E-Mail : takaro_u@yahoo.co.jp → Quick Access

Contents

Click the chapter or step title in the table below for a glance.
Glossary of Key Terms

Introduction
(1) The Mystery about the Failure Mode
(2) Infants Cut Fingers with Shredders
(3) The True Cause of the Derailment
(4) The History of FMEA
(5) The Defects of the Conventional FMEA

Reference

Chapter 1 : What Is the FMEA
1-1 The Relevant Concepts
1-2 Outline of the Design
1-3 Need of the FMEA Execution
1-4 The Basic Requirements of FMEA

Chapter 2 : The Purpose of FMEA
2-1 The Purpose of the FMEA

Chapter 3: The FMEA Process
3-1 Outline
3-2 Preparing Documents
3-3 Listing Failure Modes
3-4 Listing Current Controls
3-5 Listing Effects of Failure Mode
3-6 Evaluating from the 3 Viewpoints
3-7 Calculating Risk Index

Chapter 4: Who Executes an FMEA?
4-1 A Tool for Design Aid

Chapter 5: The Product and Process
5-1 Designing the Product and Process
5-2 Product FMEA
5-3 Process FMEA

Chapter 6: The Case Study
Step 1 : Preparation of Documents
6-1-1 In the Case of Product FMEA
6-1-2 Reliability Design
6-1-3 In the Case of Process FMEA

Step 2 : Listing Failure Modes
6-2-1 Defining Failure Mode
6-2-2 Failure Mode of Product
6-2-3 Failure Modes of Process

Step 3 : Listing Effects and Factors
6-3-1 Introduction
6-3-2 Case Study of Product FMEA
6-3-3 Meaning of Effect
6-3-4 Meaning of Factor
6-3-5 Examples

Step4 : Listing Current Controls
6-4-1 Meaning and Purpose
6-4-2 Process of Reliability Design
6-4-3 Case Study of Product FMEA
6-4-4 Case Study of Tablet Bottle Cap
6-4-5 The Purpose of Describing "Current Controls"

Step5 : Evaluating Three Elements
6-5-1 General Criteria
6-5-2 Evaluating Measures for Severity
6-5-3 Evaluating Measures for Occurrence
6-5-4 Evaluating Measures for Detection

Step6 : Application of Risk Index
6-6-1 Calculating Risk Index

Step7 : Optimization of RI
6-7-1 Optimization for Product FMEA

Appendix
Appendix 1 : Factor and Effect Diagram
Appendix 2 : Case of Gauze Collection
(a medical case of FMEA)
Appendix 3 : Case of Gauze Collection Process
Appendix 4 : Adding Risk Index




Glossary of Key Terms


Bill of Material (BOM) - a complete list of a product's components, subassemblies and assemblies that comprise a material product (A structural bill of material consists of a detailed list of all the components combined with their quantities, and information as to their relation to one onother.)

Cause - the event that has created a failure mode in a particular element of a product or process (Note the difference between Cause and Fctor. Cause is a cause-and-effect relationship that is going on past or present, Factor refers to the causal relationship that may occur in the future.)

Current Controls - every systematic means to prevent failure, to ease the influence of failure, or to detect failure or failure mode (This word may be replaced with "Reliability Design".)

Customer - any person or/and process that utilizes the product along the chain of procession and may be adversely sffecterd by a product faiure

Detection - degree of insufficiency of the measures taken to ease identifying failure cause, its consequent failure mode or failure prior to delivery to the customer (any follow-on user)

DFMEA - the acronym of Design FMEA. It should be correctly called "Product FMEA"

Effect - each worst impact of each failure mode which is worth evaluation.

Factor - the evet that may create a failure mode in an element of a product or process (This word may be replaced with "Generating Mecahanism".)

Failure - malfuncton of a product or process.

Failure Mode - structural breakage that may occur in an element of a product or process (Structural breakage can occcur in a part or connection. For example, tthe disengagement of the mating plug, the screw loosening, the peeling of the adhesive material and the invasion of the obstacles are connection failure mode. As for a process, a breach against an instruction designated in the process design corresponds to a failure mode.)

FMEA - see "What is FMEA"

FMECA - the acronym of Failure Mode, Effect and Criticality Analisys (a type of FMEA that emphasizes the nessessity of measures to take to Severity. In most cases, as measures against high Severity rating, you can adopt any of redundant design, mistake-proof device or fail-safe device. The aircraft equipped with two to four engines is an example of redundant design.)

Function - any intended purpose of a product or process



Occurrence - degree of insufficiency of the measures taken to reduce the frequency of the failure.

PFMEA - the acronym of Process FMEA.

Poka Yoke - Japanese for "error proof" - A quality improvement strategy emphasizing preventing human errors at their source by making changes so that nobody could make a mistake

QC Process Table - a chart that describes the process design (QC Process Table shows how to control and assure the quality of output from the process. This is comprised of two parts. The left side is so-called "procedure for creating quality" and the other is so-called "procedure for quality assurance.")

QS 9000 FMEA - a conventional type of FMEA that has been developed by GM, Ford and Chrysler(The recall number of the car which rose in 2009 was as below).
 - Toyota: 4,870,000
 - Ford: 4,520,000
 - GM: 2,230,000
At the beginning of 2010, Toyota and GM added millions of collection. You should immediately abolish the conventional Relative FMEA which caused such a result.
Rlative Evaluation FMEA - a type of wrong FMEA analyzing and ranking the risks associated with various producut failure modes in order to prioritize the sequence of corrective action (The correct type is the Absolute Evaluation FMEA.)

Risk Index (RI) - the index which is calculated from numerical S, O, and D, and shows the risk of failure mode synthetically: RI = (S x O x D)1/3 (This is used in the Absolute Evaluation FMEA.)

Risk Priority Number (RPN) - the mathematical product or the numerical Severity, Occurrence and Detection ratings: RPN = S x O x D (This is used in the Relative FMEA, which is wrong and impossible to conduct appropriately - see Risk Index.)

Root Cause Analysis - the process of repeatedly asking why a specific cause occurred until a fundamental defect in the management system is disclosed (This is conducted in order to prevent recurring of a certain trouble.)  Note: Don't ask why a specific problem occurred. It is nothing more than a cause investigation and is not related with recurrence prevention. After the cause clarifies, ask what is the managerial system defect which resulted it. This is the recurrence prevention.



Severity - degree of insufficiency of the measures taken to ease the impact of a certain failure mode (Note: "Severity" does not mean the degree of the impact.)

What is FMEA - FMEA is the acronym of Failure Mode and Effect Analysis. Those reading this word for the first time may think this to be the name of an technology that they have not experienced.
  However, without knowing the language of FMEA, skilled designers are implementing an FMEA substantially unconsciously. The expert designer studies each screw for its best size, materials, heat-treatment, surface treatment, to prevent the damages that may occur during use by the customer,and confirms them. Assuming planned measures, the designer should evaluate the severity of effect, the likelyhood of occurrence and the advance detection of the failure.
  This corresponds to an FMEA aside from the form substantially. However, you carried out this activity in your head conventionally. You did not recorded it according to a constant style, and a third party was not able to inspect it.
   FMEA is an activity to express the above-mentioned reliability evaluation or a reliability design to a worksheet according to a constant style.
As shown above, FMEA is a tool for evaluating relability (i.e., property not to break down) of product or process.
  In contrast, the conventional Relative Evaluaton FMEA insists that an FMEA is a systematic method of identifying and preventing product and process problems before they occur. Namely, they say, an FMEA is conducted not only for evaluating reliability but also for evaluating other possible risks. As a result, difference among a failure, a failure mode and a manufacturing defective, have become vague.
  The point which is the hardest to understand especially is that a manufacturing defective belongs to failure mode. However, FMEA is conducted to evaluate reliability of a design. Do you design a product includeing a manufacturing defective?

Purpose - An FMEA is executed:
  1. to predict all failures and disasters that may occur by use of the product or by running the process, and
  2. to assess the inadequacy of measures to prevent each harmful failure mode.


Since preventing the unexpected failure is the purpose of FMEA, you should adopt the Absolute Evaluation FMEA. It is because it starts from each breakage of each component and connection in the bottom-up way and never creates missing failures.
  In contrast, the conventional Relative Evaluation FMEA starts from the failure on the top and ask how it can occur. This means that you may ommit enumerating some failures which did not come up in mind by chance, and these ommitted failures could become the unexpected failures.

Science all failures should be predicted with FMEA and approppriate measures should be taken to every failure modes, there must not exist omission or priority in the failure modes.
  In contrast, the conventional Relative Evaluation FMEA rates the priority of failure modes and take measures only against the priority failure modes and unimportant failure and disaster are disregarded. This is not approved by our society.
  It is necessary to become clear whether sufficient measures have been already taken against each problem. Moreover, the designer has to evaluate the reliability of his own design during the design process.
The kind of FMEA suitable for this purpose is the Absolute Evaluation FMEA, whitch leads to the most suitable reliability. And you can learn it at this home page.
  In contrast, the conventional Relative Evaluation FMEA insists that the FMEA team which consists of people who were not taking charge of the design should performs FMEA. If so, by what kind of method should the designer evaluate reliability in the design process?

The image below shows that you should start from each failure mode in bottom-up way as in the Absolute Evaluation FMEA, and should not start from failures in top-down ways as in the Relative Evaluation FMEA.
  In contrast, in the case of the mistaken method, you will ask the cause of each failure and answer with three categories of cause;
 - the structural breakage of each part or connnection,
 - the faiure (=malfunction) of each sab-function, and
 - the manufacturing defective.

However, only structural breakage is the right failure mode. The figure below shows the relation between these concepts mentioned above.



Failure and Failufre Mode
Failure means malfunction, which must not be confused with failure mode. A failure mode means a structural breakage of an element consisting a product or process. A function is realised through many elements, and each element can have several manners of breakage (i.e., failure modes). Therefore, each malfunction can be caused by many failure modes. In other words, any failure is a terminal of the analysis and any failure mode is the starting point. You have to enumerate all failure modes and you can reach all faiures without missing.
  In contrast, in the conventional Relative FMEA, they enumerate failures first and ask why they occurs, and the answers are the failure modes. Then, structural breakages of parts, failures of subfunctions and manufacturing defectives will be enumareted. However, you enumerate what come up by chance and there will be left some enumeration missing in failures and factors as well.

Failure Mode
Each breakage that occures in the structure of each system element is called a failure mode.
 - As for a product, each breakage of each part or interface correspond to a failure mode.
 - As for a process, each breach against each instruction corresponds to a failure mode.

Any manufacturing defective cannot be a failure mode.It is because:
 - there are no manufactureing defectives in the design as for an product FMEA.
 - any manufactureing defective is the results of a malfunction of the manufacturing process, and any manufacturing defective should be regarded as an effect of a certain failure mode included in the failed process.

Why, do you think, manufacturing defectives are not failure modes? It is because FMEA evaluates only the design and the design does not include defectives.
  In contrast, the conventional Relative Evaluation FMEA insists that:
 - each manufacturing defective can be a failure mode of a product, since it causes a failure.
 - each manufacturing defective can be a failure mode of a process, since the process should not produce it.
This is a result of the mistaken failure mode concept.


Evaluation
In the reliabiity design, measures are taken to each failure mode. And you have to confirm the sufficiency. For the pirpose of confimation, you have to evaluate the insufficiency of the measures taken to each failure mode from the three viewpoints, i.e., Severity (S), Occurrence (O) and Dtection (D).
 - Severity(S) of effect that the customer (any follow-on user) might receive.
 - Occurrence(O): the frequency of the failure,
 - Detection(D): difficulty of the sign of failure before reaching the customer.

Evaluation point of these three will be integrated into a single index, Risk Index (RI). You must determine the degree of insufficiency of reliability measures in accordance with the index. This decision will indicate the insufficiency of the measures for each failure mode. Optimization of the design by evaluation on an absolute scale is the only useful method.
  In contrast, the conventional Relative Evaluation FMEA using RPN determines the priority among the failure modes listed. And the boundary line which classifies failure modes with an insufficient measure and failure modes with a sufficient measure does not exist.
Moreover, judgment of priority is very difficult and is unrelated to judgment of yes-no, it is not helpful after all. You have to perform independently the yes-no decision of the measure for each failure mode.

Process FMEA
What is the structure of a process?  In a product design, you define a required function and design structure to achieve the function. You design structure and it achieves the function. You cannot achieve the function directly. All the designed contents are structures. The above is completely the same also about a process.

Failure modes of a manufacturing process are breaches of instructions shown in the QC process table. You should abolish the old type of FMEA in which manufacturing defectives or other accidents are treated as failure modes.

→ Quick Accsess to FMEA


Introduction


(1) A Mystery about the Failure Mode

My First Encounter with the FMEA
When I was young, in 1960s, working as a novice engineer for a machine manufacturing company in Japan, I happened to have an opportunity to read a reference book of the FMEA with a commentary about the meaning of failures mode and failures.
Look at the following table, which is quoted from a certain book.

FMEA Worksheet for an Extinguisher
Line Component and function Potential Failure Mode Potential Effect of Failure
1 Hose;delivers extinguishing agent Craks  
       
7 Charge gauge;determine remaining volume of agent Inacurate reading  
       

A Problem Was Found
I could not reach a complete acceptance, however, at the inconsistency that "Inaccurate reading" is a functional trouble, i.e., a failure, whereas "Cracks" belongs to structural breakage. Even a reference book written in Japanese, my mother tongue, did not improve my sense of satisfaction.
The Harder Study, the More Problems
The portion of a FMEA worksheet shown below has been quoted from Professor (H. Makabe p.134), who had been one of the leaders of this field. The scope of the product covers "a brake subsystem" including a pedal assembly and a pressure oil pipe assembly.

[Note] Blue color and yelow boxes are added by us.

  1. "Can't be pushed down" and "Doesn't return" are functional troubles, which belong to failure. Why are they listed in the spaces for failure modes?
  2. "A pedal crank broken" must be a failure mode, however, it is listed in the column of cause. The same applies to the remaining two.
  3. "Joint separations" and "cracks" are acceptable as failure modes, however, this conflicts with the contents in the place of (2).
  4. It is wrong to consider "an incomplete weld" to be a failure mode, because it occurs prior to handing the product to the customer. And it is a problem which you should handle in a process FMEA. Failure modes occur after the product is handed to the customer, whereas manufacturing defectives take place beforehand.
  5. Three causes are listed here, but the grounds that there cannot be additional causes are not clear. Aren't these just what came in mind by chance? If you want to list all causes without missing, you should enumerate all stuructural breakages of all parts and interfaces beforehand and select related causes from among these. In other words, you should take a bottom-up approach.
  6. They place the end product with a system and perform a top-down approach to subsystems, components and parts downward from there. By the way, where, how can you list the misuse (very important failure mode) of the customer?
    This is one of the problems caused by the top-down approach. You cannot list the misuse unless you take the bottom-up approach.

A New Discovery after 40 Years
Approximately 40 years passed and I solved the problem.I have noticed that there are two methods to reach the failure modes.
  One is the top-down method: you start at the failure and go down to the causes. When the top-down approach is adopted about a certain failure, the following matters are certain to be enumerated as failure modes without distinction. The other is the bottom-up method: you start from the damages on components and go up to the results(i.e. effects). In the case of bottom-up approach, the failure mode concept is comprised of only structural breakage of components or interfaces, and never includes dysfunctions nor manufacturing defectives.
  The failure mode concept depends basically on which of these two is chosen and the bottom-up search is the right choice.

(2) Finger Cutting Accidents of Infant with Shredders

The Outline of The Accident
In 2006 Japan, the accidents that infants played with shredders and cut their fingers often occurred. I wondered why these could not have been prevented through execution of FMEAs.
  Then, the following thought occurred to me. The shredders were not designed to include infant's fingers, and hence no finger cut accidents could be anticipated however deep traditional FMEAs were executed.

Defect of Conventional FMEA
It is a decisive defect of conventional FMEA. How can you necessarily detect the failure mode that an infant is inserting its fingers into a shredder in the top-down way?
  It is merely each potential failure of a shredder as the end products that can be a starting point of analysis as far as top-down FMEA is carried out. And this accident cannot be foreseen even no matter how deep you examine it with in a shredder by the top-down method.
  In order to handle the misuse that a small child inserts a finger in a shredder, breaches of insruction of user's manual should be listed, and a kind of fool-proof measures should be taken to prevent serious results.
Toward Solving
In contrast, if an FMEA is executed in a bottom-up way, failure modes, i.e., breakage in parts and interfaces become the starting points to look for failures and disasters, and the search goes up to components, sub-systems, and the final product.
  Furthermore, the bottom-up search will go up to the breaches of the instruction manual for the user concerning the use and maintenance. Here everyone will reach the following understanding.

(3) Causes of Derailment Accidents

The concept of "cause in terms of management" has to be strictly defined and identified.

Another Serious Accident
On April 25, 2005, near Osaka, Japan, an overturn derailment occurred at a sharp curve in the railway of the West Japan Railway Co. It was the accident that resulted more than 100 dead people, more than 500 injured people and compensation amount of money of more than 1 billion dollars.


An Unacceptable Description of the Cause
Then the company, after setting up a speed control system of one billion dollars, has resumed the commercial operation. The cause, the government official reported, was that the driver had exceeded the speed limit. However, I was not able to accept the following railroad managements.
Measures must Be Economic and Effective
It is very hard in general, however, to expect a prior enforcement of measures on a small local line at the expense of one billion dollars. The point is to develop low-cost measures.Probably, such an accident might have been prevented if the guide fence (a fail-safe system) along the exterior of the curve by the cost of 250,000 dollars had been adopted.

Ignorance of the Economic and Effective Measures
The true cause of this accident from a viewpoint of a managerial system was that this cheap and effective means was not performed. The Japanese authorities prosecuted the presidents of the company of three generations on the suspicion of having neglected construction of 100 million dollars, but this public prosecutor's irrelevance action is remarkable.
As mentioned above, designing the process of "use and maintenance" of the railway system and its FMEA are necessary for attending to the failure mode of overturn, leading to economic and effective measures.

Switzerland Showed a Similarly Absurd Settlement
On July 23, 2010, a derailment overturn happened to the rear three cars of the tourist train "Glacier Express" in the Swiss Alps, and a lot of Japanese tourist were killed or injured. The cause, the government official reported, was that the driver had accelerated before the tail vehicle passed the curve. Then the company has resumed the commercial operation with no measures against recurrence. And the guide rail was not installed at the curve. Swiss railway officials has learned nothing from the very similar accident in Japan five years before.


