Content uploaded by Tamotsu Murakami
Author content
All content in this area was uploaded by Tamotsu Murakami on Feb 10, 2017
Content may be subject to copyright.
Abstract
The authors propose computerized support for Fault Tree
Analysis (FTA) based on new knowledge management in
product design. FTA is a method of analyzing and visualizing
the causes of a fault event by Fault Tree diagram (FT diagram)
that has a tree structure with logical step. Many methods of
support for FTA are studied, but they are not effective for our
FTA. We have developed the system of Computer-Aided FTA
(FTAid) for design engineers by the collaborative research
group (JATCO Ltd, The University of Tokyo and National
Institute of Advanced Industrial Science and Technology). We
reported this system in SAE2012 World Congress. After that,
the knowledge management for FTA and new functions of this
system have been studied continuously. In this paper we report
that FT diagrams used FTAid improved to be useful for design
skill of FTA, we study improvement of FTAid and the FTA
education system.
Introduction
A car is forced to global competition of the automotive industry
in the world. Europe, America, and Japan remain at the same
level or these car demand decline, and China, India, and
another market expansions for Asia are predicted.
Furthermore, the automotive industry of the countries such as
China, India will develop more after Europe, America, Korea
and Japan. Accordingly, the future world automotive industry
largely changes by global competition. Therefore, new systems
developed with high functionality in a vehicle will be necessary.
The drive system of a vehicle is equipped with the systems
which became complicated newly, for example an assistance
brake system which is a safe support system, Continuously
Valuable Transmission (CVT), Dual Clutch Transmission
(DCT), a hybrid system, and an electric vehicle.
Many engineers make full use of many methods and examine
it to achieve the targets of quality, performance, durability of
the complicated system in a vehicle. The methods of the
design engineering include design tools and process such as
reliability engineering : ex. Fault Tree Analysis (FTA) [4],
Failure Mode Effect Analysis (FMEA) [4], statistical analysis,
and axiomatic design [1]. In reliability engineering, the
Accelerated Reliability Testing / Accelerated Durability Testing
(ART/ADT) strategy components is advocated by Klyatis Lev
[21]. This is a method of product quality guaranteed by testing
process that uses FMEA and FTA to investigate the testing
condition of ART/ADT.
FTA is a method to analyze a cause event of faulty
phenomenon based on Fault Tree diagram (FT diagram),
which visualizes the causes of fault events by expanding a fault
event hierarchically to its possible cause events and
representing the entire structure of the problem. FTA is used
aircraft, nuclear power, etc.) and is a general analyzing tool in
problem resolution of products. Miyamura [5] reported that the
knowledge management of design information using FTA is
possible and effective in product design. JATCO also makes
use of FTA in problem resolution of product and design stage
processes: the P·FTA-D·QFD (Perfect · Fault Tree Analysis
- Design · Quality Function Deployment) method [6] (Fig. 1)
which was developed by the power train development division
Method of Computer-Aided FTA (Fault Tree Analysis)
in Reliability Design and Development: Analyzing FTA
Using the Support System in Actual Design Process
2014-01-0747
Published 04/01/2014
Youji Hiraoka and Katsunari Yamamoto
Jatco Ltd.
Tamotsu Murakami
University of Tokyo
Yoshiyuki Furukawa and Hiroyuki Sawada
National Institute of Advanced Industrial Science and Tech.
CITATION: Hiraoka, Y., Yamamoto, K., Murakami, T., Furukawa, Y. et al., "Method of Computer-Aided FTA (Fault Tree
Analysis) in Reliability Design and Development: Analyzing FTA Using the Support System in Actual Design Process," SAE
Technical Paper 2014-01-0747, 2014, doi:10.4271/2014-01-0747.
Copyright © 2014 SAE International
Downloaded from SAE International by Youji Hiraoka, Wednesday, April 02, 2014 11:15:09 PM
of Nissan Motors. Therefore the knowledge management
based on FTA was practically effective in problem resolution of
products and design stage processes.
