sources contained within a vehicle [1, 18]. These
factors are important because the large amount of ignitable
materials often results in a large amount of damage that
eliminates the evidence, and allows the fire to propagate at a
rapid pace making the area or point of origin difficult to
determine [9, 20]. Since the specific cause of a failure is
directly related to the function of that system, an analyst must
first understand the design and function of all the various
automotive systems before their failure modes and their
corresponding causes can be well understood [21].

1.
2.
3.
4.
5.
6.
7.

Recognizing the need

Defining the problem
Collection of data
Analyzing the data
Developing a causal hypothesis
Testing this hypothesis
Arrive at a final hypothesis

This process is then repeated as new or additional information

becomes available until a cause or list of potential causes is
established. In circumstances where all potential causes have
been eliminated, the cause must then be identified as
undetermined. From a vehicle fire cause analysis standpoint,
steps one and two are given and always the same. A vehicle
fire occurred and we need to understand how and why the fire
happened, and we can do this by conducting a fire origin and
cause investigation [23]. Since this is generally always the
same, this methodology focuses on steps 3 through 7.

The first step in a fire investigation failure analysis is always

to systematically establish an area or point of origin by way of
the traditional fire investigation tools such as fire pattern
analysis, heat and flame vector analysis, damage survey, event
sequencing, fuel load considerations, and witness accounts [1,
22]. The second step in fire investigation failure analysis is
cause determination, which is the essential element towards
understanding the sequence of events in product design,
product use, and/or environmental conditions combined in
unexpected ways to produce the fire. Cause determination
involves identifying the first fuel ignited, the ignition source,
and the circumstances that resulted in the fire [1, 22].

Effect

Failure

Cause

Figure 4

Copyright 2011 by ASME

Downloaded From: http://proceedings.asmedigitalcollection.asme.org/ on 03/05/2016 Terms of Use: http://www.asme.org/about-asme/terms-of-use

Using the reversed Effects - Failure - Cause model

described in the chronology of a product and highlighted in
Figure 4, we will now describe how this sequence can be
applied to steps 3-7 of the scientific method as described
above. The first step in the reverse FMEA is to determine
the observable effects of the fire in the area or point of origin.
This correlates to Step 3, the collection of data. Collecting
data on a vehicle that has been involved in a fire consists of
noting all the observable effects of the fire within the area or
point of origin. Typical observable effects can be things such
as shorted and beaded wires, ruptured hoses, broken
turbocharger shafts, transmission fluid missing, cracked or
broken components, etc. This is the information that is then
listed in the first column of the rFMEA form in Appendix A.

probable root cause should become apparent after no more

than the 5th order cause according to the 5-Whys
methodology. To quote Taiichi Ohno, father of the Toyota
Production System, the 5-Why method is the basis of
Toyotas scientific approach by repeating why five times,
the nature of the problem as well as its solution becomes
clear [24]. The causal sequences that have to stop because
they no longer make sense are the hypotheses that fail. They
can be thought of as the causes that might have occurred, but
did not. Another way to think of this Effect - Failure - Cause,
Cause of Cause, etc logic path is to think of it as a tool to test
your causal hypothesis. The hypotheses that fail will have to
stop, and the hypotheses that do not fail will identify the most
probable root causes. This step in the procedure correlates to
steps 5-7 of the scientific method; developing and testing the
hypothesis, and arriving at the final hypothesis. Using FMEA
terminology, the final hypothesis can be describes as the most
probable root cause. This information can be added to
columns 4-7 of the form. The added advantage of this
methodology is that all hypotheses that were considered are
now well documented for future reference.

Remember from the discussion section that every effect is

associated with a failure. Therefore the second step in the
process is to analyze the observable effects and to consider
what the failure was that caused the observable effect. This
failure for each observable effect is listed in the second
column. Keep in mind during this stage that failures
essentially constitute a lack of function. A short circuit, for
example, is a lack of isolation or insufficient insulation of the
conductor. It could be related to factors such the type and
thickness of the insulation, how it is clipped to prevent
chaffing, or the appropriateness of the insulating material to its
environment. This is why it is so important that an
investigator must be familiar with the design function of the
components under evaluation in order to conduct an effective
analysis. This step in the process correlates to step 4 of the
scientific method, analyzing the data.