Managerial Cause
This is rooted in the wrong concept of the cause. We are not searching the natural cause but the managerial cause (defect). The answer to "what kind of defect of the current controls is there?" is the managerial cause. Thus the cause of the derailment havs proved to be the absence of a fail-safe guide rail.

Fail-safe Measures Is the Key Technology
No effective measures are provided even if they investigate the responsibility of the driver, and who is responsible becomes clear.
The fact means that the negligence of the driver is not the cause of the accident from the viewpoint of management.The cause is that measures were not carried out.
  The low cost failsafe measures to prevent a serious result should be carried out as first priority. This is because the absolute method to preventing derailment is not obtained yet. And derailment in particular is nonavoidable at a curve yet. It should be thought that the true cause of the specific accident is in the negligence of enforcement of the fail-safe means of budget prices.

(4) History of FMEA

Beginning in the Aerospace Industry
In the aerospace industry of the mid-1960s, the safety problem was focused on in particular, and the FMEA was a key tool to be performed formally for the first time for increasing the safety of the chemical plant. The engineer always looked for a possible trouble beforehand and analyzed a product and a process. On the other hand, it is said in general, the procedure of the FMEA was standardized, and common language spoken among the employee of various companies and all levels was established.

Improved by U.S. Auto Industry
The FMFA was improved by the U.S. auto industry to meet a reliability improvement tool. In 1996, the U.S. auto industry constructed QS-9000 supplier requirements.
International Automotive Task Force (IATF) added other international quality standards to QS-9000 and ISO/TS 16949 was developed. The ISO/TS 16949 standard requires that suppliers to the automotive industry conduct product and process FMEAs to prevent failure.

Just a Typical Propaganda
Generally it was explained as follows conventionally.Unlike a many other quality improvements technology, FMEA does not need complicated statistics mathematics. Moreover, the remarkable business advantage which reduces the quality problem resulting from a process design or a products design is expectable.
  However, the above-mentioned reference is only mere typical advertisement.

(5) Defects of Conventional FMEA

Merit Is Extremely Poorer than Officially Publicized
FMEA technology contains many problems in spite of the shining history. In order to carry out FMEA, a certain amount of resources are required. However, the conventional FMEA wastes time and human resources too much. Furthermore, the strong point is very poorer than having been announced officially.
  This is caused by the mistakes of content instructed by book authors, workshop trainers and colleges mainly.

Robin E. McDermott p.4 introduces itself saying;
"This booklet was designed to help shorten the learning curve for FMEA teams and to help them conduct effective and efficient FMEA even if it is their very first FMEA."

Commentary:
This is not true. The people who learned the outdated FMEA wear the front as the end that it succeeded. The main errors of outdated FMEA are those points from following (a) to (f).
 (a) Designer's Duty
It is said that FMEA is activity of the teams consisting of the members of the experts of the various quarters that are not in charge of the design.

However, it is necessary for the designer to check reliability by carrying out FMEA while pushing forward the design. How do you show reliability if the designer must not carry out FMEA?
  This is similar with the custom that the carpenter measures the size of wood. The FMEA is a process to dentify the reliability of the designed product or process. However, this expertise-like duty is not so easy as a makeshift team can perform successfully.
 (b) Failure and Failure Mode
About the difference between a failure and a failure mode, the explanation is uncertain, and hence many engineers lose their ways in the desert. Although an FMEA neede to be performed in bottom-up way from parts up to a completed product, the wrong theory goes conversely. This mistake produces an uncertain concept of failure mode.
 (c) RPN Is an Inappropriate Index
It is useless to show the risk priority numbers (RPNs), because lack of measures for each failure mode cannot be identified. A risk evaluation should be the answer to the question whether the measures are short or enough against a specific failure mode. Otherwise, measures against each failure mode might run short or excessive.
 (d) Ten-Point Ranking
The conventional FMEA uses the RPN method, and a lot of chairs are needed for a queue in order of priority of failure modes, and the ten-point scoring is helpful for this purpose, as follows;

Severity × Occurrence × Detection = 10 × 10 × 10 = 1,000


Thus one thousand chairs are prepared. On the other hand, accurate evaluations of severity, occurrence and detection are quite difficult and cannot avoid an error of 1 or 2 points at least for each evaluation. This results in an error of around 100 points in terms of RPN, thus a practical use becomes hard.

(e) Business Retention
The order of priority is not clear unless the overall evaluation is over. Thereby even a good idea is forced to hold. In addition, even when the priority in the whole is determined, the shortage of measures for each failure mode is not clarified; hence the delay of work becomes hard to accept.
(f) Failure Mode Concept Is not Clear
Should we look for defects of an actual product or process? Or should we look for those of a design? The misunderstanding of this point produces a funny idea called an FMEA of "a de facto process" which has not been designed formally in the field of medical FMEA in particular.

This mistake is due to the fact that the failure mode concept of process is not clear. This homepage touches the serious problems of the conventional FMEA, showing case examples. And all mistakes of the conventional FMEA can be reviewed in this homepage.

Reference
Authors Title Publisher
(Language)
Year
Robin E. McDermott
Raymond J. Mikulak
Michael R. Beauregard
THE BASICS OF FMEA
(2nd Edition)
CRC Press
(English)
2009
Kenneth W. Dailey The FMEA Pocket Handbook D W Publishing Co.
(English)
2004
Hitoshi Kume Quality Management of Design and Development Union of Japanese Scientists And Engineers
(Japanese)
1999
D.H. Stamatis Failure Mode And Effect Analysis ASQ Quality Press
(English)
1995
Hajime Makabe The Basic of Reliability Engineering Japanese Standards Association
(Japanese)
1995





Chapter1: What Is FMEA


1-1 The Relevant Concepts

Not Easy to Foretell Failures of New Designs
About a new product or a newly-designed process, it is not easy to foretell what kind of function accident and disaster may happen. To that end, what kind of method is thought about? You cannot enumerate all failures and disasters if you merely look at specifications or the actual products.

The Two Typical Methods
There are two representative ways of thinking as follows.

  1. In the top-down method, all the functions are developed from the end products, each failure or a disaster is drawn, and the routes in which each failure or a disaster occurs are asked. This has supported the alignic FMEA coming from QS 9000 and has led to the wrong procedure.
  2. In the bottom-up method, we enumerate all breakages of the minimum structural elements. For example,
    • such part breakages as rust of screws, broken electric wires, deformed pins, worn gears, etc., and
    • such interface breakages as blockage, slack, separation, disengagement (of floor mat of a car), etc.
    and we ask what failure or disaster can be caused by each of these breakages.
The reason why functions or malfunction should not be the starting points is because failures beyond prediction cannot be the starting points.
  The reason why functions or malfunction should not be the starting points is because failures beyond prediction cannot be the starting points. You cannot but recognize it if you watch the number of recalls that Toyota and GM have caused in the past.
  This homepage adopts the latter contrary to QS 9000. As a result, various difficulties caused by the past method of the top-down approach are overcome.

Definition of Reliability
The quality of a product or a process that can maintain the intended functions under specific conditions and within a prescribed period is called reliability. Here,
What Is Failure?
A functional problem of a product or a process is called a failure.

As for a product, after the use of a certain time, it may be caused by damaging parts or interfaces (connections) in the product. As for a process, breaches of instruction cause failures producing output troubles.

Distinction between Failures and Manufacturing Defectives
On the other hand, a malfunction of a product before use is not a failure but a manufacturing defective (including an assembly defective). Similarly, a failure of a process may happen later because of a breach of an instruction of the process specification or the equipment maintenance even if the process works normally at first.

In 2005, several employees of a certain maker joined in our align. We answered their question by an E-mail for several times after the align, but were not understood. We visited them in their factory at last and had a talk with them for explanation.

The problem was that the automatic machine which they had designed and produced often had stopped during use. We asked how many months it had taken from the installation to the outbreak of the phenomenon, and they answered that it had begun at the beginning of use. The existing trouble at the beginning of use is usually a defect of the production or the design, not trouble by the use. In other words, this is not a theme treated in FMEA.

As for a Process
The reference above applies also to a process. If the design of function of the process is poor, it will produce various failures even if there is no breach of the instructions. Moreover, this is not a problem to be treated in FMEA.

Definition of Failure Mode
Manners in which a product or process may fail are called failure modes, which should be identified by a bottom-up approach.

In general, a breakage of a part or an interface is a failure mode of the product. And a violation of an instruction of a process is a failure mode of the process.
  This technical term usually is easily misunderstood because of the careless definition that has existed for a long time. This definition does not always lead to the bottom-up approach and it is inclined to produce the wrong way, the top-down approach, in which they ask why the failure occurs.
  The typical wrong definition is as follows:
"Ways in which a product or process can fail are called failure modes."

As a result, malfunctions, manufacturing defectives and failure modes are listed in the failure mode description column without distinction.
The Right Definition
The Failure mode concept should be definede as follows.
"What are the ways a product or process can fail, and are identified bottom-up are called failure modes."

Even when a failure occurs in a complicated product, the product is accompanied by breakages of some kind of parts or connections by all means. And failures or disasters can never be prevented unless these are prevented, and hence, only the failure mode of that meaning are worthy targets of FMEA.

Usefulness of the Failure Mode Concept
It is not difficult for a skilled designer to enumerate all failure modes of parts and interfaces of his own design. For example, rust, deformation, form collapse, etc. are immediately visualized as failure mode of the item called a screw.

Moreover, a crack, a jam, comming-off, etc. are mentioned as failure modes of a hose.
  A skilled designer knows what changes parts and connections may receive after the product is handed to the customer and used for long.
  All failures can be trodden by enumerating all these changes of parts and connections, and making them starting points of search for malfunctions and disasters.
  As for failure modes of a process, they are easily enumerated as breaches of instrutions.
Starting Points
After all, the starting points (failure modes) of FMEA are always decided as the breakage of the smallest structural element of the system as follows. And neither malfunctions nor manufacturing defectives are included.
It is hard even for experts to list up potential failures of a complicated product directly without overlooking.
  However, it is not difficult for experts to list up changes (failure modes) that may occur in parts and interfaces while being used long.

What Does Top-down Approach Bring?
In the case of a complicated product, a lot of subordinate functions exist in the product.
  If a top-down approach is taken, both the theory and the practice will certainly fail, because all of failure modes, failures and manufacturing defectives are mixed and confused.

Objection Theory

Kenneth W. Dailey p.6:
"What is FMEA?
Definition: A systematic method of analyzing and ranking the risks associated with various product failure modes (both existing and potential), prioritizing them for remedial action, acting on the highest ranked items, reevaluating those items, and returning to the prioritization step in a continuous loop until marginal returns set in."
Commentary (What is wrong?):
Robin E. McDermott p.1
"An FMEA is a systematic method of identifying and preventing product and process problem before they occur. FMEAs are focused on preventing defects, enhancing safety, and increasing customer satisfaction."
Commentary:


It should be emphasized that engineering technology peculiar to each field can prevent problems, and FMEA can only help it.

  • FMEA has to be focused not only on safety and customer satisfaction, but also on quality, cost, delivery (time and quantity), and environmental protection(Five Elements of Output) Simultaneously. For example, if a breach of a certain instruction maight lead to a outflow of poison, it has to be listed in the process FMEA and necessary measures have to be taken to it.
  • D. H. Stamatis p.25
    "A failure mode and effect analysis (FMEA) is an engineering technique used to define, identify, and eliminate known and/or potential failures, problems, errors, and so on from the system, design, process, and/or service before they reach the customer."
    Commentary:
    (1)FMEA is a management technique, and can help discover the existence of problems and evaluate them, but cannot eliminate them.
    (2)The practice of an FMEA helps management only, but cannot eliminate defects by itself. An engineering technology for each field is necessary to remove defects.

    1-2 Outline of Design Procedure

    (1) In General
    There are two purposes in a design. The first purpose is to give a function required for a product or a process. The second purpose is giving reliability in order not to fall into a functional disorder or misfortune simply.

    This means that the design of a function and reliability must be performed simultaneously. Therefore, although a skillful engineer's design has very few defects of reliability, amateur or a beginner engineer is in the tendency to think only a function as important and to neglect many reliability defects.

    (2) Simultaneous Activity
    The designs of the function and reliability have to be performed at the same time and pushed forward gradually. In other words you push forward some function designs and check reliability immediately and take necessary measures. Then you push forward some function designs again and check reliability immediately and take necessary measures. You repeat this cycle after this.

    When the design of a function precedes too much, accumulation of the previous work may be turned for correction of the reliability defect discovered after that.

    (3) The Function Design
    You have to design the structure of a product or process to work appropriately.It has to be carried out to prevent dysfunction that might happen during use.

    The Five Elements of Output
    The required outputs of the design procedure are related to five elements listed on the right. And troubles of a product or process occur when all these requirements are not satisfied. The designer should examine the effectiveness of measures to achieve those requirements before the completion of the design.

    1. Quality: A faulty product or service must not be produced at all.
    2. Delivery (time and quantity): The daily production quantity will advance as planned.
    3. Cost: The quantity planned on the date according to a plan must be produced.
    4. Safety: There must not be any risk of getting injured for workers or a customer.
    5. Environment: The environmental impact should be restrained as planned.
    (4)The Reliability Design
    The designer (the designer in charge, or the group in its duty) has to submit the following documents to the section manager.
    In contrast to function design, reliability design is to give durability to the structure and take measures to reduce the risk of failure. And various reliability tests support this accordingly.
    (5) Confirmations by the Division Manager
    When a subordinate engineer submits a process specifications to the section manager as the completion of a process design, how should the manager act? The manager has to ask the designer and confirm about the five elements.
      The designer has to submit "the characteristic and factor diagarams for management" and the measures list for prevention to insist that any problem cannot happen about quality or other requirements. When the designer answers "All necessary conditions are fulfilled", the head of the design division has to demand proof. When the designer answers "He is uncertain", the head of the design division has to refuse the receipt of the design proposal.
    Disclosure of evidence of Function Design
    When the subordinate designer replies to the question in the affirmative, the manager has to demand the display of proof. Do not recognize the design, when a subordinate does a negative reply or when the section chief discovers a matter which is not acceptable in proof.

    Disclosure of evidence of Reliability Design
    At the next stage, similar confirmation is necessary about reliability. If the designer submits the QC process table and the process FMEA worksheet and passes the manager's check, the manager should declare the start of the Design Review (DR) procedure. However, the design plan should not be accepted unless FMEAs are already completed.

    1-3 Need of the FMEA Execution

    (1) Role of FMEA, and Designer's Duty
    FMEA is a tool for a designer to check reliability while designing, and helps the designer assess reliability to ensure that no reliability defect exists. The means of verification for reliability includes the prototype testing, strength calculation, simulation, and others.

    (2) Before the Introduction of FMEA
    Without an execution of an FMEA, a systematic approach of reliability is impossible.
    Even without an FMEA execution, however, experienced designers have added measures to the weak points where tests, calculations, experience, knowledge or intuition indicates.
      On the other hand, the fact is widely known that a serious failure or disaster does not necessarily arise even without enforcement of FMEA if a expert designer designs. Why does it turn out such?
      This is a firm evidence that their unconscious activities mentioned above are very much like FMEAs. That is, FMEA is exactly performing examination performed unconsciously also conventionally with a clear fixed system.


    (3) Peace of Mind

    Various Changes Worry the Designer
    In the case of a product, a physical or chemical change may occur in a part or an interface as a result of use.
      For example, when a certain component is attached to the main part of a product with bolts, the bolts may loosen by transmitting a strong continuous vibration. Or exchange of the component may become difficult because of the corrosion of the bolts.

    The Designer will Apply Measures
    When such a tendency as above is felt, the designer will apply measures to the bolts to prevent those failure modes, for example, the bolts made of stainless steel which applied adhesion material may be used.



    However, some important failure modes may be overlooked and lead a serious result because these conventional activity does not obey a constant format procedure.

    Designer's Peace of Mind
    Even when a design passes verification, its strictness may not be sufficient, and the designer can not always get peace of mind. If an FMEA is conducted, the designer can maintain the current controls at an appropriate level. And he can get a peace of mind unless a remarkable shortage of measures is shown for any failure mode.




    (4) Toyota's Big Recall
    For example, a lot of verification tests are implemented by Toyota, but sometimes very small tests are overlooked. On 29 September, 2009, Toyota announced the largest recall of 3.8 million cars.
      The reason was that the removable floor mat under the driver's seat might have interfered with the movement of the accelerator, and might have lead to serious accidents.

    The failure modes
    Three failure modes can be identified about the case of the Toyota's floor mat accident, and this failure or accident can be predicted from those failure modes, If the right FMEA had been executed, the result could have been prevented perfectly.
    1. It is a customer's essential duty to check that the floor mat is being fixed to the floor when using a car. In this case, violation of directions used corresponds to a failure mode of the use process.
    2. The parts securing the mat to the floor may be broken, and hence the brakage of them corresponds to the failure modes of a part.
    3. The fixing may become loose and the floor mat may come off, and hence this corresponds to the failure mode of an interface.
    The Failure Mode of Use and Maintenance
    In this actual case, Toyota must have neglected FMEA about use and maintenance. Generally, it can be said that responsibility for problems by misuse, such as an increase in fuel cost, degree of comfort or operativity, belongs to the user.
      However, the world does not seem to admit the Toyota's claim that the user's misuse caused the fatal accident. The fact that the free repair was possible can be regarded as an evidence of Toyota's having been able to prevent the accident by enforcement of an appropriate FMEA.


    (5) Number of Recalls in 2009: At the beginning of 2010, Toyota and GM added millions of collection. And the collection of the car of the United States of America reached a vast number. This fact shows that Relative Evaluation Method(REM) of FMEA which they had carried out is not anything else except the simple decoration vividly.
    (6) Recent Advancement of FMEA
    The improvement of FMEA has accomplished notable progress recently in Japan. As a result, the Absolute Evaluation Method (AEM) has been established by OTL Institute in Japan, and many companies have abandoned old FMEA. The designers have come to carry out FMEA of Absolute Evaluation Method (AEM) now voluntarily and positively. This fact proves the following matters which designers feel.

    1-4 Basic Requirements of FMEA

    Based on the above-mentioned study, the conclusions are as follows.

    (1) Designer's Tool to Check Reliability
    FMEA is a tool for a designer to investigate the reliability of the product or process of own design. In other words, the person who conducts a design should perform its FMEA simultaneously.