Fig.1. Design stage processes (the P·FTA-D·QFD method)
company few years ago. About 60% design engineers in
there is individual difference of FT diagrams even if a design
engineer was educated in our programs. Therefore we have
developed FTAid using design knowledge management for
improvement of the quality of creating an FT diagram. Many
methods of support for FTA are studied [7][8], but they do not
effective for our FTA.
We propose a new approach of computerized support for FTA.
We would like to make an accurate FT diagram to use the
approach. We have developed the system of Computer-Aided
FTA (FTAid) for design engineers by the collaborative research
group (JATCO Ltd, The University of Tokyo and National
Institute of Advanced Industrial Science and Technology).
FTAid is a support tool for design engineers to create an
accurate FT diagram of failure event of products. Design
engineers using FTAid will be not only our company's but also
our supplier engineers in the future. Our strategy is to aim for
all design engineers engaged in our products to use a common
platform of this system and to establish the database of many
FT diagrams. It is easy to predict that the working hours should
be decreased with FTAid, and the quality of FT diagram be
improved with it. Therefore we will be able to improve higher
quality of our products.
We reported that the engineer's working hours of creating an
FT diagram decrease by 50 % with the support system in SAE
2012 World Congress [20]. In this paper we report that FT
diagrams with FTAid are analyzed based on individual
education for FTA are studied.
Previous Study
Many computerized design support systems for design
knowledge management which can be used to organize design
data and share the data among the related engineer groups,
are studied. Tamura et al.[2] proposes a design support system
based on the knowledge-structured database established with
Stress-Strength Model (SSM) to reuse failure data as design
knowledge, and they put the system in practical use. Kitamura
et al. [3
knowledge structuring applications. These methods are the
design support systems that require the database to be
Converting enormous amounts of corporate design data into
the notations to establish the database requires a lot of labor
hours and days of work.
Education and Qualification for FTA
and table.1
level(A,B,C), has an education system in each. Engineers have
to acquire that of rank B in our company. That of A rank is an
instructor of FTA. The education system imposes solving a
common practice problem for an FT diagram on rank C, and an
his section.
Table.1. The education and qualification system for FTA
Structure of FT Diagram and Support
Method
Fault tree analysis (FTA) is one of such systematic method in
which the possible causes of fault events can be analyzed and
visualized during a product design and development process in
the form of an FT diagram called a “fault tree”. Even with the
fault analysis of a single component, the scale of FT diagram
frequently becomes large, since it follows through the logicality.
Therefore, the FT diagram of such a actual product is on the
much larger scale, and check and correction of development of
they can be easily handled by design engineers:
• FT diagram of system failure
• FT diagram of component failure
• FT diagram of control failure
Downloaded from SAE International by Youji Hiraoka, Wednesday, April 02, 2014 11:15:09 PM
Fig.2. System failure FT diagrams and support method
of failure which occurs between multiple components and
component failure or a control failure. With an FT diagram of
component failure, the breakage, wear, etc. of a component is
system and an electronic system, and the control logic, its input
(sensor, etc.), etc. are factors of failure. According to the
complicated combination of top events in the FT diagrams of
component failure and control failure, as shown in Fig.2.
In this study, we propose a support method of creating FT
diagrams by the interactive arrows in Fig. 2, because of the
Since the FT diagram of component failure follows the physical
phenomena in many cases, we supports the FT diagram
created based on the quantity dimension indexing method.
Since the engineers majoring the system and control use a
block diagram to analyze the behaviors of system and control,
we support the making FT diagrams based on the block
diagram for the FT diagrams of system failure and control
failure. It is possible in this way to support creating the FT
diagrams of a large-scale system product, by combining these
support methods.
The Method of Support Creating FT
Diagrams
As the method of support creating FT diagrams, we set the
concepts of the system to be developed as follows:
The concepts of the system developed for support creating FT
diagrams.
• The design engineers shall be users.
• The system can point out mistakes in newly creating FT
diagrams.