Another concept from advanced FMEA methodology of

importance is that of hierarchical alignment. Think of it as a
tool to help the analyst determine if events are linear
succeeded causes, or preceding events. It therefore helps to
differentiate between effects and causes, and helps to assure
that causes immediately succeed each other. Lets consider an
example where a shorted and beaded wire has been found in
the area of origin in a vehicle. This is our initial observable
effect. Since a beaded wire is an indication of a short circuit
that cannot happen on a properly insulated wire, the failure is
a lack of isolation of the conductor. For the conductor to have
been exposed to a ground and short it must have worn through
the wire insulation, and through the harness protective
covering (if used). Wear between surfaces generally does not
occur without relative motion, and relative motion cannot
occur if the harness is properly secured. Therefore the effect,
failure cause sequence looks as follows.

Next, the analyst has to consider that each failure had a cause.
Think of a cause as the immediate reason for the failure.
Remember that failures can have more than one cause, so all
potential causes have to be listed and then each one will be
further evaluated later to determine its corresponding causes.
By doing this, the methodology considers every cause that
In NFPA 921 terminology, every
may have happened.
hypothesis is being developed and evaluated, i.e. tested. The
cause or alternate potential causes are listed in the third
column of the form in appendix A.
The next step in the methodology is to determine the cause of
the cause (2nd order cause), and then the cause of the cause of
the cause (3rd order cause), and so on until the point where the
most probable root cause is found or to the point where the
analysis no longer makes sense and has to stop. The most

Copyright 2011 by ASME

Downloaded From: http://proceedings.asmedigitalcollection.asme.org/ on 03/05/2016 Terms of Use: http://www.asme.org/about-asme/terms-of-use

Conductor short to bracket

Failure to insulate conductor

Wear through wire insulation

In this example, the hierarchical alignment also helps to assure

the analyst that each step in the causal sequence is linear, i.e,
the very next immediate reason for the failure. If the causal
sequence is truly linear, then the Effect - Failure - Cause
template illustrated above can be shifted to the right and still
make sense. This is illustrated in Figures 5-7 above.

Harness loose

Cause

Relative motion - wire / bracket

Failure

Wear through harness covering

Effect

One of the most common criticisms of a 5-Why technique

incorporated in to this reverse FMEA is that it needs to be
conducted by engineers familiar with the parts or components
being studied and sometimes only identifies symptoms instead
of a true root cause. Although this criticism is valid, following
the linear hierarchy methodology explained above helps the
investigator to focus on immediately preceding causes of
failures and not on symptoms of failures. However, a sound
understanding of the function of the components and systems
being studied is still essential. The analysis should be
conducted by experienced and knowledgeable individuals.
Colin Gagg, Forensic Engineering Lecturer at the Open
University in Milton Keynes, England noted: "To recognize
how a component or system failed, the engineer must
understand how it worked and was manufactured in the first
place."

Failure

Cause

Wear through harness covering

Relative motion - wire / bracket

Harness loose

Effect

Wear through wire insulation

Failure to insulate conductor

Conductor short to bracket

Figure 5

Traditionally, FMEA concepts in failure analysis (including

fire cause determination) have been utilized in two different
ways. Previous design and process FMEAs have been
reviewed post-failure to see if the predictive analysis of the
FMEA properly considered the cause(s), effect(s), and failure
mode(s) found in the failure and possibly suggest variations in
these characteristics which were not previously included.
Secondly, FMEA concepts were used as a tool to assist in
analysis of the failure, using the information discovered in
analysis of the failure to construct a FMEA model for the
failure, but actually working in the reverse direction, whether
distinctly recognized or not, since the new information is the
effects, leading to failures, and then causes. The rFMEA
methodology characterizes and formalizes the common
application of FMEA and Advanced FMEA concepts and
methodology to failure analysis by using them in a reverse
analysis. The rFMEA would include new failure information
of the effect(s) and failure mode(s) provided by the specific
fire investigation analysis including factors, circumstances,
and possible causes, that would be outside the scope of
previous product design or process FMEAs (such as arson, for
example).