    If a team takes charge of a design, it should carry out an FMEA at the same time, but it should not be called an FMEA team. It is a design team that executes an FMEA. Don't give priority over FMEA by a designer to an FMEA team which consists of outsiders of a specific design. The reason is as shown below. (2) Using a Fixed Format
    The procedure is important to FMEA and you have to learn the format of the FMEA worksheet to prevent mistakes of the procedure.

    The Worksheet Format for the Product FMEA
    Product:Gas kitchen range Product FMEA Review
    Line Article or
    Interface
    Failure Mode (Breakage) Effect Mechanism Current Control (Grounds) a b c RI Action Added a b c RI
    1                            
    2                            
    3                            

    The meaning of the identifying information of the FMEA worksheet is explained as shown in the table below.

    Identifying Information
    (↓Order of the column almost reflects the procedure of FMEA)
    Part or Interface
    Location for failure modes ( Here must not be a column of "function" )
    Failure Mode
    Manners of failure: such as rust, breakage, transformation, melt, coming-off, etc.
    Effect
    Consequences of each failure mode, i.e. potential failures and/or disasters.
    Factor
    Potetial cause: For example, fall to floor, collision, child's mischief, etc.
    Current Controls
    Measures for each factor: for example, reliability tests, inspection, strength calculation, simulation, fool-proof instruments, fail-safe system, redundant design, etc.
    Severity (a)
    Insufficiency degree of measures to ease severity of effect
    Occurrence (b)
    Insufficiency degree of measures to decrease frequency of occurrence of effect
    Detection (c)
    Insufficiency degree of measures to detect factors, a failure mode or effect, according to the situation
    Risk Index
    RI=(a × b × c)1/3 : Insufficiency degree of measures as a whole

    Caution
    You must not put the column of "Function of Component" in a worksheet. Otherwise you may fall into the top-down method. The details are explained in Appendix2.

    (3) Design Expressed on Documents
    FMEA technique is systematic means to investigate the reliability of a designed product or process. FMEA is carried out only about a design written in the document and cannot be carried out unless the design by the document exists.

    A document that describes the actual situation of a workplace is not a design document, but a mere copy document. The wrong FMEAs based on a above-mentioned copy document are frequently seen in hospital risk management. The recognition of hospitals seems to be related to it.


    (4) Difference between Failure and Failure Mode

    Concept of Failure
    A functional trouble is called a "Failure" in terms of FMEA.

    Concept of Failure Mode
    Manners in which a product or process may fail are called "Failure Modes," which are needed to be identified in bottom-up ways. When pointed out concretely, the smallest structural element of a product or process may receive a physical or chemical change due to a use or running, and this structural change is called a "Failure Mode."

    In the case of a product, a crack may occur in a shaft, and an electric wire may break. And in the case of a process, by breaching an instruction, a wrong speed setting of a lathe may occur. An FMEA execution enables you to identify all harmful failure modes.

    Reason for the Strict Distinction
    The reasons why a failure must be distinguished from a failure mode strictly are as follows.

    (5) Evaluating Lack of Measures
    Failure modes and its effects are identified and enumerated in an FMEA practice. Then, measures to the effects described in the current controls column are evaluated in terms of the following viewpoints.

    How insufficient are the current measures,

    (6) Purpose of FMEA
    FMEA must be carried out to predict all failure and disasters, and to judge whether measures for every failure are enough or not. The purpose of FMEA is not to decide priority of the importance of the failure mode. Otherwise measures become insufficient or excessive.




    Chapter2: Purpose of FMEA




    2-1 Purpose of FMEA

    The purpose of FMEA is to check reliability of a product or process that he designs, in order to enable himself to ensure reliability by a rational cost. In other words, the designer has to check his own design about whether measures are sufficient to prevent the potential failures and disasters beforehand.

    Hence, it is the necessity of additional measures to failure modes that has to be clarified by FMEA. It is not priority order of failure modes. This will be explained concretely as follows.

    To Clarify the Potential Failure and Disaster
    It is difficult to imagine what kind of potential failures or disasters may be cause in the future when you design a new product. And the same applies to a manufacturing process and other processes in general.

    The FMEA technology has been intended to identify potential failures and disasters. Hence these potential failures and disasters are the targets of an FMEA, but not the starting points. This means that you must not treat these failures or disasters as starting points of an FMEA. Starting points must be the failure modes.

    To Clarify Failure Modes not Enough in Control
    To prevent failures, you have to identify failure modes not enough in control. In order to ensure the reliability, failures and disasters must be predicted by means of searching from failure modes (the starting points). And hence, the failure modes should be enumerated first.

    And the current design controls should be listed and evaluated for each failure mode (cause and effect) in terms of severity of effect, frequency of occurrence, and probability of detection. And if necessary, measures should be added to the current controls for each element.

    Searching Failure Modes Is not the Purpose
    Because the measures are impossible to take unless a failure mode is clear, they are apt to think that the purpose of FMEA is to clarify potential failures first, and to search for potential failure modes from there. However, this is the biggest mistake of the top-down method.

    The purpose of FMEA is not to search failure modes from failures. Conversely, it is right approach to look for failures and disasters from each failure mode enumerated first.

    When FMEA Is Helpful
    The execution of an FMEA will help prevent reliability problems. For example:

    When FMEA Is Useless
    Execution of an FMEA will not help prevent troubles other than mentioned above, which should be improved by attending to the design of function itself. For example:

    Objection Theories

    Robin E. McDermott p.3:
    "Preventing process and product problem before they occur is the purpose of Failure Mode and Effect Analysis (FMEA)."
    Commentary:
    This description is not precise, and cannot avoid a correction about the next two points. And this recognition is extremely important to understand FMEA.
    Robin E. McDermott p.4:(About documented procedures)
    "This is especially critical with a process FMEA. In the absence of documented procedures, people working in the process could be introducing significant variation into it by operating it slightly differently each time the process is run. In this case, the FMEA is aiming at a moving target because each time the process is run, it produces different results."
    Commentary:
    This opinion should be corrected as follows. If there is no document :
    S. Iida. P.34:(About a process of the hospital without documented procedures)
    "All factors of the medical accident exist in the medical shop.It is essential to grasp who performs what how when precisely in preventing a medical on-site accident. The grasped things become the base of the process table." (Translated by OTL Institute)
    Commentary: This description is wrong at a basic point.

    (1) Which should follow the other?
    Which of process design and the on-site action should follow the other? If the on-site work is not performed as instructed in the process design, the process control is quite impossible.

    It is often carried out that staffs in charge of FMEA copy processes which occur at the shop. "The copied process table" made in this way is a copy till the last and it is not a designed specification. The evidence is as follows:

    (2) The Difference between Copying and Designing
    Think of the difference between copy and design. An FMEA should not be performed based on the document which is comprised of copy of on-site work.

    Since what is grasped on the site is not necessarily accompanied by a design of reliability, it should not be copied into QC process table.

    Internet site: Click Failure Modes and Effects Analysis (FMEA) - ASQ

    The purpose of the FMEA is to take actions to eliminate or reduce failures, starting with the highest-priority ones.
    Commentary:
    There are three mistakes in the opinion above.
    1. The main purpose of the FMEA is to predict serious failures or disasters.
    2. FMEA is a technique for management, and points out defects of a designed product or process. The actions to eliminate or reduce failures are taken, however, by the expertise peculiar to each field, not FMEA.
    3. The priority is nonsense after all because all necessary measures must be carried out. The reason why the opinion mentioned above requires priority is because wrong FMEA which cannot examine the degree of the need of measures every failure mode is adopted.



    Chapter3: FMEA Process



    3-1 Outline of FMEA

    Process of FMEA consists of seven steps as follows.
    1. Preparing Documents
    2. Listing Failure Modes
    3. Listing Current Controls
    4. Listing Effects of Failure Mode
    5. Evaluating Current Controls from Three Viewpoints
      - Severity (a) of Effect
      - Occurrence (b) of Failue
      - Detection (c) of Effect, Failure, Failuree Mode, or Factor
    6. Calculating Risk Index     
    7. Optimizing Measures

    3-2 Preparing Documents

    The purpose of FMEA is prediction of failures and disasters from failure modes and the evaluation of insufficiency of measures against them. Hence,
    The Worksheet for the Product FMEA
    Product:Gas kitchen range Product FMEA Review
    Line Article or
    Interface
    Failure Mode (Breakage) Effect Mechanism Current Control (Grounds) a b c RI Action Added a b c RI
    1                            
    2                            
    3                            

    The Worksheet for the Process FMEA
    Process: Medicine Bottle Ass'y Process FMEA Review
    Line Step No.
    Name
    Failure Mode (Violation) Effect Mechanism Current Control (Grounds) a b c RI Action Added a b c RI
    1                            
    2                            
    3                            



    3-3 Listing Failure Modes


    3-3-1 Concept of Failure Mode The Starting Points
    An FMEA process starts with listing potential failure modes of a product or process. In other words, the failure modes are starting points of the FMEA.
    Failure and failure mode completely belong to a different concept mutually, and should never be confused. We need to know all failures that may occur by use in the future. There are two ways to search failures. One of them is top-down approach and the other is bottom-up approach.

    The Definition of Failure and Failure Mode
    A failure occurs when a product or a process does not function properly. A failure can occur in a variety of manners. And ways, found when searched by a bottom-up approach, in which a product or process may fail are called failure modes.

    For example:

    Caution
    The error is not a failure mode without the violation for instructions written in QC process table or other associated documents even if it leads an accident that somebody did some kind of errors.

    3-3-2 Bottom-up Approach
    Bottom-up Search
    The most important point in listing up failure modes exists in the bottom-up search. One of the reason is to prevent overlooking of failure modes. Breakage in the smallest structural elements are defined and identified as failure modes and treated as starting points of searching failures.

    In other words, the damages of parts or the connections (with a product) or breaches of instructions (by a process) are the failure modes and are the starting points of all FMEA activities.

    Failure modes can easily be listed from a parst and connections or a process instructions, if expert designer is in charge.

    Top-down Search
    In the top-down approach, the failures are necessarily the starting points to search failure modes, and efforts are paid to investigate downward from the potential failures. This is because failures are listed first, and the questions how they does occur are considered next. As the result,

    are enumerated without distinction all as failure modes. On the other hand, the failues which are missed become what you call unexpected failures that no precautions are taken beforehand.

    The Merits Brought by Bottom-up Approach
    The expert designer can enumerate all possible potential damage about parts and the connection that oneself adopts without exception in the future.For example, there may be enummerated such thinsg as;

    Because an FMEA starts with a question what failure will be caused if this failure mode occurs, failures are not included in the starting points. A failure (even a dysfunction of the sub-function) does not belong to the failure mode concept.

    The Problems with Top-down Approach
    When top-down, failures are enumerated first. Next, it shifts to investigation of the routes in which failure occurs for every failure. And the activity which looks for failure modes starts. Various phenomena are listed as failure modes as follows if the failure of the windshield wiper of a car is taken as an example.

    - The turn of the wiper shaft may not be smooth.
    - The switch may not work.
    - The motor may not work.
    - The wiper shoe may be deformed.
    - The arm of the windshield wiper may be broken.
     
    (A failure or manufacturing defective)
    (A failurfe or manufacturing defective)
    (A failure or manufacturing defective)
    (A failure mode)
    (A failure mode)

    When failure is shown as above, it is not the end of the search of failure mode, and you have to continue looking for the failure modes of the failures further. This means the failures are not failure modes.

    When a manufacturing defective is pointed out as above, this is not the end, too. You have to search defects existing in the manufacturing process, further. You notice that you do not know which of a product FMEA and process FMEA you are in charge of.

    Failure Mode must Be Starting Point
    Failure modes (structural breakeges) should be the starting points because:

    1. The purpose of FMEA is to look for all failures and disasters, and hence failures or functions must not be starting points.
    2. It is not difficult for an expert engineer to enumerate failure modes of a part or interface that is adopted in the own design. And from the failure modes enumerated, it is not difficult to approach all of the corresponding potential failures or disasters by bottom-up search.
    3. Without these starting points, it would be very difficult to imagine all failures and all disasters that are potential.
    4. Disasters are not always caused by failures (non-functions or malfunctions). Hence starting from failures will not cover all the disasters. For example, even if leakage of electricity occurs, it is not necessarily accompanied by functional disorder, but is in danger of a fire.

    3-3-3 A Useful Theory by H. Kume: p.142: (Translated by us from Japanese)
    Difficulty in Predicting Failures
    The difficulty in prediction of potential failures is caused by necessity of qualitative prediction. In contrast, a weather forecast gives a quantitative prediction to the events which are limited to several like fine weather, cloudy weather, rain, snow, and the storm, etc.

    For Example
    Concerning new technologies or new products, listing of potential failures is not an easy task. For example, Japan is proud of the Tokaido Shinkansen System, which is a high technical level railroad train. Initially, however, they encountered many unexpected problems.

    Solution
    Failure range is diverse and systematic analysis needs a special concept which represents as much diverse failures as possible. It has enabled us to predict potential failures by establishing the failure mode concept. Failures (= effects) of those failure modes are different depending on products. And picking all failures in general is impossible and meaningless. However, once a system (= product or process) is decided, searching failures will become easy starting at the failure modes. On the other hand, it also will become easy to identify causes and effects.

    Commentary:
    The above-mentioned theory by Professor Kume (the former professor of University of Tokyo, Japan) is appropriate and splendid truly.


    3-4 Listing Current Controls

    Identification and Listing of Current Controls
    The current controls (i.e., measures) to improve the severity of effect, the frequency of occurrence, and the provability of detection should be identified and listed respectively.

    One description space
    In many cases, one measure is effective in more than one of severity, occurrence, and detection. The measures written are treated as the measures for each of severity, occurrence, and detection, and hence one description column is enough.


    3-5 Listing Effects of Failure Mode

    Identification and Listing Effects
    All failures and effects resulting from each failure mode should be identified, however, only effects that are obviously harmful should be listed.


    3-6 Evaluating from the Three Viewpoints

    Three viewpoints
    Current controls for each failure mode are evaluated in terms of three viewpoints i.e., The evaluation is carried out on the absolute four-point scale regarding necessity in adding measures to the current controls.

    Evaluation by the Engineering Technology
    The current controls to each failure mode should be evaluated from the three viewpoints on the four-point absolute scale which is based on the engineering technology for each field. In other waords, this rating is not related wity management technology.

    Severity (a)
    Evaluate the degree of insufficiency of measures in current controls to ease the severity of the effect. If a fail-safe system is taken, the severity takes a very low value.

    Occurrence (b)
    Every cause, evaluate degree of the lack of measures of the current controls to reduce outbreak frequency of the failure. If a fool-proof system is taken, the occurrence takes a very low value.

    Detection (c)
    Every cause, evaluate the degree of insufficiency of measures in current controls to improve the probability of detection of failure. However, as an object of detection, the thing that detection is the easiest can be chosen among potential cause, failure, and effect by circumstances.

    The influence of the trouble mode may gradually develop into serious influence according to the flow of the process under light influence first. You may choose the method before developing into a serious accident by the detection of early light influence if measures are possible.

    Exceptional Evaluation Criteria
    When the frequency rating is low, the necessity of detection is also low. Hence when occurrence takes such low evaluation as 2 or 1, the detection takes the same rating as that. And this exceptional treatment accelerates FMEA very much.


    3-7 Calculating Risk Index

    Numerical value of Risk Index (RI) is calculated using the ratings as shown above and an expression shown below.
    - Severity value= a
    - Occurrence value= b
    - Detection value= c

    Thr ratings a, b, and c are obtained on the basis of engineering technology for each field, and RI is the index from the three viewpoints of management technology to indicate the necessity of additional measures. The outline of this principle is shown in the figure below.



    3-8 Optimizing the Measures

    Getting RI Closer to the Optimal Value 2
    The numerical value of risk index (RI) must be appropriate, i.e., as close as possible to its optimal value 2. And its measures are adjusted, and this treatment is called optimization. However, the content of optimization varies according to the value of RI as shown below.

    When RI Exceeds 2.3
    As RI exceeds 2.3, the need of additional measures increases.

    When RI Is around 2.3
    Actions will be suspended, until the result of test is obtained or a better measures is found.

    When RI Is Considerably Lower than 2
    Because reliability is superabundant, consider whether therefore extra cost does not exist. When reduction of the cost is difficult, aim at value-added increase appropriate for the cost increase. Consider how to create a selling-charm (for example, improvement of appearances, convenience of use, or safety) to compensate for the cost increase.

    Defect of the Conventional Evaluation
    Since RPN, which the conventional FMEA is proud of, does not indicate the necessity for measures, adjusting of measures can not be performed on the basis of this numerical value. Besides, the comparison between first RPN and RPN after the improvement is totally nonsense because neither the cost nor the time delay is considered.


    Objection Theories Robin E. McDermott p.9:

    "The objective of an FMEA is to look for all of the ways a process or products can fail."
    Commentary:
    This opinion cannot be said to be a mistake. However, McDermott's is missing the turning point of bottom-up or top-down approach about "looking for ways a process or a product can fail". This book itself has taken the wrong choice, and as a result, all ways (trouble mode) are not found and other failures and manufacturing defectives are mixed with failure mode.

    Robin E. McDermott p.9:

    "Failures are not limited to problems with the product. Because failures also can occur when the user makes a mistake, those types of failures should also be included in the FMEA."
    Commentary:
    This description is correct, but has missed some explanation points. For example: Robin E. McDermott p.9:
    "Ways in which a product or process can fail are called failure modes."
    Commentary:
    This mention is also true; however, the turning point "top-down or bottom-up" has been missed. And this book itself has chosen top-down identification. The mistake of this choice seems to have led to the mistake of the whole FMEA.

    Robin E. McDermott p.10:

    "The relative risk of a failure and its effect is determined by three factors:
    • Severity: The consequence of the failure, should it occur.
    • Occurrence: The probability or frequency of the failure occurring.
    • Detection: The probability of the failure being detected before the impact of the effect is realized."
    Commentary:
    Explanation mentioned above has serious mistakes as follows.

    (1) What should be determined?
    It is wrong to determine the relative risk of failure, because it does not become clear which failure modes need additional measures and which need not. Correctly, the way that a failure mode needing additional measures is clarified, in other words, the absolute risk evaluation must be carried out.

    (2) The Relative Risk Evaluation is Useless
    The relative risk evaluation is useless because the necessity of addition of measures for each failure mode does not become clear. Only for the reason that the risk of failure mode A is larger than that of B, measures should not be immediately added to A. Excessive measures lead to waste of resources (person, time, equipment), and shortage of measures also increase waste for the troubles.