• The system can show some correction lists of a mistake,
and the correction is easy for users.
• The system has system extensibility in the future (generally
available the system, Internet, etc.).
while the support system is based on the judgment of designer
engineers.
Method of Support Creating an FT Diagram of
Component Failure
Nowadays, the design of an advanced engineering product
includes an increasing number of issues such as product
liability and environmental impact as well as the functionality
and performance of artifacts. Knowledge management is a
promising approach for supporting such complicated design
processes, by which design knowledge and information are
compiled using information technology and used. Although
many knowledge management techniques for literal expression
have been developed, they are not necessarily suitable for
managing the engineering design knowledge of physical
phenomena. For example, a design engineer may store
knowledge about a faulty state of a machine as “… the
temperature of the part is too high …”, whereas another design
engineer may search for relevant knowledge about a failure of
a machine with the description “… the part becomes too hot
…”. The relevance between these two descriptions, however,
cannot be estimated from their literal expressions and a
thesaurus or concept dictionary must be prepared in advance
(e.g.,[13][14][15][16]). On the other hand, a physical quantity
approach to knowledge management has been mentioned in
some studies (e.g.,[17]). If we select the physical quantities
appearing in these two descriptions and represent them using
SI units (the International System of Units) [18], we can
estimate their possible relevance because they have a
common physical quantity: temperature (K). This suggests a
possible advantage of physical quantity representation as an
index for estimating the relevance between design knowledge
descriptions. On the basis of this observation, the authors
propose quantity dimension indexing for design knowledge
management [11].
Quantity Dimension Indexing [9]
In SI Units, all physical quantities describing physical
quantities (and their units): length (m), mass (kg), time (s),
electric current (A), thermodynamic temperature (K), amount of
substance (mol) and luminous intensity (cd). The unit of a
quantity is represented as a seven-dimensional vector in such
a space:
For example, Force [N] = [kg · m/s2] = [m1 · kg1 · s] is
vector a “quantity dimension vector”. Examples of quantity
dimension vectors for various quantities are listed in the
Appendix.
Downloaded from SAE International by Youji Hiraoka, Wednesday, April 02, 2014 11:15:09 PM
Because of the generality, objectivity and universality of SI, this
covers all physical quantities that may appear in any design
knowledge or design problem in the past, present and future,
and the same physical quantities are represented as the same
vectors regardless of the differences between people,
products, domains, organizations, nations and languages. This
is expected to be a promising method of indexing design
knowledge from the viewpoint of physical phenomena.
Application of Quantity Dimension Indexing and
the FT Diagram for Component Failure
The FT diagram for component failure places a faulty event of
component in the top event. Since a faulty event of component
occurs following the physical phenomenon, the events
developed from the top event are mostly expressed by a
physical amount. In the development of an event expressed by
physical amount, unless the quantity dimension of an upper
event is expressed by the quantity dimension of a lower event,
is, the development with a mistake can be said if the quantity
dimension of an upper event can not be expressed by
multiplying or dividing the quantity dimension of an lower event,
or both are not equal (see Fig. 3). Therefore, by using the
quantity dimension indexing for each event of the FT diagram,
it can be judged whether the FT diagram can correctly be
developed, by calculating the quantity dimension vectors of
upper and lower events. Furthermore, since the indexes based
on the quantity dimension vectors of each event makes search
easy on computer, a design engineer can show the candidates
of development patterns from the database and select the
correct pattern.
Fig.3. Example of an FT diagram
Processes of Support Created the FT Diagram
of Component Failure
Fig. 4 shows the computer-aided support processes examined
from the concepts of the FT diagram creation support system.
newly created FT diagram on computer, user checks and
corrects it to determine the quantity dimension indexing of each
event. The system judges the logicality of event developed by
quantity dimension indexing of each event, and displays the
results. At a mistaken point, the system searches for a
development pattern similar to the event of FT diagrams on the
database. It calculates the similarity [9] [10] between the
mistaken event and the quantity dimension of the patterns
registered in the database, and presents the patterns in the
order from the highest similarity to the user. The user selects
the pattern judged to be correct among the presented the
candidates, and corrects the event developed.