Failure

Cause

Harness loose

Wear through harness covering

Wear through wire insulation

Failure to insulate conductor

Conductor short to bracket

Effect

Relative motion - wire / bracket

Figure 6

Figure 7

Copyright 2011 by ASME

Downloaded From: http://proceedings.asmedigitalcollection.asme.org/ on 03/05/2016 Terms of Use: http://www.asme.org/about-asme/terms-of-use

NFPA 921 discusses the use of FMEA as a potential fire

analysis tool, but the usage discussed and demonstrated there
is not consistent with the Cause - Failure - Effect concept of
FMEAs as used in predictive automotive design and process
FMEAs. Rather, it appears to be an attempt to fit fire
investigation effects reactive information into a standard
FMEA format.

origin. This area of origin could not be narrowed down any

further.
As background information, also consider that as this is a
commercial vehicle, it has undergone annual DOT inspections
in addition to a well documented quarterly preventative
maintenance program. The operator of this vehicle was
required to hold a valid commercial drivers license (CDL)
and to abide by the license requirements. The primary
requirement of interest here is that the vehicle must undergo a
pre-trip inspection every day before it can be driven, and that
these reports must be kept on file. The inspection report for
this truck on the day of the fire shows all the engine
compartment fluids were full, and that the lights, gages, and
electrical system were in good working order. The inspection
reports for prior days also did not reveal any electrical or
mechanical issues with the truck.

If the fire area or point of origin has been correctly

determined, the final, correct hypotheses in a fire investigation
failure analysis must contain a cause that provides failure(s)
and effect(s) consistent with that point or area of fire origin
but the final root cause may be in another area of the vehicle
and the rFMEA analysis should provide that cause of a cause
using all of the information available to a fire investigation
analyst or engineer. An example would be an electrical short
remote from the fire area of origin, that caused the current to
be carried through a wire-reinforced hose transporting a
flammable fluid, which in turn caused a fire in the area of
origin.

Within this area of origin, three significant observable effects

have been identified.

The rFMEA concept and methodology proposed is a tool to

assist fire investigation analysts and engineers in the fire cause
failure analysis process and, as any other tool, should be used
by trained, knowledgeable persons familiar with the product
design and fire investigation analysis and again, as with any
analysis tool, will not succeed if the analyst fails to understand
the tool, pose the right questions in the investigation, and
provide the correct information to the methodology.

1.

The positive cable of the alternator charging circuit

was found to have signs of arcing and beading where
it short circuited to a steel bracket. This is a SAE 2
gauge conductor running from the positive terminal
on the alternator to the positive terminal on the starter
solenoid.
This circuit is covered with nylon
convoluted conduit for abrasion protection.

2.

The negative cable of the alternator charging circuit

that also runs between the alternator and a ground
stud on one side of the engine block was also found
to exhibit signs of arcing and beading against the
same steel bracket. This is also a SAE 2 gauge
conductor protected with nylon convoluted conduit.

3.

Engine was excessively low on motor oil (SAE 15W40). Only two quarts remained from a sump capacity
of 24 quarts.