    (3) We Want to Know the Absolute Necessity of Measures

    Evaluating the Degree of Risk Is Absurd
    It is absurd to evaluate the "relative degree of risk" of failure modes.

    For example, the effect of the crack of the airplane body should be estimated at a high level (probably "10"). However, the effect of the cracks of the paper cups for passengers will be evaluated by a low level as the very trifling problem, compared with an aircraft accident. However, on the other hand, the paper cup maker shuold give the greatest point to the crack.

    The same inconsistency is seen also by evaluation of frequency. Suppose that frequency is once a year. The crack of the airplane body in the frequency should be evaluated at a high level (probably "10"). Moreover, it in the paper cup in the frequency should be evaluated at a low level.

    Because of such complex conflicts, it is difficult to establish common criteria and make smooth use. In addition, each numerical value obtained by the relative evaluation can not be utilized as the data to determine excess or shortage of measures.

    Robin E. McDermott p.10:

    How to Calculate RPN
    "Using the data and knowledge of the process or product, each potential failure mode and effect is rated in each of these three elements on a scale ranging from 1 to 10, low to high. By multiplying the ranking for the three factors, a risk priority number (RPN) will be determined for each potential failure mode and effect."
    "RPN=Severity × Occurrence × Detection"

    Purpose of RPN
    "The risk priority number (which will range from 1 to 1,000 for each failure mode) is used to rank the need for corrective action to eliminate or reduce the potential failure modes. Those failure modes with the highest RPNs should be attended to first, although special attention should be given when the severity ranking is high (9 or 10) regardless of the RPN."

    Resulting RPN
    "Once corrective action has been taken, a new RPN for the failure is determined by reevaluating the severity, occurrence, and detection rankings. This new RPN is called the 'Resulting RPN.' Improvement and corrective action must continue until the Resulting RPN is at an acceptable level for all potential failure modes."

    Commentary:
    In the above description, six major flaws are pointed out as follows;
    1. Absence of Common Criteria
      As mentioned above, the relative scale method has no common scale which can be applied to all products and processes, and hence smooth evaluation is quite hard to conduct.
    2. Absence of Reproducibility
      The 10-point relative rating method is very hard to execute and the result is extremely crude without reproducibility.
    3. Business Retention
      Because addition of the improvement is reserved to the end of all RPN calculations and priority decision, the large delay of the work cannot be avoided.
    4. How can "an acceptable level" be determined?
      "An acceptable level" mentioned above is vague about how to decide, and the executer cannot judge which failure mode to work on. In fact, there is no means except intuition.
    5. Priority Is Nonsense
      If "an allowable level" can be determined, it should be determined first. And it will become unnecessary to examine the priority. As the result, it becomes clear that the absolute evaluation method is the only evaluation method that FMEA should take.
    6. Cost and Time Delays
      The judgment is impossible whether RPNs after the improvement are superior to RPNs before the improvement because you do not consider cost and delay in the relative rating system.
    Robin E. McDermott p.47
    Case Study Step 8 : Prioritize the Failure Modes for Action
    One of the team members created a Pareto diagram of the failure modes so that it would be easy to distinguish virtually between the items. The team decided it would work on any item that had an RPN of 200 or higher. Two hundred was set as the cutoff point because it encompassed over half of all of the potential failure modes.

    The team rationalized that an improvement in more than half of the failure modes would be a significant step in the right direction. With the criteria of an RPN of 200 or higher, there were eight items they would need to attend.

    Commentary:
    1. The reason why the cutoff point which encompasses over half of all of the failure modes is rational is not shown. This is totally measures without grounds.
    2. About the design without need of improvement of the reliability, RPN which encompasses half of all failure modes is set as the cut-off point, and further actions will be added irrationally.
    3. Because RPN is defined for numerical value indicating the priority, it is against the definition to use it for numerical value indicating the additional need of measures.

    Internet site: Click Failure Modes and Effects Analysis (FMEA) - ASQ

    "Failure modes" means the ways, or modes, in which something might fail.
    Commentary:
    The above-mentioned definition is not touching on whether to be top-down or bottom-up in identification of failure modes. As the result, it has fallen into the top-down approach which is a serious error.




    Chapter4: Who carries out an FMEA?



    4-1 A Tool for the Design Aid

    Real Time Execution of an FMEA
    The designer of a product or process can give it appropriate reliability through real time execution of an FMEA. It is possible to perform FMEA, after all design proposals are completed. However, if an improvement of reliability is needed after a design progresses there, it may develop into a large-scale change of design, and this is not a wise method.

    Furthermore, when the design results in a high cost because of excessive reliability, a large scale change of design similarly is needed. As a result, a remarkable delay can not be avoided.

    Simultaneous FMEAs
    This problem is solved only by performing FMEA simultaneously with a design. In other words, FMEA is a tool with which a designer checks the reliability while designing.

    FMEA team
    The thought that an FMEA team who are not in charge of the design should take charge of FMEA is essentially wrong.

    Reason:

    1. Supposing an outsider's team carries out, while a designer designs, there will be no method of checking reliability.
    2. It is too much difficult work that an outsider's team understands a design, and FMEA by an outsider's team wastes time and a labor notably.

    Objection Theories

    Robin E. McDermott p.11:
    "Although one person typically is responsible for coordinating the FMEA process, all FMEA projects are team based. The purpose of FMEA team is to bring a variety of perspective and experiences to the project. It is helpful also to have people on the team who have different levels of familiarity with the product or process. Those who are most familiar with it will have valuable insight, but may overlook some of the most obvious potential problems."
    Commentary:
    (1) Nobody but the expert designer can point out failures most precisely.
    The description "Those who are most familiar with it (the design) may overlook some of the most obvious potential problems." is a groundless assertion, because overlooking of problems is due to the defects of the conventional FMEA. It is nobody but the designer that can point out failures and disasters from failure modes most precisely.

    (2) As for Variety of Knowledge and Experience
    The opinion which supports the team-based FMEA has the tendency to assert various knowledge and experience as the reason. If various knowledge and experience are needed, however, not only FMEA but also design should be performed by the same team. It is more suitable to have sufficient capability beforehand rather than compensating man's lack of ability about a design later.

    We can not find the reason why a special FMEA team is necessary in addition to a Design Review team which is similar to an FMEA team.

    Hitoshi Kume p141: (Translated by us from Japanese)

    "Design review (DR) enlarges the width of check of the product design. In contrast, FMEA increases the depth. DR and FMEA are activities to be carried out both by groups, but DR performs examination about the whole of the product. In contrast, FMEA examines the prevention of the trouble of the product, safety improvement mainly, and participants are engineers mainly."
    Commentary:
    Professor H. Kume has explained a failure mode concept by a wonderful theory. In contrast with it, his theory about who performs an FMEA is very poor.

    (1) Why does Only Reliability Need a Deep Rview?
    The design review can include a FMEA review. It is not fully explained why the review of FMEA of the designer by DR team is irrational. It is hard to accept the idea that a deep examination is required only for FMEA and the depth is not required for other matters.

    (2) Reliability Check by the Designer
    The upper description has forbidden the natural procedure of performing a reliability design, while a designer checks reliability at the time of a design. This is not rational. However, when a team designs, the design team should perform FMEA. In this case, you should not call it a FMEA team.

    D. H. Stamatis p.32:

    "Who conducts the FMEA? The FMEA is a team function and cannot be done on an individual basis."
    Commentary:
    There are two grate mistakes. This is a extremely poor opininon that does not examine even such extremely rudimentary problems as shown below.

    (1) How Can the Designer Check Reliability?
    If the FMEA cannot be done on an individual basis, how could a single designer check the reliability of regarding his own design? The answer is not shown.

    (2) Which Is More Difficult, Designing or Executing FMEA?
    If the FMEA cannot be done on an individual basis, why can design be done on an individual basis? The answer is not shown.

    Which is more difficult to perform, an FMEA or a design? Of course design is ten times more difficult than conducting an FMEA, because the designer must contrive function mechanisms and structure to support these taking into account production methods, quality, costs, time-quantity demands, safety for workers and customers, environment, and reliability.

    If a team executes an FMEA, the same team should take charge of the design. Similarly, if design is done by one person, the same designer should execute the FMEA alone. In addition, even if a design is executed by a team, an FMEA for each member's portion of design should be performed by each designer beforehand.




    Chapter5: The Product and Process




    5-1 Designing the Product and Process


    5-1-1 General Description For the Benefit of All People
    The product must be designed to be appropriate to the benefit of customer, manufacturer, cooperating company, neighborhood inhabitants, and environmental protections. Hence, the outcome of the design must meet the demand about quality, time and quantity, cost, worker's safety, user's safety, and environment.

    Purpose of Design (Function and Reliability)
    Design is carried out so as to ensure functions and reliability of a product or process to meet the design intent and requirements. The design of function and reliability should be performed simultaneously. If a defect of reliability is detected after the design of function advances too much, previous accumulation gets overturned, and duties delay remarkably.

    Definition of Design
    Design Is Bunch of Measures
    Why do both the product and the process appoint detailed structure specifications by a design? The design must consist of measures to prevent all defects. In other words, not only the product design but also the process design must be a bunch of measures. Otherwise incoherent works are carried out, and innumerable troubles happen.

    Structure of Process
    The process consist of specifications of the five-M resources as shown below. These are the requirement about,

    Machinethe equipment and its maintenance
    Manthe people in charge or related
    Methodthe work, operation, engineering, management
    Measurementthe collection, recording and treatment of information
    Materialthe subject which receives processing

    Generally, the consideration about the element of five resources is necessary so that a process functions.These placement and preparations to carry out a process constitute the structure of the process.


    Definition of Function
    Function is the ability of a product or process to work and output based on the design intent. Output of function consists of five elements as described in the table below.

    Quality Q Products should have quality conforming to requirements, and processes should have quality appropriate to output products of quality conforming to requirements.
    Delivery
    (time and quantity)
    D As for product design, sale time and the volume of the product should be considered. As for process design, daily amount of production should be considered.
    Cost C Initial cost and running cost should be considered in both product and process design.
    Safety S Both of worker's and customer's safety should be considered in both product and process design.
    Environment E Environment protection should be considered in both product and process design.

    Principle of QDC Inseparability
    The details of 5M resources are strictly and definitely specified as design specification so that the five required characteristics shown above are completely satisfied. The design specification of a product is mainly expressed by drawings or specifications. The design specification of a process is mainly expressed by QC process tables.

    It is necessary to manage the characteristic as what five results cannot divide. If only one of these is managed, the other 4 will fall into a non-controlled state. When these five characteristics are managed by the separate administrator, each plan collides and problem always arise.

    Verification of Functions

    Definition of Reliability
    The quality of a product or a process that can maintain the intended functions under specific conditions and within a prescribed period is called reliability.

    Verification of Reliability Definition of Failure
    The inability of a product to work properly after normal use for a time is called a failure (non-function or malfunction). This is what we call a failure which should be prevented by reliability design, and reliability should be evaluated by conducting an FMEA to check whether reliability is sufficient or not. In this case, an FMEA can play an important role.

    Poor Function Existing before Use
    When by contrast not functioning appropriately already prior to handing to the user, it is a defect to which it comes not from failure but from a poor manufacturing process. These problems are unrelated to reliability or FMEA.

    How to Start an FMEA If Prevention Is not Sufficient
    How can we identify all failure modes when instructions for preventing failures are not sufficiently specified in the process design? In that case, the failure modes which can be enumerated will also run short corresponding to shortage of instructions.This is the defect of design, and it is necessary to do the following thing prior to creating the QC process table so that all necessary instructions are described.

    All the factors (potential cause) about all output characteristics should be enumerated to the cause-effect diagram. And all necessary measure about those factors should be taken in the QC process table to be created.

    Factor and Effect Diagram (FED)
    All failures (function troubles) about output should be enumerated first. And at the next stage, all factors (potential causes) should be enumerated toward each failure to form a factor and effect diagram (FED).

    FED and CED
    The factor effect diagram (FED) and the cause effect diagram (CED) are fundamentally different.


    Creation and Use of FED
    The process design which has not sufficient instructions for the satisfactory result is called an imperfect process design. In the case of an imperfect process design, listing of failure modes cannot fully be performed. It is necessary to begin the solution by creating perfect FEDs.
    1. Enumerate all potential trouble from knowledge and experience.
    2. Enumerate all factors (potential causes) of each trouble from knowledge and experience.
    3. Plan suitable measures to all the factors according to the importance.
    4. Describe factors and their controls concretely and definitely in each step of the QC process table.

    5-1-2 Failure and Failure Mode

    Failure
    Malfunction or nonfunction is called a failure. Both of a product and a process can produce failures by breakages of the smallest structural components.

    For example:


    Opposite Views

    Robin E. McDermott p.19:
    "The principles and steps behind all FMEAs, whether they are focused on the product or the process, are the same even though the objectives may differ."
    Commentary:
    Correctly, steps vary depending on which of a product or process is treated.

    Difference of Means and Timing of Detecting Failure Modes

    Product FMEA
    The main means of detection are strength calculations, reliability tests with prototypes, and computor simulations. These primarily are performed during design control stage. And the second is a certain periodical inspection by the user (or a trustee) after sale acording to the user's manual.

    Process FMEA

    Timing
    The failure method of the process is found during process preparations and running.

    1. Detection of a small effect may be enough if you can find the small influence that can be easily settled before developing into the serious effect.
    2. It is necessary to detect the stage of the trouble mode if too late by the detection of the stage of the influence.
    3. It must be found at the time of the factor (potential cause) of the failure mode if too late by the detection at the stage of the failure mode. Or a mechanism which the failure mode does not occur is necessary.
    Means
    Most means of detecting failure modes or reducing need of detection are as follows.
    1. Means of detecting failure mode
      The confirmation of meeting a process condition in most cases is carried out using checksheet.
    2. Means of reducing need of detection
      The means shown below prevent errors perfectly, and as a result, reduces the necessity of detection.
    3. Means of detecting abnormality
      When the abnormalities of the characteristic are detected at an opportunity as shown below, generating of failure mode may be detected.


    5-2 Product FMEA

    5-2-1 General Description
    Purpose of Product FMEA
    Product FMEA is conducted for the purpose of uncovering potential failures and disasters and evaluating the insufficiency degree of measures for prevention.

    Failure Modes is Easily Misunderstood
    Failure modes are manners in which a product can fail. This definition is the generally received view and is right. However, it is easily misunderstood, and in fact, most of the researchers and trainers misunderstand it. And the misunderstanding leads to greater mistakes further.

    Should manners in which a product may fail be searched in a top-down approach or bottom-up? This is the most important turning point. However, few researchers have been aware of this. As a result, they have fallen into the wrong way, the top-down approach.


    5-2-2 Bottom-up Approach
    Trying a Bottom-up Search
    Ask like in the following way : what (a failure and disaster) may happen if a crack (a failure mode) is created in this part? This way of asking is the bottom-up approach. In this approach, an approach to a failure starts from a breakage that may occur in parts or interfaces (connections). This kind of breakage belongs to the structural defect and is called a failure mode of a product.

    The failure mode of a product is limited to a structural defect and is definitely distinguished from a failure (functional defect).

    Enumerates All Harmful Failure Modes
    The failure mode has to be the starting point of searching failures or disasters which is the object of FMEA. The designer should be an expert capable of enumerating all harmful failure modes easily without overlooking. The bottom-up approach will lead to the failure and disaster which should be prevented.

    For example, when attaching a certain unit to a main part of a machine using bolts, the expert designer must immediately enumerate the failure modes of the bolts as follows.

    Bottom-up Approach
    Potential Cause Failure Mode Potential Effect
    Vibration What may cause this failure mode? Slack What may be caused by this failure mode? Unit may move and electric wires may cut. And what?
    Repetition of Impact load Broken
    Excessive dead load Bent It may become hard to loosen bolts and to replace unit.
    Repair may become hard at last.
    Dew condensation Rust

    A Novice Engineer
    It may be worried that all harmful failure modes are not enumerated if difficult for a novice engineer. However, it is not a problem which arises only about FMEA that it is impossible to leave a design to beginners. Beginners have to work under an expert supervisor.

    The Principle of the FMEA
    All the failure and the disasters which are due to a certain failure mode can be searched for from that failure mode. This is the technical foundation stone of FMEA.


    5-2-3 The Top-down Approach Is Wrong
    Trying a Top-down Search
    "What is the failure mode which draws this failure or disaster?" This way of asking is a top-down method, and unexpected failure or a disaster will show a tendency for measures to leak.

    Let's think of the top-down approach to the manners in which the windshield wiper may fail. What can be cause? The answer may be the defects as shown below.

    The Top-down Approac (Wrong)
    Article
    Function
    Failure
    → Failure Mode
    Subfunction Failure Structural Breakage Manufacture Defectives
    Wiper ASSY Wipes window smoothly Intermittent motion Battery voltage drop Wiper arm bent
    Wiper shoes worn
    Heat wrong
    Material mistake
    Does not move Battery voltage drop Switch contacts worn
    Electric wire broken
    Coupling screw slack
    Short circuits
    Damaged wire insulation
    Problems → ↑ Othee failures might fave been overlooked. ↑ These are failures (wrong). ↑ These are failure modes (right). ↑ These belong to the process FMEA (wrong).


    Note : The resulting top-down feature:

    Objection Theories

    D. H. Stamatis. P.74:
    - Failure : The inability of the system to perform based on the design intent. Examples of failures are Broken, Worn, Noise, and Rust.
    - Failure mode : This is the physical description of the manner in which a failure occurs. Examples of failure modes include the following: Broken, Grounded, Corroded, and Leak.
    Commentary:
    Any of "Broken, Worn, and Rust" cannot be a failure, because they are not "the inability of the system." This is one of the typical theories which take the top-down approach, and evidences of confused concept of failure mode are clearly uncovered. For example, "Broken" belongs to both of failure and failure mode. This is the proof of theory collapse. The boxed comments shown below the table with arrows are the correct understanding.

    Robin E. McDermott p.45 :
    Examples of filure modes of a product (Fire extinguisher) are shown, and we have pointed out some mistakes of alignification.

    Component and function Failure Mode Correct alignification
    Hose ;
    delivers extinguishing agent
    Cracks ○ Failure mode
    Pin holes    × Manufacturing defective
    Blockages ○ Failure mode
    Canister ;
    resrvoir for extinguishing agent
    Paint coverage uneven    × Manufacturing defective
    Label not properly applied    × Manufacturing defective
    Charge gauge ;
    determin remaining volume of agent
    Inaccurate reading    × Failure
    Broken crystal ○ Failure mode
    Valve mechanism ;
    release agent
    Safety pin missing ○ Failure mode
    Handle jams    × Failure
    Commentary:
    A wrong product FMEA worksheet case is shown above. Some wrong words are listed in the column of "Failure mode". And the correct concepts are shown by us at the right end column.