Fig.4. Support process creating an FT diagram
We considered to create two databases on the past design
data and the theories. One database contains the patterns of
the FT diagrams created in the past resolving problems
(experimental patterns), and the other database contains the
patterns included in a textbook such as physical formulas and
equations (theoretical patterns). By using a database that
contains the FT diagrams created in the past; the know-hows
of an enterprise, design engineers can easily refer to the
patterns of FT diagrams in the past failures and the past
resolving problems. In addition, the patterns of axioms and
principles in textbooks and literatures can also be referred to,
Furthermore, by accumulating the FT diagrams in a database,
it is possible that the know-hows which FT diagrams are
compiled in a database. This means that we can update a
database of FT diagrams in sequence by registering the FT
diagrams which have been made by actual designing and
Improving the System of FTAid
FTAid has been improved to be useful for design engineers
with continuous studying a process of FTA and results of user
questionnaires.
In the process using the FTAid, there was a work load
increased that a user (design engineer) has to input a quantity
dimension into a computer. We considered that a user would
not use the FTAid if he increased a work load. Therefore, the
system of FTAid have improved that a computer analyzes
Japanese of each events of a FT diagram, and supposes those
candidates of quantity dimensions and display it [22]. Therefore
a user can reduce a work load to input a quantity dimension in
the system of FTAid improved. In the results of the
Downloaded from SAE International by Youji Hiraoka, Wednesday, April 02, 2014 11:15:09 PM
questionnaires in JATCO, there was the most indication of the
convenience of the software. Because a user feels it is hard to
use it if he has worse convenience than the software using
conventionally (Microsoft Excel). We have improved the
system of FTAid more convenient too.
Furthermore, it was predicted that quantity of database of the
measures that a computer reduced the data which searched a
some technical areas, and the user has some areas selected
in a database. The user does not have unnecessary data of
another technology areas and can have minimize the
searching data. This measure allows user not to increase
processing time of searching the candidates of FT diagrams.
Consideration
We considered the following points of effect using FTAid in
actual design process because of the concept of FTAid.
1. FTAid is effective for an expert engineer
2. FTAid in not effective for an engineer who don't understand
engineering and FTA
Our company has education system of FTA, FMEA, and design
engineering, but not all engineers understand everything.
Therefore, we consider that the quality of an FT diagram
depend on personal capability of understanding the
engineering knowledge and FTA.
Result and Future Study
We had more data for the user (design engineers) to create FT
diagrams with FTAid compared to it in 2012. Fig. 5 shows the
relation of working hours creating an FT diagram and quality
level of FT diagrams evaluated by domestic standard. This
created an FT diagram and the quality of it which has used
FTAid in actual design process. Particularly, FTAid allow user
to decrease the working hour of creating an FT diagram
greatly. However there are some FT diagrams less than the
point of an FT diagram is necessary to be valid for design
process and resolving a problem.
Fig.5. Relationship of accuracy and working hours
Fig. 6 shows the result that we have analyzed the working
hours divided into two personal ranks of FTA. Chief engineers
of rank B and engineers of rank C decrease working hours of
creating an FT diagram with FTAid by more than 50% (target
value). Particularly the effect of engineers of C is larger than
our prediction. As for this, some engineers of rank C create FT
diagrams in a short time and they judged enough. Therefore,
we consider that the FT diagrams of them are not accurate. In
the actual process of creating FT diagram, it takes more time
that chief engineer improve it with engineer.
Fig.6. Results of working hours of FT diagrams divided into two
personal ranks
Fig.7 shows the result that we analyzed the evaluated point of
FT diagram divided into two personal ranks of FTA. The effect
of the evaluated point of an FT diagrams by rank B with FTAid
is +23% on the average, but the effect of it by rank C with
FTAid is +12% and is much lower than B rank. We consider
that some engineers of rank C judge that the FT diagrams are
right even if they have little improvement of it in short time.