EXEMPLAR rFMEA FIRE INVESTIGATION CASE

STUDY
As an example, lets look at an illustrative, hypothetical case
in which the area of origin has been determined to be on right
side of an engine compartment of a commercial DOT
(Department of Transportation) class 8 truck that caught on
fire while operating after several years of service. The process
of determining the area of origin for a fire in a commercial
vehicle is outside the scope of this paper and therefore will not
be discussed here. For the purpose of this case study, the fire
damage was concentrated only on right side of the engine
compartment with only minimal melting observed in other
areas, and the fire did not propagate into the cabin of the
vehicle. Since one entire side of the engine compartment was
substantially involved, it was determined to be the area of

Following our Effect-Failure-Cause model for the first

observable effect, the failure is a short circuit to the bracket.
This is the same sequence as depicted in figures 5-7. The
short circuit could only have happened if there is no insulation
on the wire, so we have two viable options for our first order
cause. Either the insulation was worn through in an area
where it readily visible during a pre-trip inspection or it

Copyright 2011 by ASME

Downloaded From: http://proceedings.asmedigitalcollection.asme.org/ on 03/05/2016 Terms of Use: http://www.asme.org/about-asme/terms-of-use

melted off in as a result of the fire. Other environmental

issues such as rodents chewing on insulation are generally not
an issue on vehicles where the engine compartment is
inspected daily and are therefore not considered here. For the
insulation to have worn through, it would first have to wear
through the convoluted nylon conduit that covers and protects
this circuit in the wiring harness. This is the second order
cause. Wearing through the harness cover requires relative
motion between the harness and the bracket, which is the third
order cause. This cannot happen unless the harness is loose.
This is the forth order cause. At this point it no longer makes
sense to follow this causal sequence as background
investigations on the design of the harness determined that this
harness is clipped at regular eight inch intervals, and no
repairs had ever been made to the harness. This information
was also verified with the pre-trip inspection reports and the
preventative maintenance records. A harness that is well
secured and has no relative motion cannot chafe. Chafing
means to wear down or rub away a surface by scraping. There
is no fifth order cause and this hypothesis does not provide a
probable root cause. If instead the insulation melted off in the
fire, the short circuit is obviously an effect of the fire and not
its cause. This causal sequence also stops, and this hypothesis
fails.

engine block itself, and the engine oil cooler. These are the
first order causes. The turbocharger in turn has three major
potential leak points which are its second order causes. The
engine block can only leak if it is cracked. This is its second
order cause, and the oil cooler has two potential leak points,
the base plate and the seams for the cooling fins. The
turbocharger was inspected and no problems were found, so
this causal sequence stops. No cracks were found on the
block, so this sequence stops as well. The engine oil cooler
was removed and found to have a cracked base plate. Further
investigation into the manufacturing process revealed a recent
design change to the thickness of the base plate material. This
is a third order cause. This crack allowed motor oil to be
sprayed on either the turbocharger housing or on the exhaust
manifold. Since either of these surfaces is sufficiently hot to
allow for a hot surface ignition of the motor oil, this is a
probable root cause for the fire and we have arrived at a final
hypothesis in compliance with NFPA 921 and 1033.

The second observable effect is arcing and beading on the

negative alternator cable to the same steel bracket as the
positive cable. Just like before, the failure here is a short
circuit to a bracket. This failure allows for two potential first
order causes. The first cause is that the ground cable short
circuited, but as it is a ground, this causal sequence makes no
sense and has to stop (negative common ground design). The
second is that the ground cable was energized by a prior short
circuit. This means this circuit has made contact to a positive
cable and is now energized. This in turn can only happen if
the insulation on the ground wire and its harness covering
were both worn or melted (second order cause). Again this
sequence has to stop because if this circuit was already
exposed to a fire long enough to burn through the harness
covering and the wire insulations, it is an effect and not a
cause.

This paper proposed that the advanced FMEA techniques and

methodology currently used by the automotive industry for
product and process design can be reversed and then used as
an effective failure/root cause analysis tool to help determine
the root cause of vehicle fires. It is particularly useful tool in
situations where multiple potential root causes can be
observed within the area of origin. We are calling this
methodology reverse FMEA denoted as rFMEA. The
fundamental insight that Causes have Failures and Failures
have Effects and that this relationship holds true in either
direction, (i.e., Effects involve Failures and Failures have
Causes). This insight allows the rFMEA methodology to be
an effective tool to aid in the determination of root cause in
vehicle and equipment fires, and also constitutes a systematic,
well-documented application of the Scientific Method as
prescribed in NFPA 921 and NFPA 1033.