    Look at the left edge column. The word hose and the function are listed there. The function is to supply extinguishant. How can this function be spoiled? Yes, the cracks, the pinholes and the blockage can spoil the function of the hose.If you take this way of approach, however, other important failure modes can be overlooked. For example, coming-off of the hose from the connector can be another failure mode.

    Furthermore, those pinholes are production defectives. Because they are not matters caused by the use of the product, they are not failure modes.

    Because beginners are instructed to conduct the traditional FMEA by their leader, they tend to conduct a top-down FMEA. They are not grown up enough to consider things independently, and are easily affected by the people around. They are familiar with the traditional FMEA, and repeat errors without hesitation or doubt.


    5-2-4 Manufacturing Defectives
    The manufacturing defectives including assembling defectives are one of the results derived from the process design or the workshop management, and cannot be failure modes for the three reasons shown below.

    1. Products must Conform to the Design
      A product FMEA checks a performance of a product design in the aspect of reliability and hence the product under the evaluation of an FMEA is limited to one that conforms to the design. And nobody designs manufacturing defectives. In other words, a manufacturing defective cannot be a failure mode.
    2. Failure Modes Are Caused by the Use
      Manufacturing defectives are produced prior to use, in contrast to failure modes.
    3. The Role Outside of Product FMEA
      Limitlessly enormous numerical potential failure modes can be enumerated if manufacturing defectives belong to the failure mode concept of a product. And FMEA comes to a deadlock.

    Robin E. McDermott p.9:
    "Failures are not limited to problems with the product. Because failures also can occur when the user makes a mistake, those types of failures should also be included in the FMEA."
    Commentary:
    This is explanation in the sections of product FMEA and this explanation is based on a wrong theory. The mistake of the user must be handled in use process FMEA not product FMEA. This book does not seem to understand it.

    Designer's Duty
    Accidents caused by incorrect use or maintenance can not always be prevented without an FMEA of the customer's instruction manual. The designer of the final product must design a customer's instruction manual and should carry out its FMEA further.

    Case of Finger Cut Accident by Shredder
    The product of shredder has to be safe even if a small child inserts a finger. A warning label to put on the product is insufficient as means of prevention. It is necessary to determine whether means of the accident prevention is enough by carrying out FMEA.

    Examples of Measures
    Some concrete plans over the accident of such a product shredder can be devised easily.

    1. A couple of parallel switch buttons to be held down at the same time,
    2. The narrow loading slot which inserts the paper to discard into,
    3. Structure where its finger which a small child inserts does not reach a cutter,
    4. One leg is attached to the bottom of a product shredder in order to protect from being put on the floor. On the other hand, the attachment for fixing it to a desk is attached.
    In conclusion, to attend to the measures as above, a design should be considered about use and maintenance. And an FMEA about use and maintenance should be performed to judge the degree of insufficiency of measures.


    5-3 Process FMEA

    Failure Modes of a Proces
    Failure modes of a process correspond to a breaches of each process specification. All specifications of the five-M resources are defined at the stage of process design and are realized at the stage of setting and running of the process. And all five-M requirements have to be maintained faithfully for the process to work properly and meet all output requirements about quality, delivery, cost, safety, and environment.

    For example, a wrong torque setting of an electric screwdriver is a failure mode if it is a violation of process design (i.e., QC process table). The outcome of the manufacturing defective shows that a failure is taking place in the process. And this defect may lead to the next effect that is more serious.

    Dysfunction of Equipment
    When a machine using in a certain process broke down, is it a failure mode of the processes? It is not a failure mode, but a failure of the process, because process does not work properly. Then, what is the failure mode in this case?

    Equipment Maintenance Process
    Equipment should be maintained in accordance with maintenance plans, in which requirements are defined in detail. Breaches of instructions of the maintenance plan is called failure modes. When a failure arises without violation of maintenance plan directions, failure mode does not exist. In this case, the omission in regulation of instructions exists in the maintenance plan.

    In the case of 5M alignification, the facilities maintenance belongs to "Machne".However, the facilities maintenance can make a alignification independent of preparations for use. In that case, the alignification becomes six, and the name of 5M is exchanged for 6M.


    Objection theories

    Robin E. McDermott p.20:
    Process
    "Process FMEAs uncover process problems related to the manufacture of the product. For example, a piece of automated assembly equipment may misfeed parts, resulting in products not being assembled correctly. Or, in a chemical manufacturing process, temperature and mixing time could be sources of potential failures, resulting in an unusable product.

    It is helpful when conducting a process FMEA to think in terms of the five elements of a process: people, material, equipment, methods, and environment. With these five elements in mind, ask: How can process failure affect the product, processing efficiency, or safety?"

    Commentary:
    This is all of explanation of Mr. Robin E. McDermott about process FMEA. If he cannot describe process FMEA other than this, it is thought that he has neither the understanding of process FMEA nor the experience. Such troubles may occur and should be prevented. That is right, but the author of this book should learn process FMEA from the first step again.

    The upper description has four basic mistakes.
    (1) Any definite answers to the questions below are not found, even through whole of the book. (2) Replace "Environment" with "Measurement"
    Among the five elements, it is right to eliminate "Environment", because environment conditions should be controlled by use of equipment. This belongs to the requirement of machines (or equipment). "Environment" should be taken into consideration as an object of protection, and is one of the five process output characteristics. That is, the five result characteristics of a process are quality, delivery (time and quantity), cost, safety, and environment.

    On the other hand, "measurement" should participate the five elements of process condition, because it is an important element meaning the information management. Measurement of,

    are indispensable for management. For example, process conditions and process results should be measured for control. In addition, the time of measurement, analysis of a measurement result, check, calibration of measuring apparatus, etc. are included in this process element.

    (3) Think of the Five Requiremenys When Designning
    It should be thought in terms of five condition elements and output elements of a process at the time of a process design, and it is not a time of carrying out process FMEA. The requirements of the following five elements should be decided at the time of process design to meet requirements of the result.

       Man:
       Requirements which those who take charge of processing should satisfy
      Machine:
       Requirements which equipment, machines, tools, etc. which are used for processing should satisfy
       Material:
       Requirements which what receives processing should satisfy
       Method:
       Requirements about how to perform processing
       Measurement:
       Requirements about: information gathering, measurement, record, and analysis required for execution of processing

    On the other hand, the process FMEA is the stage where the designer evaluates the effect that may result from the breaches of the five element requirements determined in the design stage.

    D.H.Stamatis.p.75:

    Level 1 Failure mode :
  • For a product FMEA: "Cannot generate pulse."
  • For a process FMEA: "Poor material feed."
  • Commentary:
    This book misunderstands the meaning of failure mode basically.



    Chapter 6. The Seven Steps / Case Study

    Step 1. Preparation of Documents


    6-1-1 In the Case of the Product FMEA

    For an product FMEA, the product design documents such as below should be prepared.

    (1) Drawings
    An example of assembly drawing shown below indicates the "introduction hose and gas range assembly", which is comprised of a gas kitchen range set and such connection parts as the introduction hose and the hose band sub-assembly.


    The number in the parentheses shows the necessary number of the part.
    • The introduction hose (1)
    • The hose connector (1)
    • The hose band (1)
    • The retaining screw (1)
    (2) Structured BOM (or, P/L)
    These parts are enumerated to a part list as shown below. Each part should be listed in the place indicating assembly or component it belongs to.

    (3) User's Manual
    The manual for users is the document which listed a process of use and the maintenance of the product. It should be made under the guidance of the designer of the end product, and hence the designer should also perform FMEAs of those processes.

    A Case of Panasonic's Recall
    A breach of instruction about detergent by a user of the dishwasher made by Panasonic (Japan) caused generating of the superfluous air bubbles which invaded the coil domain of the fan motor and caused a fire, 2020. Since the same accident occurred several times, Panasonic began a free part exchange as measures of the accidents of the fire.


    Necessity of Reconsideration of FMEA
    Panasonic (Japan) realized the appropriate solutions in a short time with an outstanding technology. However, because it has been responsible for similar incidents about many kinds of products, Panasonic's FMEA seems to have a problem.

    (4) FMEA Worksheet for Product FMEA
    A form of FMEA worksheet for the absolute evaluation method is shown next.

    The Worksheet for the Product FMEA
    ( a:Severity, b:Occurrence, c:Detection, RI: Risk Index )
    Product:Gas kitchen range Product FMEA Review
    Line Article,
    Interface
    Failure Mode (Breakage) Effect Mechanism Current Control (Grounds) a b c RI Action Added a b c RI
    1                            
    2                            
    3                            


    This form is simpler and more sophisticated compared with the traditional form shown bellow. The differences will be explained in details in the Step 4 (current controls).

    6-1-2 Reliability Design

    Do not Depend on Correction
    The reliability of a system like a product or process is defined as the property of being hard to fail in short. A product or process must be designed so that failures do not happen by normal use under a certain scope of conditions. An FMEA must be performed in order to ensure reliability in the design of every product or process by the designer.

    The reliability should be evaluated to determine whether additional measures are necessary. Otherwise, products which have low reliability may often be shipped, and a variety of process trouble may happen. And it becomes the tendency that engineers perform temporary measures such as "Band-Aid" frequently.

    Two Types of Failure Mode of a Product
    The details of the reason are mentioned afterwards, but only two kinds of failure modes can exist for the product FMEA.

    1. The failure mode which happens in a part in the form of physical or chemical change, for example, cracked, bent, rust, worn, melted, etc.
    2. The failure mode which happens in the form of physical or chemical change in a interface (connection) between parts or part and its surrounding. For example, wet, dirt, separation, coming-off, blockage, slacked, etc.
    Thus, the parts list and the drawings are important in carrying out a product FMEA.


    Objection Theories

    Robin E. McDermott p.24:
    (Quoted format from Robin E. McDermott p.24)
    Commentary:
    This worksheet is accompanied by some serious difficulty in enforcement. The details are as follows.

    Items IdentifiedCriticism
    LineIt is the number of the line, and there is not the problem.
    Component and Function The purpose of indicating a function is to draw failure from it. As a result, failures get mistaken as if failure modes.
    Potential Failure Mode Failures will be listed here by mistake.
  • Since it is original to carry out FMEA at the time of a design, all are potential, and hence, it is more desirable not to use the term "potential."
  • This is useful for the simplification of the format of worksheet.
  • Potential Effect (s) of Failure This column has no problem.
    Severity Degree of lack of measure should be written here, not the degree of severily.
    Potential Cause (s) of Failure This column has no problem.
    Occurrence Degree of lack of measure should be written here, not the degree of probability or frequency.
    Current Controls, Prevention
  • These two columns should be combined. Moreover, since one measure may affect all the severity of effect, frequency of occurrence and, difficulty of detection, all measures should be described in one column.
  • You should not rate degrees of severity of effect, frequency of occurrence, nor difficulty of detection. The objective of evaluation has to be measures (Current Control)and lack of measures should be evaluated.
  • Current Controls, Detection
    Detection Degree of lack of measure should be written here, not the degree of dificulty or detection.
    RPN The numerical value of RPN shows the size of a risk. However, since the success value of RPN is completely unknown, it is unsuitable as an index.
    Recommended Action Since FMEA is a tool for a designer to rate his own design and add measures if necessary, the term "recommendation" nor "responsibility" is not suitable.
    Responsibility and Target Completion Date



    6-1-3 In the Case of the Process FMEA

    (1) Documents for Process FMEA
    For the process FMEA, the process design documents such as below should be prepared. The principle of the quality control that forms the base of these documents is explained next briefly. The explanation about the contents of documents will be shown after that.

    (2) The Principle of Process Management
    A process is comprised of five-M resources, namely, Man, Machine, Method, Material, and Measurement. And these requirements about five-M conditions should be designed to meet the predetermined output requirements.

    The design of a process is constituted in a QCprocess table as follows.

    1. The left half:
      In order to fulfill the 5M resource conditions necessary to produce the prescribed results, provide the relevant specification concretely and in detail. In other words, show how to use 5M resources, to control all of the quality, time-quantity, cost, safety, and environment.
    2. The right half:
      The following matters are specified on the right-half of a QC process table about output quality concretely and in detail.
    The process design should take the form of QC process table. Moreover, in order to manage a process, maintain the 5M conditions and output characteristics, and modify continuously for improvement or correction.


    (3) Factor and Effect Diagram (for Prevention)
    The specifications of process condition comprise two kinds of measures.
    1. Positive Conditions:
      5M details which surely realize the result with the intended characteristics are specified.
    2. Negative Conditions:
      You have to add the measure which prevents the negative characteristic like quality defects to the current controls of each step of the process. For example, it is required to prevent cost rises, time-for-delivery delay, injuries, and environmental pollution besides quality problems.

      The Factor and Effect Diagram (FED) should be created in order to prevent all the problems, and for that purpose, all the factors related to undesirable results should be enumerated.

    Notes
    There are two types of diagrams that are similar to each other. and special attentions are necessary because purposes, methods of creation, and methods of use are totally different from each other.
    • The Factor and Effect Diagram (FED)
      This is used to enumerate the potential causes for the purpose of prevention. Since the troubles is not existing, related work process data do not exist, too. Hence, you must create it in a top-down approach supposing all the troubles and factors which are likely to happen based on knowledge and experience.

      For all factors listed here, it is necessary to take measures accordingly. This makes it possible to design a fully functional process.

    • Cause and Effect Diagram (CED)
      When a trouble occurs, this is created in order to enumerate the candidates suspected as causes in order to perform correction. Since it is the correction purpose, the trouble is existing and related work process data also exist. Hence, you should creates in a bottom-up approach supposing the cause candidate based on data, using knowledge and experience.
    The following diagram indicates that the FED is composed of the same structure as the process table QC.

    (4) Outline of QC Process Table
    A process design is created in the form of the QC process table comprising the following sections:
    1. The section to list bibliographic data (general description),
    2. The section of 5M specifications, which are described for the purpose of creating the prescribed outcome characteristics, and,
    3. The section to prescribe the results and describe how to confirm the results.

    (5) A Case Study of the Process FMEA
    Here is shown the outline of an assembly process of a medicine bottle with a certain fixed number of tablets in it. The screw cap is assembled to a medicine bottle at a fixed torque at the step No.3.

    The QC process table comprises two sections except the general description as follows.

    (a) The Part of QC Process Table for Producing
    The 5M requirements, the check matters and the record method are listed.

    QC Process Table for Producing, Case study
    No. Equipment Man Material Method Check Record
    3 Cap tightenning machine
    (1) Spindle:60 RPM
    (2) Torque: 5kg-cm
    (3) Lubricant:#23
    (4) Put cushions to appointed places to prevent acratch, dent
    (5) Cleaning to prevent dirt

    Product Feeder

    Maintenance ST-124
    (6)
    One-day experience
    Fed from the last step Shown below Time: A1, P1 Form:
    R-41
    Item: (1),(2),(3),(4),(5),(6),(7),(8)
    Method of work
  • Pick up work from preceding step
  • Put work on tightenning machine, switch on
  • Confirm signal buzzar, and take out
  • Put work on feeder
    (7) Wear gloves when touching product
    (8) Bring itemds falling to appointed box

  • (In this QC process table, only a function is designed and the reliability design is not carried out. Click here, in order to refer to the measure which needs to be filled up for reliability.)

    (b) The Part of QC Process Table for Confirming Results
    A detailed description is described in the designated space, about each of below.

    QC Process Table for Confirming Results, Case study
    Spec./Limit Inspections Record Time QA Section
    Tightening torque 5±1 kg-cm Number of samples : 5
    Measure torque by hand torque wrench (10kg-cm)
    Plot data on process capability chart
    Process capability chart A2, P2 0/18 (AOQL=2%)
    Calibration Form : R5 W1
    Dirt,scratch,dent are not acceptable 0/18 (AOQL=2%), visual Form : R4 A2, P2 0/18 (AOQL=2%)


    These two tables mentioned above are standardized into one oblong QC process table.


    (c) Design of Reliability Is Suspended for the Present
    The function of the process that contents vary with instructions violation easily is not maintained well. It is necessary to carry out a reliability design so that the contents of the process are hard to change. Otherwise, workers may often commit an error and the various makeshift measures may be frequently performed by engineers.

    The designer must carry out FMEA to confirm whether the reliability design that oneself performed is appropriate. It goes without saying that reliability must not be insufficient, but it is necessary to avoid an excessive reliability by an excessive cost.

    However, the example of the process design without a reliability design is shown below for study. We transfer to FMEA prior to improving reliability defects, which are corrected after performing this FMEA case study.

    (d) FMEA Worksheet for Process FMEA
    An FMEA worksheet should be created simultaneously with a QC process table by the process designer. The form of Process FMEA worksheet for the absolute evaluation method is shown below. While the designer supplements the weak pointe of the reliability through the FMEA evaluation, this worksheet helps to push forward the design.

    The Worksheet for the Process FMEA
    Process: Medicine Bottle Ass'y Process FMEA Review
    Line Step No.
    Name
    Failure Mode (Violation) Effect Mechanism Current Control (Grounds) a b c RI Action Added a b c RI
    1                            
    2                            
    3                            


    Objection theories

    Robin E. McDermott :
    "For a process FMEA, the team should physically walk through the process exactly as the process flow." (p.25)

    "The best way to create a flowchart is to walk through the process as if you were the thing being processed or created. The process steps should be followed sequentially, and notes should be taken during the walk-through." (p.55)

    Commentary:

    The Phisical "Walk-through" Is Impossible
    Since execution of a process FMEA and creation of a QC process table are performed before actually arranging a process, following the steps of a process physically could not be performed.

    What Is the Failure Mode for a Process?
    Since 5M specifications of the process are not determined yet, listing of failure modes is completely impossible.