Therefore the effect of the evaluated point of some FT
diagrams by rank C is lower quality.
Fig.7. Evaluated results of FT diagrams divided into two personal ranks
Accordingly, the FTAid is effective for chief engineers of rank B,
but is few effective for engineers of rank C. Furthermore, we
analyze below this problem of FT diagrams created by
engineers of rankC. We have investigated some FT diagrams
of lower 60 point (our standard level) with FTAid. Fig.8 shows
the points of FT diagrams which were deducted. The top three
causes are
• Failure modes were not expressed correctly
• Events developed from failure mode were not correct
• Events of another precondition for design were not
expressed
Downloaded from SAE International by Youji Hiraoka, Wednesday, April 02, 2014 11:15:09 PM
Fig.8. The points of FT diagrams deducted
Table.2. The index of Fig. 8
In order to create FT diagrams correctly and effectively, the
following measures are studied in these results.
• Increasing correct failure modes in the database
• Increasing FT diagrams of some technical area
• Improving the lecture for FTA and the textbook
• Improving the database system for FTAid
First is a measure that the FTAid supports failure modes by
searching a Japanese word. As FTAid also have the software
of searching Japanese word, it is necessary that the database
of failure modes have to be increased according to engineering
knowledge and our know-hows. Secondly FT diagrams of
some technical area have to be increased more and more. This
measure is for engineers to search correct expanding events of
failure mode using FTAid. However few inexperienced
engineers do not know how to use a failure mode in an FT
diagram. It is necessary that we improve our education for
inexperienced engineers. Finally, it is necessary to improve the
database system, because an FT diagram of the database will
be increased accumulatively. Our database system will be
changed to relational database system to prevent increase of
the processing time used FTAid.
Conclusions
diagrams in actual design process.
engineers but is uneffective for engineers not to understand
engineering and FTA.
3. We have found that the causes of lower point of FT
diagrams were three following.
Failure modes were not expressed correctly.
Events developed from failure mode were not correct.
Events of another precondition for design were not
expressed.
4. We have studied to take the following measure of FTAid and
domestic education.
Increasing correct failure modes in the database
Improving the lecture for FTA and the textbook
Improving the database system for FTAid
References
1. Suh. Nam P., “Axiomatic Design”, Oxford University Press,
2001.
2. Tamura, Y., etc., “A Study on the Method to Manage Design
Knowledge on Failures - Construction of the Knowledge
Structure of a Causal Chain of Failures -”, The Journal of
Quality, Vol.32 (1), 2002, pp.122-135.
3. Kitamura, Y. and Mizoguchi, R., “An Ontology of Faults
-Articulation and Organization-”, The Journal of JSAI,
Vol.14, No.5, 1999, pp.828-837.
4. Shiomi, H., Shimaoka, J. and Ishiyama, K., “Utilization of
FMEA · FTA”,JUSE Press, Ltd., 1983
5. Miyamura, T., “Effective Use of FTA to Join Design
Technology Information and Make it Organization Reusable
Knowledge”, The Journal of Quality, Vol.27 (2), 1997,
pp.69-77.
6. Ooki, Y., Murakami, Y., etc., “P·FTA-D·QFD method for the
short development process and high quality production”,
Nissan Technical Review, No.44, 1999, pp.50-55.
7. Ooyama, A, etc., “Introduction of IT Technology into D&D
Process of JAXA's Rocket Engine Valves foFr Reliability
Improvement”, Proc. Of Design & Systems Conference '07
(JSME), 2007, pp.201-205.
8. Yoshimua, T. and Tada, N., “Function and Feature of FTA
System at Toyota Motor Corporation”, The 9th Symposium
of Application Case with JUSE-QCAS.