This information has been tabulated in a sample form in

Appendix A.

CONCLUSION

The third observable effect was that the engine was low on
motor oil. The failure mode here that causes a fire is if hot
motor oils sprays on the hot surfaces of the turbocharger
turbine housing or the exhaust manifold which are located in
close proximity to each other. The engine components in this
area that can leak hot motor oil are the turbocharger, the

ACKNOWLEDGEMENTS
We would like to acknowledge Stahl Engineering & Failure
Analysis, LLC for all the support, as well as all the reviewers
for their helpful comments.

Copyright 2011 by ASME

Downloaded From: http://proceedings.asmedigitalcollection.asme.org/ on 03/05/2016 Terms of Use: http://www.asme.org/about-asme/terms-of-use

NFPA is a registered trademark of the National Fire Protection

Association.
SAE is a registered trademark of SAE
International. AIAG is a registered service mark of the
Automotive Industry Action Group.

[12.]
Senders, J. W. "FMEA and RCA: the Mantras of
Modern Risk Management." Qual. Saf. Heath Care 13 2004:
249-50. Print.
[13.]
Nicholson, John. "Vehicle Fire Problem." NFPA
Journal Jan/Feb 2007. Print.

REFERENCES
[1.]
"NFPA 921: Guide for Fire and
Investigations." NFPA. Web. 16 Nov. 2010.

[14.]
Long, Jr, R. T., Jeff D. Colwell, Rose Ray, Helene L.
Grossman, Ben Thomas, and Robert Strassburger. "Section
21, Chapter 1: Passenger Vhicle Fires." Fire Protection
Handbook 2008. 2008 ed. Vol. I. Quincy, MA: National Fire
Protection Association, 2008. Print.

Explosion

[2.]
"NFPA
1033:
Standard
for
Professional
Qualifications for Fire Investigator." NFPA. Web. 16 Nov.
2010.

[15.]
Tewarson, Archibald, James G. Quintiere, and David
Purser. Fire Behavior of Materials in Vehicle Crash Fires and
Survivability of the Passengers. Tech. no. 2005-01-1555.
Warrandale, PA: SAE International, 2005. Print. SP-1939.

[3.]
Chrysler Corporation, Ford Motor Company and
General Motors, 1998, Quality System Requirements QS-9000
3rd Edition, AIAG

[16.]
Tewarson, Archibald. Fire Behavior of Automotive
Polymers. Tech. San Fransisco: 8th International Fire and
Materials Conference, 2003. Print.

[4.]
Procedure for Performing a Failure Mode Effect and
Criticality Analysis, November 9, 1949, United States Military
Procedure, MIL-P-1629

[17.]
Colwell, Jeff D., and Ali Reza. "Hot Surface Ignition
of Automotive and Aviation Fluids." Fire Technology 41.
Second Quarter 2005. Print.

[5.]
(R) Potential Failure Mode and Effects Analysis in
Design (design FMEA), Potential Failure Mode and Effects
Analysis in Manufacturing and Assembly Processes (process
FMEA), Warrendale, PA: Society of Automotive Engineers
for Advancing Mobility Land Sea Air and Space, 2002. Print.

[18.]
Kirk, Paul Leland, and John D. DeHaan. Kirk's Fire
Investigation. Upper Saddle River, NJ: Brady, 1997. Print.
[19.]
Digges, Kennerly H., and R. Rhoads Stephenson.
Fire and Toxicity Properties of Underhood Fluids and Plastics.
Tech. MVFRI Research Summary. Print.

[6.]
"Quality." Automotive Industry Action Group. AIAG.
Web. 20 Apr. 2011. <http://www.aiag.org>.
[7.]
Potential Failure Mode and Effects Analysis:
Reference Manual. Southfield, MI: Automotive Industry
Action Group, 1995. Print.