    Robin E. McDermott p.25:

    "If a print or flowchart is not available, the team will need to create one prior to starting the FMEA process."
    Commentary:
    There are three great mistakes in the upper description.
    1. The Absence of Design Documents
      It is an unbelievable story, but the above description says as follows. When the designer neglects indispensable design documents, it is instructed that the FMEA team consisting of outsiders should create necessary documents in behalf of a designer. But, the FMEA is to the designer what the ruler is to the carpenter. Since the designer pushes his work when checking reliability by an FMEA as a carpenter works checking dimensions with a ruler, an FMEA team comprised of outsiders must not try to create those documents or cannot create appropriate documents.
    2. Appropriate Documents for Processes Design
      "A print or flowchart" quoted from the upper description is not clear in content because any example is not shown in the whole book. Because the QC process table or the five-M process specifications are not touched at all, Mr. McDermott does not seem to know the meaning of the process design.
    3. Conventional Process Design without Document
      When a virtual process without specific design documents has been generated in a longtime convention, the wrong instruction may often be given that the FMEA team should prepare necessary documents. However, by following reason, these should not be adopted.

      Design consists of function design and reliability design.


    S. Iida, p.88. (A Case of Medical FMEA)
    An FMEA Worksheet for the Gauze Supply and Collection Process
    (O; Occurrence, S; Severity, D; Detection)
    Line Small alignification Unit Duties Failure ModeO Potential Effect S D RPN
    1
    Preparation Suply of gauze Not done 2 Little effect
    (The operation performer will point out)
    1 1
    2
    2
    Preparation Write down number of suplied Not done 2 Little effect
    (After suture, body is checked by X-ray)
    2 1
    4
    3
    Wrong 2 Serious
    (Abdominal pain and fever)
    4 4
    32
    4
    Operation Collected gauze Not done 2 Little effect
    (Start collection)
    1 2
    4
    5
                   
    Commentary:
    Many hospitals have received wrong FMEA instruction and are performing FMEAs using based on the copy of field operations.

    The upper example of a model was quoted from Shuhei Iida.p.88. This is an FMEA case of abdominal surgery for preventing the accident in which a piece of gauze is left behind to a patient's body. And two serious mistakes are pointed out.

    Outline of Worksheet
    Let's try the interpretation of the top three lines of the worksheet mentioned above. It seems to be based on the Four-point Relative Evaluation Methods.

    Line 1: The nurse's duty is to supply the opetation performer with peaces of gauze. Because the opetation performer demands them even if it is forgotten, the effect is negligible. Hence;

    Occurrence (2) × Severity (1) × Detection (1) = RPN (2)

    Line 2: Duty is to write down the number of the supplied gauze. However, it is forgotten. Since record does not exist, there is not other way than checking gauze by X-rays after suture. This method of detection is effective. Hence;
    Occurrence (2) × Severity (2) × Detection (1) = RPN (4)

    Line 3: Duty is to writing down the number of supplied gauze. Howefer the number is wrong. And when the number of the collection accords with this wrong number, an accident happens. The stomachache and heat by the gauze remainder are expected and the effect of the record mistake is serious. Hence;
    Occurrence (2) × Severity (4) × Detection (4) = RPN (32)

    Because similar work and evaluation continue, after that it is omitted, and we have the following questions.
    1. Why X-rays Check Is Not Performed Always?
      As shown in the line 2, if a good evaluation is provided by an examination through X-ray, you should treat all cases likewise.

    2. Reliability Design Has Been Omitted
      This medical process includes a serious problem. As long as this procedure is seen, only functional design fas been is performed and the desin of reliabilityis has not been performed at all. It is clear because the identification columns dof "Current Controls" is not found in the top of the worksheet. This fact shows that it is necessary to learn a process design from a beginning again before describing FMEA.
    Recommendation for Process Design and FMEA:
    The leader of the medical process FMEA should study how to design a process, prior to starting FMEA execution. According to the importance of work, you have to design a process to be hard to generate a human error. Do not think that a poor process design can have sufficient reliability by enforcement of Process FMEA. The duty of FMEA only is the evaluation of reliability and FMEA does not have the power of changing the framework of a process design.

    The Example of a Model Process design
    The gauze delivery and collection process should be designed as follows.

    In the case of this design, become all needless, and there becomes little possibility that the mistake happens. It is right to do an evaluation by FMEA after having done such a reliable design.




    Chapter 6. Seven Steps / Case Study

    Step 2. Listing Failure Modes


    6-2-1 Definition of Failure Mode
    Manners in which a product or process can fail are called failure modes, however, you should identify them in the bottom-up way, on account of the reason shown below.
    1. According to the bottom-up method, failure modes are easily extracted from a product design or process specifications. As for a product, breakages of parts or interfaces are the failure modes, and as for a process, breaches of process instruction are the failure modes. Moreover, it is very rare to overlook those failure modes which are defined as above.
    2. The purpose of the FMEA is to look for all failures and disasters. Among those failures and disasters,some are unknown. Hence, failures cannot be the starting points of an FMEA. The unknown failures and disasters are identified as the result of searching from failure modes as the starting points.
    Otherwise (if a top-down approach is chosen), the following defects will appear as serious problems.
    1. Failure Mode Concept will Be Confused
      It becomes impossible to distinguish a trouble mode from failure, and FMEA comes to a deadlock.
    2. Overlooking will Increase
      When the number of people in the company begins with the president and is counted, the numbers of people to belong to each hierarchy as rank falls down to directors, managers, section managers, employees increase. As above, the progress to the direction where numbers increase is called top-down approach.
      When top-down is adopted for the activity that all matters belonging to a certain range are enumerated, it is easy to produce enumeration omission. Hence, on the occasion of the enumeration of failures and tfailure modes, the top-down approach must not be adopted.

    6-2-2 Failure Mode of Product

    Expert Designers Easily Find All Failure Modes
    The designer of a product can imagine easily through an appropriate effort all future changes that may happen to the components which oneself have adopted. For example, if screws, shafts, bearings, electric wire coils, keys, gears, boards, wire connectors or others are designed or adopted in the design, the designer in charge knows what kinds of change those components may catch in the future in detail.

    Amateurs (Novices)
    The designer who does not know all of these is not a full-fledged designer. Novices confuse failures and failure modes, and cannot point out failure modes easily. They tend to demand performing the brainstorming to make up for the ability lack. However, brainstorming only induces the custom of guessing things without a basis. Furthermore, even by performing brainstorming. beginners cannot do the same work as professional engineers.

    It should be prohibited that outsiders or novices of design hold a brainstorming meeting and that they enumerate a large number of groundless failure modes. Amateurs easily miss important targets of failure modes by firing rifles blindly.

    Starting Points
    The failure mode in a product is defined as the breakage of a component (part or interface) that is the smallest unit of a product.

    Examples of Product Failure Mode
    Part Breakage:
  • Crack,
  • Broken (wires, shafts, screws, pins),
  • Fusion (of metal parts, rubber parts),
  • Corrosion,
  • Abrasion, worn (sliding parts, brake shoes),
  • Transformation,
  • Deterioration (of material), and so on.
  • Interface Breakage:
  • Coming-off, or peeling off (of an assembled or adhered part),
  • Blockage (in a pipe or under a brake pedal),
  • Pulled up (of a plug, a nock pin, or a clamp),
  • Slack (of a screw, a rivet, or a clamping),
  • Adhesion (of liquid, or dirt),
  • Pressure change (inside or outside),
  • Getting with water, and so on.


  • Listing Failure Modes - The Case Study of the Product FMEA
    ( a:Severity, b:Occurrence, c:Detection, RI: Risk Index )
    Product:Gas kitchen range Product FMEA Review
    Line Article or
    Interface
    Failure Mode (Breakage) Effect Mechanism Current Control (Grounds) a b c RI Action Added a b c RI
    1 Introduction hose Cracks                        
    2 Retainer screw Slack Introduction hose coming-off                        
    3 Hose band Ruswt                        
    4 Cracks                        
    5 Hose connector Rust,
    Collapse of shape
                           



    6-2-3 Failure Mode of Process FMEA
    Starting Point to Failures
    Failure modes of a process are defined as a breachs of instructions that are the smallest units to constitute a process, and they are the starting points of approach to failures.

    No Failure Mode without Violation
    There are important cautions. The work error in the process does not correspond to a trouble mode unless it violates the instructions of the process design (i.e., the QC progress table).

    For example, an operator may begin a work without turning on the switch of the machine and this may cause a trouble. However, unless the right action is shown in the written procedure, the operator's action is not a failure mode since there is not a breach. And this problem is not associated with the FMEA at all. It is a problem of a poor functional design of the process.

    On November 05, 2008, Central Japan International Airport was making a great uproar. The security inspector realizes that he was inspecting the passengers without turning on the power switch of security inspection equipment. All the passengers of the airplanes before a takeoff were taken down, and the additional inspections were carried out in a great hurry.

    The security inspection is conducted in order to prevent passengers carrying dangerous objects into the airplane. In this case, the functional design of the process was perfect. However, since there was no reliability design, violation arose easily.

    The inspector had been directed under the rule as follows. Before he begined the security inspection, he had to perform a test of the metal detection performance. However, neither the reliability design nor an FMEA had not been performed. Directions had been able to be broken even if specified by a rule. Furthermore, there had been also no mechanism in which generating of a breach of the rules was displayed.

    Expressing this state of reliability in terms of FMEA, the following measures are insufficient:

    The entrance of conveyer should not be able to open before the metal detection examination is finished. The fact that even an extremely easy fool-proof system which prevents start of the conveyer unless the power supply is turned on is not provided with reveals the fact that FMEA has not been carried out.

    Case Study of Process FMEA
    The failure modes of the case study process are shown below.

    Listing Failure Modes - The Case study of the Process FMEA
    ( a:Severity, b:Occurrence, c:Detection, RI: Risk Index )
    Process: Medicine Bottle Ass'y Process FMEA Review
    Line Step No.
    Name
    Failure Mode (Violation) Effect Mechanism Current Control (Grounds) a b c RI Action Added a b c RI
    1 Step 3
    Cap twisting
    Change of 60 rpm                        
    2 Breach of 5 kg-cm                        
    3 Wrong oil                        
    4 Using up oil                        
    5 Omitting work                        


    Objection Theories

    Robin E. McDermott p.25:
    "Once everyone on the team has understood the process (or product), team members can begin thinking about potential failure modes that could affect the manufacturing process or the product quality.
    A brainstorming session will get all of those ideas out on the table. Team members should come to the brainstorming meeting with a list of their ideas. In addition to the ideas members bring to the meeting, others will be generated as a result of the synergy of the group process."
    Commentary:
    This description is the same as the contents which not only this book but also many other researchers describe. However, it contains three errors.
    1. Impossible without QC Process Table
      This description is about an execution of FMEA under the conditions without creation of a process design document, i.e., QC prcess table. The members who are outsiders of the process design are unable to understand the process design, and it comes to be in overlooking the great portion of the failure modes.

    2. Brainstorming Is Unsuitable
      Brainstorming may be useful as a method of producing a completely new novel idea, when all possible ideas are examined and it does not succeed. It is a method for conceiving freely without being restrained by the established concept, and inducing a novel idea.

      On the other hand, any criticism is forbidden to others' idea and creation of another idea which uses others' idea is recommended. Since the idea created there does not need to be useful, brainstorming is liked by the incompetent. However, it will be very rare to obtain useful fruit probably. The procedure of creating an idea carefully by FMEA later without considering good ideas sufficiently in the stage of a process design is an entire error.

    3. Delay of Duties
      It may take several months for the FMEA to understand the design of a process or product in detail, even if the team members stop their own original duties. The complex duties like designing contain few things which outsiders can easily understand even if the outsiders are researchers of each special field.

      Since the time limit of the design comes before FMEA team reaches a conclusion, the remarkable time delay arises as a serious problem.

    D.H. Stamatis. p.166:
    "For each process function identified you must list the corresponding failure of the function. There can be more than one failure from one function. To help identify the potential failure mode one may think of the negative or loss of the function.(1)"

    Examples include(2)

  • Fails to open
  • Part leaking
  • Broken
  • No pressure
  • Cannot control speed
  • Hole is not round (eccentric)
  • Part undersized, oversized, porous, damaged
  • Another way to identify the failure mode anticipated is by asking the following questions (partial list).(3)
  • How could this process fail to complete its intended function?
  • Why could this part be rejected at this operation?
  • What does the customer find unacceptable?
  • How would the part not conform to specifications at this operation?"

  • Commentary:
    This description is entirely wrong because:
    1. According to the above-mentioned way, failures are derived from functions, and the failure modes are derived in turn from the failure. And "there can be more than one failure from one function" mean the tendency that some failufres may be overlooked. However, the main purpose of FMEA is to anticipate enexpected failures and disasters. Hnece, if you start an FMEA with enumerating functions or failures, unexpected trouble hide behind and these lead to big disasters.
      Correctly, failures have to be derived from failure modes listed beforehand in the bottom-up way.
    2. Examples include function troubles and structural breakages, and these can be failures, manufacturing defectives, structural breakages. Thus concepts are confused.
    3. Those questions should be asked at the time of designing process. If such a thing is taken into consideration in this stage (after a design), you have to reply to the question what to take into consideration at the time of a design. A design is a bunch of measures for trouble solutions.
      Because there can be more than one answer to each question, overlooking occurs easily. This is the defect peculiar to the top-down approach, and hence FMEA should adopt the bottom-up approach.


    Robin E. McDermott: p.44: (A portion of the worksheet has been quoted.)

    Product FMEA of a Fire Extinguisher
    Line Conponent and Function Potential Failure Mode ( Our opininons ) Potential Effect(s) of Failure Severily
    1 Hose; Delivers extinguishing agent Cracks ( ○ Failure mode ) Misfire 10
    2 Pinholes ( Χ Manufacuring defective ) Low discharge pressure 8
    3 Blockages ( ○ Failure mode ) No discharge 10
    4 Canister: reservoir for extinguishing agent Paint coverage uneven ( Χ Manufacuring defective ) Bare spots rust weakening metal; possible explosion 10
               
    9 Charge gauge;determine
    remaining volume of agent
    Inacurate reading ( ΧFailure ) Overfill if gauge reads low; underfill if gauge reads high 10
    10 Broken crystal ( ○Failure mode) Injury to user from cut glass 8
    (The indications in the parenthesis are our opinions.)


    Commentary:
    Looking at the description column of "Potential Failure Modes", it is understood that there is a mixture of failure mode, failure, and manufacturing defective. Naturally, things other than a failure mode must not be described here.




    Chapter 6. Seven Steps / Case Study

    Step 3. Listing Effects and Factors


    6-3-1 Introduction
    In this third step, a potential effect of failure modes is listed in each appointed column.
    6-3-2 Case Study of Product
    (1) The Product FMEA
    The worksheet for the product FMEA with descriptions of effects and their factors (mechanisms) in it is shown below.

    Listing Failure Modes - The Case Study of the Product FMEA
    ( a:Severity, b:Occurrence, c:Detection, RI: Risk Index )
    Product:Gas kitchen range Product FMEA Review
    Line Article or
    Interface
    Failure Mode (Breakage) Effect Mechanism Current Control (Grounds) a b c RI Action Added a b c RI
    1 Introduction hose Cracks Gradual gas leakage Aging                    
    2 Retainer screw Slack Introduction hose coming-off Sudden gas leakage Collision with cooking devices

    Spilling salty broth

    Child's mischief

                       
    3 Hose band Ruswt                    
    4 Cracks                    
    5 Hose connector Rust,
    Collapse of shape
                       


    Line 1:
    Some cracks are generated and raised by aging in an introduction hose, and fuel gas comes to leak from there gradually.

    Line 2 to 4:
    The introduction hose may come off from the hose connector sometime after all. And the fuel gas comes to blow suddenly. Failure modes are put between failures and causes in order to serve to narrow a distance between a failure and causes, and hence these three are usually listed at the same time.


    As shown above, the intervals between (1), (2), (3), (4) and (5) are very short. However, that between (1) and (5) is very long. This is actually the very important role of the failure mode concept. If the role of failure mode is understood, it becomes clear that neither a failure nor a production defective belongs to the failure mode concept.

    Factor and Effect Diagram
    When there are various factors, it is useful to create a Factor and Effect Diagram (FED) about every effect in order to prevent overlooking of factors.

    (2) Process FMEA
    The process FMEA worksheet in which a descriptions fills in the space of "effect and cause" column is shown below.

    Listing Failure Modes - The Case study of the Process FMEA
    ( a:Severity, b:Occurrence, c:Detection, RI: Risk Index )
    Process: Medicine Bottle Ass'y Process FMEA Review
    Line Step No.
    Name
    Failure Mode (Violation) Effect Mechanism Current Control (Grounds) a b c RI Action Added a b c RI
    1 Step 3
    Cap twisting
    Breach of 60 rpm Slack of screw cap Human error                    
    2 Breach of 5 kg-cm                    
    3 Wrong oil                    
    4 Using up oil                    
    5 Omitting work                    

    Line 1 to 5:
    All the failure modes are brought about by the same factor of a human error, and bring about the same effect called slack of a screw cap. In fact, not only the case where the tightening is loose but when too tight, it should be taken into consideration, but it has been simplified on account of space.


    6-3-3 Meaning of Effect
    Issues in this step are as follows.

    (1) Difinition of Effects
    The disadvantageous result which stakeholders including customers receive is called Effect in FMEA terms. (2) Multiple Effects
    When there are several independent effects that are not suitable to be described in the same space for a failure mode, they are divided into separate mention spaces of the worksheet if given different measures. For example, the abrasion of a bearing in a instrument will lead to following effects. If same measures are given to these failure modes, the measures can be described in the united explanation space. Supposing these failure modes are given separate measures respectively, the measures are indicated to a separate space respectively.

    (3) Tandem Chain
    When the tandem chain continues so that failure mode A may generate another failure mode B and B may generate another failure mode C, and so on, where is the boundary of failures mode and effects?

    The example of the electric washing machine is shown below. This chain seems to show many effects, but, in fact, the chain of the failure mode merely exists towards one final effect. Of these, only changes to occur in parts and interfaces of the product are the failure modes.

    Factor (A) : Aging of the water tank gasket,
    Tandem chain of failure modes (B) : Gasket cracks,
    (C) : Leakage of liquid detergent (C is not a failure mode, but a function trouble),
    (D) : A drop of detergent attachment to the motor wires,
    (E) : Penetration from cracks of electric wire covering,
    (F) : Adhering to core wires,
    (G) : Penetration inside the wire by capillary action,
    (H) : Corrosion of the electric wire,
    ( I ) : Breakage of electric wires by vibration.
    Effect (J) : Outbreak of the sparks by the vibration, which is   the initial effect.
    (K) : Heat generation by the spark,
    (L) : Melting of the coating,
    (M) : Generation of flammable vapors,
    (N) : Ignition by the apark,
    (O) : Generation of a fire, which is   the final effedct.