9. Murakami, T., Kikuchi, Y. and Hiraoka, Y., “Quantity
Dimension Indexing for Design Knowledge Management”,
Proc. Of ASME 2008 IDEC&CIEC Conference (DETC2008-
49737) (CD-ROM), (2008), 1.
10. Murakami, T. and Kikuchi, Y., “Knowledge Management
for Fault Tree Analysis Based on Quantity Dimension
Indexing”, Journal of JSME C ed., VoL.75, No.756(2009),
pp.2171-2180.
11. Hiraoka, Y., Murakami, T., etc., “Development of the support
tool for fault tree analysis (FTA)-Application of knowledge
management based on Quantity Dimension Indexing-”, The
39th Symposium on Reliability and Maintainability, 2009
12. Ichikawa, T., “Thinking (Cognitive Psychology 4)”,
University of Tokyo Press, 1996
13. Szykman, S., Sriram, R. D., Bochenek, C., Racz, J. W. and
Senfaute, J., “Design Repositories: Engineering Design's
New Knowledge Base”, IEEE Intelligent Systems, 15(3),
2000, pp.48-55.
14. Hirtz, J., Stone, R. B., McAdams, D. A., Szykman, S. and
Wood, K. L., “A Functional Basis for Engineering Design:
Reconciling and Evolving Previous Efforts”, Research in
Engineering Design, 13(2), 2002, pp.65-82.
Downloaded from SAE International by Youji Hiraoka, Wednesday, April 02, 2014 11:15:09 PM
15. Kitamura, Y., Kashiwase, M., Fuse, M. and Mizoguchi, R.,
“Deployment of an Ontological Framework of Functional
Design Knowledge”, Advanced Engineering Informatics,
18(2), 2004, pp.115-127.
16. Ahmed, S., “Encouraging Reuse of Design Knowledge: A
Method to Index Knowledge”, Design Studies, 26(6), 2005,
pp.565-592.
17. Gruber, T. R. and Olsen, G. R., “An Ontology for
Engineering Mathematics,” Fourth International Conference
on Principles of Knowledge Representation and Reasoning,
Bonn, 1994 pp.1-18.
18. Bureau International des Poids et Mesures, SI brochure,
8th ed., 2006 pp.94-180.
19. Sawada, H., Matsuki, N., Tokunaga, H., Furukawa, Y.,“A
Manufacturing Software Development and Operation
Framework ‘MZ Platform’ and its applications in Industry”,
Proc of Advanced Engineering Design, 0093.pdf (CD-
ROM), 2004.
20. Hiraoka. Y., Yamamoto. K., Murakami. T., Furukawa. Y.,
Tokunaga. H., and Sawada. H., “Method of Computer-
Aided FTA in Reliability Design and Development -Using
Knowledge Management based on Quantity Dimension
Indexing and Block Diagrams -”,SAE Technical Paper
2012-01-0203, Journal of Commercial Vehicles, 5(1), 2012,
pp.72-82
21. Klatis. Lev M., “Accelerated Quality and Reliability
Solutions”, ELSEVIER, 2006.
22. T., Furukawa. Y., Tokunaga. H., and Sawada. H.,
“Development of the support software created an FT
diagram (FTAid)”, 22th Design engineering & System
section Conference of JSME, 2012.
Contact Information
Youji. Hiraoka (JATCO Ltd.)
youji.hiraoka@nifty.com
The Engineering Meetings Board has approved this paper for publication. It has successfully completed SAE’s peer review process under the supervision of the session
organizer. The process requires a minimum of three (3) reviews by industry experts.
All rights reserved. No part of this publication may be reproduced, stored in a retrieval system, or transmitted, in any form or by any means, electronic, mechanical,
photocopying, recording, or otherwise, without the prior written permission of SAE International.
Positions and opinions advanced in this paper are those of the author(s) and not necessarily those of SAE International. The author is solely responsible for the content of the
paper.
ISSN 0148-7191
http://papers.sae.org/2014-01-0747
Downloaded from SAE International by Youji Hiraoka, Wednesday, April 02, 2014 11:15:09 PM