[20.]
Routhier, Brian. "Section 21, Chapter 2: Fire Safety
in Commercial Vehicles." Fire Protection Handbook 2008.
2008 ed. Vol. I. Quincy, MA: National Fire Protection
Association, 2008. Print.

[8.]
Parker, Sybil P. McGraw-Hill Dictionary of
Scientific and Technical Terms. New York: McGraw-Hill
Book, 1989. Print.

[21.]
Lindland, John. System and Design Failure Mode and
Effects Analysis. Ann Arbor: QualSAT, 2005. Print. Q-25.3.

[9.]
Mattes, Pat. "FMEA Methods Seminar." FMEA
Methods Seminar. Hypersonic Technologies, Fort Wayne. 26
Oct. 2010. Lecture.

[22.]
Fire Investigator: Principles and Practice to NFPA
921 and 1033. Sudbury, MA: Jones & Bartlett Learning, 2012.
Print.

[10.]
Lindland, John L. 7fm: the Seven Failure Modes.
[S.l.]: Bella Group, 2007. Print.

[23.]
Mifsud, L. The Scientific Method. The National Fire
Investigator. NAFI, 2001. Print.

[11.]
Cote, Arthur E., John Raymond Hall, Pamela A.
Powell, Casey C. Grant, and Robert E. Solomon. "Section 3.8
Fire Risk Analysis." Fire Protection Handbook. Vol. 1.
Quincy, MA: National Fire Protection Association, 2008. 3135. Print.

[24.]
no, Taiichi. Toyota Production System: beyond
Large-scale Production. Cambridge, MA: Productivity, 1988.
Print.

10

Copyright 2011 by ASME

Downloaded From: http://proceedings.asmedigitalcollection.asme.org/ on 03/05/2016 Terms of Use: http://www.asme.org/about-asme/terms-of-use

APPENDIX A: rFMEA / CASE STUDY

Effect
Observed

Alternator
positive
1 cable short

Alternator
2 negative
cable short

Failure
Short to
bracket

Short to
bracket

1st Order Cause

2nd Order
Cause

Oil spray
on turbocharger
or
exhaust
manifold

4th Order
Cause

5th Order
Cause
Stop.
Harness tied
down in 8"
increments good repair

1. Wear through
wire insulation

Wear through
harness covering

relative motion
between
harness and
bracket

Harness loose

2. Wire insulation
melted in fire
3. Environmental
causes

Harness covering
melted in fire
Stop. (vehicle
inspected daily)

Exposed to fire

Stop. (effect,
not cause)

a. Wear through
wire insulation

Wear through
harness
covering

b. Wire insulation
melted in fire

Harness
covering
melted in fire
Stop. (vehicle
inspected daily)

1. Stop. Ground
not energized
2. Ground
energized by prior
short

1. Leak at
turbocharger

a. Leak at shaft
seal
b. Leak at oil
supply
c. Leak at oil
return

2. Leak from
engine block
3. Leak from oil
cooler

Probable
Root
Cause

NO
NO
NO

NO

c. Environmental
causes

3 Engine low
on oil

3rd Order
Cause

Stop. Block found

OK
a. Cracked oil
cooler base plate
b. Leak at seam

11

relative
motion
between
harness and
bracket
Exposed to a
fire

Stop.
Harness tied
down in 8"
incrementsgood repair
Stop. (effect
not cause)

NO

NO

Stop.
Turbocharger
OK
Stop.
Turbocharger
OK
Stop.
Turbocharger
OK

NO

NO

NO
NO

Design change
in material
thickness
Stop. Seams
are OK

YES
NO

Copyright 2011 by ASME

Downloaded From: http://proceedings.asmedigitalcollection.asme.org/ on 03/05/2016 Terms of Use: http://www.asme.org/about-asme/terms-of-use