    Where to take measures
    This pursuit ends at the stage where it becomes able to evaluate the potential worst effect. In this example case, evaluation of effect is terminated when some detergent liquid reaches an electric component area where the potential worst effect can be anticipated. And measures should be taken for the failure mode not to reach B or at the latest D.

    Determine the effect
    When a certain failure mode x leads to its worst effect y, y is an effect if it is a function trouble or a disaster which is worth evaluating. If y is not worth evaluaing, it is one of the trival failure modes and can be neglected. When many failure modes are derived in a queue, this is called a chain-reaction (a billiards-phenomenon, in Japan) of failure modes.


    6-3-4 Meaning of Factor

    In the field of management technology, when a phenomenon may cause a specific result, it is called a potential cause or a factor of the specific result or effect. More than two factors can coexist about one effect.

    However, factos and result have to have the quality to be controled by a company. Neither a natural phenomenon nor a social phenomenon cannot be a factor of a specific result, if it cannot be controled by a company.


    6-3-5 Concept of Cause

    When a certain potential result is realized by one of the potential causes, only this cause is identified as a true cause. Other potential causes are still potential.

    (1) A Big Disaster in Japan
    On March 11, 2011, a big earthquake and tsunami occurred in East Japan, and it attacked the Fukushima Japan Nuclea Power Plant of TEPCO (Tokyo Electric Power Co.). After the automatic stop of the nuclear reaction, as an unexpected accident, the electric power supply necessary for cooling the fuel stopped entirely and hydrogen generation and explosion happened. This explosion began to blow a large quantity of radioactivity and caused a serious social problem.

    It was well known among geologists since ancient days that a major earthquake and huge tsunami as big as those of this time attack this district every 1,000 years. And a geologist clarified after the construction of the nuclear power plant, the fact that a tsunami of more than 10 meters above sea level was generated about 1,200 years ago, in an investigation done 10 years before the tsunami day.

    Was the earthquake or the tsunami a cause of the disasters? The answer may changes by how an answering person was concerned with this problem. Geologists may answer that the tsunami was the cause, however from the viewpoint of management, it is not correct.

    1. It was a ridiculous story, but Nuclear Safety Commission of Japan (NSCJ) announced since before the construction of the nulear power plant that risk management should be performed by each electric co. independently and NSCJ could not decide what to do.

    2. TEPCO built the nuclear power plant so as to endure a tsunami up to 5 meters in height saying that this height was decided in consideration of the record-high tsunami in this district having margin.
      TEPCO itself had predicted a tsunami of over 10 meters in height by computer simulations three years before this earthquake, but did not report either to NSCJ nor to the nation, for the reason that it was an uncertain fact.

    3. The media reported an above-mentioned fact after the disaster, but hardly reported anything before the disaster.
    In consideration of the above-mentioned situation, will it be proper to think that still it is caused by an earthquake? All that matters here is not a physical factor. From the viewpoint of management, the cause of collapse of a nuclear power plant is as follows.
    1. Negligence of NSCJ,
    2. Negligence of TEPCO, and
    3. Negligence of media.
    The major earthquake and the tsunami are natural phenomena without the measures to prevent and should not be understood as a cause in the field of the management technology.

    If it is judged that the cause of collapse of the nuclear power plant is the tsunami, the cause of it is the earthquake, and the cause of it is the continental drift, and the cause of it is the convection inside the earth, and the cause of it is the astronomical collision at the time of the earth generation, and it continues in a similar manner hereafter. We will enter deeply the world of geophysics and astronomy unrelated to management.

    Human Errors Are Inevitable
    About a problem caused by an operator's error in a process, the way of thinking mentioned above is important. When an error occurs because you do not carry it out enough though measures are relatively easy, the neglect of the measures is the cause of the error. This is because it is well-known that anyone commits errors.

    (2) An Example of Educational Issue
    In 2006, finger cutting accidents of children with the shredders occurred in succession in Japan. What the cause of this kind of accident was is the problem. A teacher of a certain elementary school thinks that the main cause of the accident was to be inferior in the discipline of the parents to children. This teacher considers it to be a problem of the education responsibility, rather than a problem of the product security.

    Neglect of Measures Was the Cause
    From the standpoint of product safety, the accidents were caused by the fact that any of well known easy measures had not been carried out in the designs of shredders. Because the measures need cost, it is not proper to require measures unconditionally. A prevented disaster should be compared with the cost for measures, and it is necessary to determine how much price hike consumers expect the safety measures in exchange for.

    A legal standard or regulation is desirable in this connection.

    (3) A Case of a Subway Derailment
    On March 8, 2000, a derailment of a subway in Tokyo occurred at a sharp curve near a station. It was the derailment of the type that a flange of a wheel raised its body on the rail when going slowly at a sharp curve. It had been known that this type of derailment is relatively easy to occur when the jet of the lubricant to the flange doesn't go well.

    What was the Cause in Management?
    In this case, was the low speed one of the causes of the derailment? If it was so, the measures to prevent it must have been carried out. However, such a way of thinking is impractical. The low speed was doubtless one of the causes of the derailment from the viewpoint of physics. However the causes should be considered from the viewpoint of management, because the place of derailment was very close to the station and the low speed was inevitable.




    Chapter 6. Seven Steps / Case Study

    Step 4. Listing Current Controls


    6-4-1 Meaning and Purpose
    About an FMEA worksheet for a kitchen gas range shown below for example, what is the meaning and purpose of Current Controls ?

    Let's evaluate the failure mode of this example, to answer the above-mentioned questions. The next evaluation standard for rating of the current controls is applied to the "severity" of effect. The score indicates the necessity of additional measures to the current controls about the severity of this effect. The measures provided in the current design control are as shown in the table below.

    General Evaluatuon Criteria
    ScoreDedree of Control
    4Insufficient very much
    3Insufficienth
    2Approximately enough (the best)
    1Perfect

    A problem for beginners is shown below.

    Which score from 1 to 4 is appropriate in each of spaces x and y shown below?

    ( a:Severity, b:Occurrence, c:Detection, RI: Risk Index )
    Product:Gas kitchen range Product FMEA Review
    Line Article or
    Interface
    Failure Mode (Breakage) Effect Mechanism Current Control (Grounds) a b c RI Action Added a b c RI
    1 Introduction hose Cracks Gradual gas leakage Aging   x y              

    About Severity, a large majority answers that it is x=3 for the following reason. Since a gas leak is inevitable after prolonged use, it is not x=2. But since aging of the current high quality hose is slow, it is not x=4, either. Also about Occurrence, a large majority shows the similar reason and answers that it is y=3.

    They do not pay attention to explanation space of the Current Controls being a blank. It is evidence of not understanding the object of the evaluation. The problem above is asking which score is appropriate to be given to the current controls of the severity and occurrence. It is the insufficiency (or necessity) of measures to ease the severity of the effect in this problem case. However, the evaluation is impossible because the description place for current controls is blank, and object of evaluation does not exist.

    The absence of the current controls means that of the reliability design. If measures are unnecessary, it means no harmful effect are produced and the failure mode should be eliminated from the worksheet.


    6-4-2 Process of Reliability Design
    An expert designer performs a design of function and reliability simultaneously, so that both go successfully. In the product design, a variety of means are planned like strength calculations, testing with prototypes, computer simulations, and durability tests to ensure the reliability.

    As for the process design, various measures are taken like an fool-proof system, a fail-safe system, an auto-alert, a color-marking, fixation of setting, automatic inspection equipment, prototype production, pilot production, and so on.

    When such preventive actions are planned for the reliability before the design, these measures are called Current Controls.


    6-4-3 Case Study of Product FMEA
    Line 1:
    Annual inspection of the introduction hose was adopted as the measure against cracks due to aging.

    Listing Current Controls of the Product FMEA
    ( a:Severity, b:Occurrence, c:Detection, RI: Risk Index )
    Product:Gas kitchen range Product FMEA Review
    Line Article or
    Interface
    Failure Mode (Breakage) Effect Mechanism Current Control (Grounds) a b c RI Action Added a b c RI
    1 Introduction hose Cracks Gradual gas leakage Aging Annual check

    Stainless steel cover

                     
    2 Retainer screw Slack Introduction hose coming-off Sudden gas leakage Collision with cooking devices

    Spilling salty broth

    Child's mischief

                     
    3 Hose band Ruswt                  
    4 Cracks                  
    5 Hose connector Rust,
    Collapse of shape
                     
      ↑ These measures should be listed during the stage of the reliability design.


    Line 2 to 5:
    Annual inspection of these failure modes was planned. In addition, child mischief was thought about depending on home and, in consideration of the seriousness of the result, it was decided to use the cover made of stainless steel to cover up the whole. As a result, the introduction hose of line 1 also was covered together.


    6-4-4 Case Study of Tablet Bottle Cap
    Line 1:
    The gearshift lever of the twisting machine was planned to be fixed by means of screws or a knock pin not to move and the speed arrangement of 60 rpm will never change. This failure mode may occur, for example, when a worker temporarily changes the selector position for the cleaning of the machine and forgets to be returned carelessly afterwards, or when somebody may touch the lever for some purpose.

    Line 2:
    This failure mode may occur mostly in cleaning or repair mending of the machine. The setting of the torque was planned to be fixed by means of screws or a knock pin.

    Listing Current Controls of the Process FMEA
    ( a:Severity, b:Occurrence, c:Detection, RI: Risk Index )
    Process: Medicine Bottle Ass'y Process FMEA Review
    Line Step No.
    Name
    Failure Mode (Violation) Effect Mechanism Current Control (Grounds) a b c RI Action Added a b c RI
    1 Step 3
    Cap twisting
    Breach of 60 rpm Slack of screw cap Human error Fixation of the gearshift lever                  
    2 Breach of 5 kg-cm                  
    3 Wrong oil Color marking                  
    4 Using up oil Automatic oil dropper                  
    5 Omitting work Feeder starts with the processing completion signal                  
      ↑ These measures should be listed before rating a, b, or c.

    Line 3:
    The color identification method was planned in order to prevent use of wrong lubricating oil. The cans, oilers, oilcups, and nipples of oil of the same grade were planed to be painted in the same color, and were mutually distinguished using the colors of the traffic light.

    Line 4:
    In order to prevent exhausting oil, attachment of an automatic oil dripper was planned.

    Line 5:
    In order to prevent the accident that the worker does not tighten the screws to the end, it was planned to install an error-proof system into the feeder, which catches the signal that is emitted when the screw cap is tightened to the end, and starts sending the product to the next step automatically.

    Memorandum for Correction
    Elementally, these measures should have been planned at the step of the reliability design. In the case that these measures are added as a result of the design review, however, these correction plans are temporarily written briefly on a copy of the QC process table in red for a memorandum as shown in the examples below with the mark "✽", and then concrete measures such as equipment improvements are implemented and finally the QC process table should be corrected appropriately.

    QC Process Table for Producing, Case study
    No. Equipment Man Material Method Check Record
    3 Cap tightenning machine
    (1) Spindle:60 RPM
    (2) Torque: 5kg-cm
    (3) Lubricant:#23
    (4) Put cushions to appointed places to prevent acratch, dent
    (5) Cleaning to prevent dirt
    * Gear shift levewr to be fixed
    * Lubricant auto-drippig
    * Unification and color marking about oil

    Product Feeder
    * Erro-proofing against omission
    Maintenance ST-124
    (6)
    One-day experience
    Fed from the last step Shown below Time: A1, P1 Form:
    R-41
    Item: (1),(2),(3),(4),(5),(6),(7),(8)
    Method of work
  • Pick up work from preceding step
  • Put work on tightenning machine, switch on
  • Confirm signal buzzar, and take out
  • Put work on feeder
    (7) Wear gloves when touching product
    (8) Bring itemds falling to appointed box

  • Click here to refer to the original QC process table.


    6-4-5 Purpose of Describing Current Controls

    (1) Object of Evaluation
    The FMEA involves evaluating insufficiency of measures against each failure mode in terms of severity of effect, frequency of occurrence and difficulty of detection. Most of the researchers and university teachers lecture that it is necessary to evaluate the severity of effect, the frequency of occurrence, and the difficulty of detection, not to evaluate the degree of necessity of additional measures for these.

    That lecture, however, has the error which a person of a slight work experience is apt to fall into. Do not evaluate the severity of effect, the frequency of occurrence and the difficulty of detection, without consideration of the Current Controls. The objects of evaluation is the degree of Current Controls against those. The problem is related whether Current Controls are enough.

    (2) How to Decide Current Controls
    As for a reliability design of a product, measures should be taken so that :

    As for a reliability design of a process, measures should be taken so that :

    In most cases, one measure can be effective to any of three elements, that is, severity of effect, frequency of occurrence, and probability of detection. For example, an error-proof device prevents errors, and this contributes to improve all of the three elements.

    (3) Number of Column for Current Controls
    Description space for current controls may seem to be necessary to Severity, Occurrence and Detection respectively. There are some researchers who insist that blanks of "Current Controls" for "Severity" are not necessary. As a reason, they seems to think that Severity of the effect of a certain failure mode is decided, and is not influenced by the measure.

    However, this is an error by the people who have not acquired enough experience. By the above-mentioned reason, one description space is enough. Fail-safe technique exists to reduce influence even if a failure mode rises. An example of the worksheet based on this wrong theory is shown below.

    (Quated from Robin E. McDermott.p.44)

    Well experienced engineers will probably think as follows:

    Objection Theories

    Robin E. McDermott p.31:
    "The severity ranking is an estimation of how serious the effects would be if a given failure did occur. In some cases it is clear, because of past experience, how serious the problem would be. In other cases, it is necessary to estimate the severity based on the knowledge and expertise of the team members."
    Commentary:
    This opinion is not acceptible.
    1. This description will probably give basic misunderstanding to many engineers. It is not the right design procedure that a designer omits the reliability design which must be performed essentially, and evaluates the risk in each failure mode by FMEA based on someone's experience and technical knowledge.
    2. If it is necessary to estimate the Severity based on the knowledge and expertise that the team members have, the designer should have received instruction from the members during the design work.
    3. In other words, according to the opinion mentioned above, this FMEA team has to play a role of the designer. Probably the abolition of this kind of team will become the big one step to restore the function that a designer should originally have.
    The boss has to superintend a designer, in order to prevent each designer's negligence and carelessness. Furthermore, you should promote the design by teamwork. On the other hand, you have to abolish the FMEA team as soon as possible.




    Chapter 6. Seven Steps / Case Study

    Step 5. Evaluating Three Elements


    6-5-1 Common Criteria
    Measures which constitute current controls are evaluated from three viewpoints. The evaluation is carried out based on the four-point scale common criteria shown below.

    Common Evaluatuon Criteria
    ScoreDedree of Control
    4Insufficient very much
    3Insufficienth
    2Approximately enough (the best)
    1Perfect

    There is an exception to the above-mentioned general standard. The exceptional standard is applied to the evaluation of "Detection", which will be explained in the section 6-5-4.

    Objection Theories

    Takayoshi Narimatsu:
    A certain Japanese (Takayoshi Narimatsu) comments as follows on his homepage.
    "An FMEA is performed to narrow down subjects that need measures. If measures are to be performed for all problems, the ranking will be unnecessary. In that case, measures must be performed for all of factors included in the cause and effect diagram, and there is no need for an FMEA. To obtain the maximum effect of limited resources such as time and money for solving problems, the prioritization of the problem is inevitable. FMEA is the tool for that purpose."
    Commentary:
    This opinion accepted by some people, however, should be denied by the three reasons as follows.
    1. A Few Problems for Experts
      Amateur designers are apt to do only a functional design without a reliability design. They use some screws to attach a component, but do not almost consider the reliability problems ( such as the rust or slack of screw which may occur later). Thus their designs are inclined to include a lot of defects.

      In the design by a specialist, all the detailed reliability is examined from one screw even to one pin. Hence, experts do not adopt the policy of taking means according to the priority of a defect.

    2. Measures should Be Taken If Necessary
      Experts perform measures for all of factors included in the factor and effect diagram (FED), as far as the defect might hurt reliability and cause a cost that cannot be ignored later.
    3. Prioritization Is Meaningless
      The people who felt a sign of the influenza infection will go to the hospital and have an examination. Suppose that the results of medical examinations for a certain patient are as follows. In such a way, the patient can know neither their illness nor the necessary measures. Do you ever have looked at such a hospital?

    Robin E. McDermott p.37:
    "The team must now decide which items to work on. Usually it helps to set a cutoff RPN, where any failure modes with an RPN above that point are attended to. Those below the cutoff RPN are left alone for the time being. For example, an organization may decide that any RPN above 200 creates an unacceptable risk. This decision sets the cutoff RPN at 200."
    Commentary:
    There is not the explanation at all about the base of the cutoff point. How could the cutoff RPN decide at 200? In other words it is a poor theory that everything is decided groundlessly. As for this turning point, the Absolute Evaluation Method (AEM) is rational and should take the place of the conventional Relative Evaluation Method (REM).

    6-5-2 Evaluating Measures for Severity

    Note for evaluating Effect
    Under the assumption that the failure mode occurs in spite of the measure indicated in the Current Controls, evaluate the insufficient grade of the measure which eases the severity of the result predicted.

    (1) Case Study of the Product FMEA
    Line 1:
    Under the assumption that aging and cracks occurs in spite of the measure indicated in the Current Controls (that are annual check, and stainless steel covering), this failure modes (that are aging and crack) do not grow quickly. Since it is thought that a big problem does not arise even if an oversight of very shallow cracks arises several times, the annual check with the naked eye is enough as a measure. Since it is expected that the result is light even if failure mode occurs in this state of Current Controls, 2 is a suitable score of the Severity.

    Rating Severity of the Product FMEA
    ( a:Severity, b:Occurrence, c:Detection, RI: Risk Index )
    Product:Gas kitchen range Product FMEA Review
    Line Article or
    Interface
    Failure Mode (Breakage) Effect Mechanism Current Control (Grounds) a b c RI Action Added a b c RI
    1 Introduction hose Cracks Gradual gas leakage Aging Annual check

    Stainless steel cover

    2                
    2 Retainer screw Slack Introduction hose coming-off Sudden gas leakage Collision with cooking devices

    Spilling salty broth

    Child's mischief

    4                
    3 Hose band Rust                
    4 Cracks                
    5 Hose connector Rust,
    Collapse of shape
                   
      ↑ These measures should be listed before rating a, b, or c.

    Line from 2 to 5:
    As for Line from 2 to 5, the failure modes (rust, slack, cracks, and collapse of shape) is expected to grow slowly. But the failure mode of coming-off of the hose occurs suddenly and the expected result is serious. Hence, measures to ease the Severity of the effect are very lack, and score 4.

    Relations between "Severity" and "Current Controls"
    If the Severity of the effect is defined as the impact taking place when the failure mode occurs, that is inclined to be misunderstood as unrelated to the current controls.

    The gas range includes an introduction hose, in which cracks caused by aging are identified as a failure mode. The current controls include conducting an inspection of the hose once a year, and this seems to be enough to detect the crakks before growing to a dangerous level, should it occur.

    However, when the design controls are changed into an inspection that is conducted every five years, the effect should present different severity from the case of an inspection once a year. The depth of cracks, the quantity of the gas leakage, and Severity of the effect become different considerably. In the case of the failure mode of slow advance, generally, Severity changes by measures. It is necessary to evaluate measures to describe in the column of Current Controls in consideration of this point.


    (2) Case Study of the Process FMEA

    Rating Severity of the Process FMEA
    ( a:Severity, b:Occurrence, c:Detection, RI: Risk Index )
    Process: Medicine Bottle Ass'y Process FMEA Review
    Line Step No.
    Name
    Failure Mode (Violation) Effect Mechanism Current Control (Grounds) a b c RI Action Added a b c RI
    1 Step 3
    Cap twisting
    Breach of 60 rpm Slack of screw cap Human error Fixation of the gearshift lever 4                
    2 Breach of 5 kg-cm                
    3 Wrong oil Color marking 4                
    4 Using up oil Automatic oil dropper 4                
    5 Omitting work Feeder starts with the processing completion signal 4                
      ↑ These measures should be listed brfore rating a, b, or c.

    Line from 1 to 5:
    The outbreak of the failure modes leads to the slack of the screw cap. It is unlikely to cause a disease but it seems not appropriate to leave untouched. Then, the severity scores "3" from this aspect.

    However, this score may be criticized, because the moisture absorption of the tablet may spoil the product value, the recall loss is expected to be vast, and the company may suffer an enormous damage by rumors. Then, the severity should score 4, which indicates that the measures taken in the current controls are very short.


    Objection Theories

    Robin E. McDermott p.31:
    "The severity ranking is an estimation of how serious the effects would be if a given failure did occur. In some cases it is clear, because of past experience, how serious the problem would be. In other cases, it is necessary to estimate the severity based on the knowledge and expertise of the team members."
    Commentary:

    Indication of the Basic Problem
    At first, Severity will surely score 10 from the possibility of crash, assuming that the failure mode of a crack arises on the wings of a passenger plane first. Next, how is evaluation of Severity about the crack assumed to produce in the paper cup for passengers' meal?

    The severity ranking of a crack will score as follows, when the ten-point method of relative evaluation estimates.

    If ranking is determined by the seriousness of effect itself, the airplane may be made strongly. In contrast, goods like a paper cup, an earphone, a radio set or a paper cup will be treated as a trifling problem. However, is the crack of the paper cup trifling or serious to the user of the paper cup, or the paper cup maker?

    May a paper cup be designed rudely? The paper cup has to be designed and manufactured as it should be. The same can be said also about other goods. As a result, evaluation criteria have to be determined for every product, and it becomes impossible to obtain the criterion common to all the products.

    The Correct Understanding
    The above-mentioned problem is certainly solved by the Absolute Evaluation Method.

    An evaluation of Severity should be the evaluation of insufficiency of the Current Controls to ease it. The Severity evaluation of the cracks of the plane is to evaluate a shortage of the current management for it. And similarly, the Severity score of a crack of a paper cup is an evaluation of how insufficient the current controls to relax it. And hence, you must not evaluate how serious effect would be and must not compare the seriousness of the result between products.

    Kenneth W. Dailey p.22:

    "Utilizing the Failure Mode Effect description and the applicable Severity Criteria, assign and record a Severity Ranking. Remember this is an art not a science. There is no Correct Ranking."
    Commentary:
    The upper explanation is frank testimony about the defects of Relative Evaluation Method. However, still why is it that he maintains the wrong method?

    Kenneth W. Dailey p.29:
    A ranking standard for Severity is quoted below.

    Commentary:
    Let's examine the usefulness of this standard . Even if anything happens in a paper cup for plane passenger's meal services according to this standard, the worst marking is 6. Depending on this standard, the degree of the lack of measures cannot be evaluated. Correctly, the degree of the lack of measures for the paper cup crack has to be evaluated at 10 if measures are short very much. This problem is solved very naturally by Absolute Evaluation Method(AEM). The thought that function and reliability must be designed appropriately whatever kind of product it may be is received generally.

    6-5-3 Evaluating Measures for Occurrence

    (1) Case Study of the Product FMEA

    Line 1: The score of occurrence is b=2.
    Checks performed once a year for the introduction hose have been planned as described in the blank of the Current Controls. The cracks of the hose do not grow deep fast. And the gas leakage of a serious level would not be generated even if slight cracks passed the check by mistake several times.

    Line from 2 to 5: The score is b=1.
    As described in the current controls column, a stainless cover has been installed to protect things around the introduction hose from the collision with the kitchen utensil, from the overspill drops of salty broth and from the child's mischief.

    Rating Occurrence of the Product FMEA
    ( a:Severity, b:Occurrence, c:Detection, RI: Risk Index )
    Product:Gas kitchen range Product FMEA Review
    Line Article or
    Interface
    Failure Mode (Breakage) Effect Mechanism Current Control (Grounds) a b c RI Action Added a b c RI
    1 Introduction hose Cracks Gradual gas leakage Aging Annual check

    Stainless steel cover

    2 2              
    2 Retainer screw Slack Introduction hose coming-off Sudden gas leakage Collision with cooking devices

    Spilling salty broth

    Child's mischief

    4 1              
    3 Hose band Ruswt              
    4 Cracks              
    5 Hose connector Rust,
    Collapse of shape
                 
      ↑ These measures should be listed before rating a, b, or c.


    (2) Case Study of the Process FMEA

    Line 3: The score is b=2.
    Lubricating oil is provided with many grades. However, usually three grades are enough in the same workshop. All the oil canisters, the oilers, nipples, the oil supply cups and so on are distinguished by painting in the three traffic light colors. This grade distinction is effective to prevent mistakes.

    Rating Severity of the Process FMEA
    ( a:Severity, b:Occurrence, c:Detection, RI: Risk Index )
    Process: Medicine Bottle Ass'y Process FMEA Review
    Line Step No.
    Name
    Failure Mode (Violation) Effect Mechanism Current Control (Grounds) a b c RI Action Added a b c RI
    1 Step 3
    Cap twisting
    Breach of 60 rpm Slack of screw cap Human error Fixation of the gearshift lever 4 1              
    2 Breach of 5 kg-cm              
    3 Wrong oil Color marking 4 2              
    4 Using up oil Automatic oil dropper 4 1              
    5 Omitting work Feeder starts with the processing completion signal 4 1              
      ↑ These measures should be listed before rating a, b, or c.

    Other Lines: The scores are b=1.
    Measures taken in the Current Controls will prevent the failure modes nearly completely.


    The Objection Theory

    Kenneth W. Dailey p.23:
    Step7.- Assign Occurrence Ranking

    "From the Cause Description, the group's knowledge and experience, and the applicable Occurrence Ranking Criteria, assign and record an Occurrence Ranking."

    Commentary:
    1. The Object of Evaluation Is Wrong
    According to this opinion which is full of mistakes, assignment of Occurrence Ranking is performed depending on the three things shown below. The question which many people bear about this description is not touching on the Current Controls description. Correctly, the most important key point in evaluation is that the measure described in the Current Controls is the object of evaluation.

    Recomended Rankings for FMEA Occurrence (Design)
    Ranking Description Failure Rate
    10 Very High Failure is a Continuous Problem P 20%
    9 P 10%
    8 High Frequent Failure P 5%
    7 P 2%
    6 Moderate Sporadic Failure P 1%
    5 P 0.2%
    4 P 0.04%
    3 Low Relatively few Failure P 0.0067%
    2 P 0.0007%
    1 Remote Failure is unlucky O 0.0007%
    (Quaterd from Kenneth W. Dailey p.31)
    1. The Criteria are Useless for Evaluation
    It is quite impossible to perform rating according to these criteria. "FMEA is an art, but not a science" is Mr. Kenneth W. Dailey's thought. The FMEA here refers to the Relative Evaluation Method. These criteria cannot hold rational contents essentially. For example, according to this standard, when the failure rate of the crack of a passenger plane is 0.04%, evaluation is set to b= 4, but if common sense is followed, you have to give b=10. On the other hand, as for the crack which arises in the paper cup for a passenger's meal, if a failure rate is 0.04%, evaluation also looks appropriately by b= 4.

    In other words, evaluation should be done not by frequency, but by the degree of necessity for adding measures to improve frequency.

    Robin E. McDermott p.36:

    "The best method for determining the occurrence ranking is to use actual data from the process. This may be in the form of failure logs or even process capability data. When actual failure data are not available, the team must estimate how often a failure mode may occur. The team can make a better estimate of how likely a failure mode is to occur and at what frequency by knowing the potential cause of failure. Once the potential causes have been identified for all of the failure modes, an occurrence ranking can be assigned even if failure data do not exist.
    Commentary:
    Several mistakes are pointed out as below.
    1. Process Data do not Exist
      Before a process design and FMEA are completed, the actual process does not exist yet, and hence actual data are not available generally.

    2. Do not Use Actual Data
      You should not always evaluate based on actual data, even if a similar process exists, because most human errors happen in the operation where mistakes have not happened for many years. For example, it is clear from the facts of the driver's mistakes in the recent railroad derailment rollover accidents which occurred in Japan and Switzerland.

    3. Measures Are to Be Evaluated
      The object of the evaluation is measures in the Current Controls description as a result of reliability design. You should not evaluate outbreak frequency of the failure without considering Current Controls description.

    4. Applying Frequency Ranking Is too Difficult
      It is too difficult to apply frequency ranking from 1 to 10 from knowledge and experience based on a standard shown above, when, in particular, a new product or a new process is designed.

    6-5-4 Evaluating Measures for Detection

    (1) Outline
    The general meaning of evaluation of Detection in the Absolute Method is a judgment whether the means of detection in the Current Controls description are enough or insufficient. Be aware that the object of detection varies according to circumstances of the product or the process. Measures in the Current Controls described obtain a evaluation score corresponding to the common criteria as shown below.
    Common Evaluatuon Criteria
    ScoreDedree of Control
    4Insufficient very much
    3Insufficienth
    2Approximately enough (the best)
    1Perfect

    In the Absolute Evaluation Method, a very convenient Exceptional Standard is applied to Detection. Evaluation b=2 leads to estimation c=2 automatically, and b=1 leads to c=1 automatically again.
    The theory which draws the above-mentioned conclusion is shown below.

    For the question of whether a firebrick burns at temperature of 400 degree C, you should estimate it as b=1. In such a case as an essential nature is the basis, no measures are described in the Current Controls. If measures are unnecessary, however, it is necessary to describe it and its reason.

    Detection is evaluated at the same time as c=1, that is, the combustion examination is unnecessary. If the judgment that the firebrick does never burn is based on enough grounds, the verification is accepted as unnecessary.

    Exceptional Criteria
    Occurrence Dtection
    4
    4
    3
    2
    1
    3
    4
    3
    2
    1
    2
    2
    1
    1

    This exceptional standard facilitates FMEA practice very much. However, in the case that occurrence c is 3 or 4, the evaluation for the detection should be based on the common standard.

    (2) Evaluating the Measures for Detection in the Product FMEA

    1. Measures for Detection
      In the case of a product design, the following typical means for the detection of failures are often performed. In late years computers have come to help strength calculations and the production of resin trial products remarkably.

      Opportunity of Detection
      In many cases, detection of defecst in a product is performed during the design controls, that is, before submitting of drawings or specifications. Sometimes, periodic inspections of the product by the user or a specialized supplier may be planned.

    2. Case Study of the Product FMEA
      In the worksheet shown below, an equal score is given for Occurrence (b) and Detection (c). Here, you can watch the usefulness of a special standard.
    Line 1: The score (c) is 2, in accordance with the occurrence score (b).

    Line 2 to 5: The score (c) is 1, in accordance with the occurrence score (b).

    Rating Detection of the Product FMEA
    ( a:Severity, b:Occurrence, c:Detection, RI: Risk Index )
    Product:Gas kitchen range Product FMEA Review
    Line Article or
    Interface
    Failure Mode (Breakage) Effect Mechanism Current Control (Grounds) a b c RI Action Added a b c RI
    1 Introduction hose Cracks Gradual gas leakage Aging Annual check

    Stainless steel cover

    2 2 2            
    2 Retainer screw Slack Introduction hose coming-off Sudden gas leakage Collision with cooking devices

    Spilling salty broth

    Child's mischief

    4 1 1            
    3 Hose band Ruswt            
    4 Cracks            
    5 Hose connector Rust,
    Collapse of shape
               
      ↑ These measures should be listed before rating a, b, or c.


    (4) Means for Detection in the Process FMEA
    What are the object of detection in a process FMEA? Let's consider the process of manufacturing a shaft with strict allowable width of surface roughness with a lathe.

    You must perform the exchange of the tool every constant numerical processing to maintain surface roughness constantly. In this case it corresponds to a falure mode not to perform tool exchange at fixed time. Next is considered as measures.

    There are various detection means according to process specifications. For example, the pilot inspection done prior to running a process everyday, the automatic alarm systems against the abnormal process conditions, the product sample inspections of each step and the shipment inspection, etc.

    The means for the detection of the failure modes are decided in a process design, and the detection is carried out during the preparation for running and running of the process. The objects of detection (e.g., a factor, its sign, a failure mode or its initial effect) and the detection means should be concretely decided at the stage of the process design.

    If prevention of a human error is required, "error-proof system" is very effective and recommended.

    Breaches of a user's manual which may be generated after a product is handed to a user can be predicted easily. Furthermore, it is necessary to take suitable measures about breaches which may lead to a serious danger.

    (5) Case Study of Process FMEA

    Line 1 to Line 5:
    The detection score (c) is from 1 to 2 in accordance with the occurrence score (b)respectively. Such handling as above extremely accelerates FMEA process.

    Rating Occurrence of the Process FMEA
    ( a:Severity, b:Occurrence, c:Detection, RI: Risk Index )
    Process: Medicine Bottle Ass'y Process FMEA Review
    Line Step No.
    Name
    Failure Mode Effect Mechanism Current Control (Grounds) a b c RI Action Added a b c RI
    1 Step 3
    Cap twisting
    Breach of 60 rpm Slack of screw cap Human error Fixation of the gearshift lever 4 1 1            
    2 Breach of 5 kg-cm            
    3 Wrong oil Color marking 4 2 2            
    4 Using up oil Automatic oil dropper 4 1 1            
    5 Omitting work Feeder starts with the processing completion signal 4 1 1            
      ↑ These measures should be listed before rating a, b, or c.



    Objection Theories

    Robin E. McDermott p.36:
    "The detection ranking looks at how likely we are to detect a failure or the effect of a failure. We start this step by identifying current controls that may detect a failure or effect of a failure.(1) If there are no current controls, the likelihood of detection will be low, and the item would receive a high ranking, such as a 9 or 10.(2) First, the current controls should be listed for all of the failure modes or the effect of the failures, and then the detection rankings assigned."
    Commentary: The above-mentioned description cannot be admitted.
    The worksheet by Robin E. McDermott has a description space for generating and detection separately. The following problem is caused by this.

    (Quated from Robin E. McDermott.p.44)

    1. Identifying Current Controls
      It is sometimes dufficult to start a step by identifying current controls that may detect a failure or effect of a failure.

      One measure does not always function for one of the three elements exclusively. Since controls which may prevent outbreak may work for detection at the same time, you cannot always decide which description space you should use.

      A description space is not provided for controls of severity, hoever, a certain measure which is expected to improve both of severity and frequency has no space to be described. For example, an inspection to detect cracks in a gas hose performed once a year may detect cracks, and may prevent gas leakage and may ease severity of effect.

      From the above-mentioned consideration, there should be prepared one mention column for "Current Controls" , and it is right to list all measures taken for the failure mode here in this column.

    2. The blank Description Space
      If there are no current controls, do not evaluate the likelihood of detection. Is it approved that a description column of "Severity" is missing or the blank description column of "Occurrencr" or "Dtection" occurs? If the designer judges that means is not necessary at all, he should list it in a blank column with an "Unnecessary".



    Chapter 6. Seven Steps / Case Study

    Step 6. Calculating Risk Index (RI)


    6-6-1 Calculating Risk Index (RI)
    (1) Significance of RI

    Estimation of 3 Elements ( a, b, and c)
    The score of each element is estimated by the expertise.

    1. Severity--------Measures to relax the severity of the effects of the failure,
    2. Occurrence---Measures to reduce the frequency of occurrence of the failure,
    3. Detection------Measures for detecting each failure mode or its factors or initial effect, while restoration is easy.
    Common Criteria for Estimation of Elements
    If measures are taken for elements, and if evaluation are given as shown on the left side, each score is given as shown on the right side.

    Common Criteria for Element Scores
    Estimation of Current Controls by Expertise Element Score
    Very short
    4
    Insufficient
    3
    Sufficient ( the optimal control level)
    2
    Perfect
    1

    The Exceptional Criteria for Detection
    When occurrence b is 2 or 1, detection c is equal to b. This is according to the exceptional criteria as shown in the table on the right (refer to 6-5-1).

    The Global Evaluation
    When rating of any element exceeds 2, measures are insufficient or very insufficient. Conversely, when score does not exceed 2, measures are sufficient or excessive.

    Necessity of a New Numerical Value Index
    When each element takes respective value, however, the global assessment needs a special calculation from the viewpoint of management. For example, when a combination shown next is coming out, how should the global evaluation be done?

    a=3, b=3, c=1, RI = 2.1
    The Exceptional Criteria (again)
    Occurrence Detection
    4
    4
    3
    2
    1
    3
    4
    3
    2
    1
    2
    2
    1
    1

    (2) The Calculation of Risk Index (RI)
    New index and technique are necessary in consideration of a viewpoint of the control technology. Furthermore, we announce a new technique to calculate a value of risk index (RI) as follows here.
    Risk Index
    Each element's score a, b and c is obtained from the expertise of the product or the process, and is transformed in to the new synthetic evaluation index RI, which is a numerical value indicating the degree of lack (i.e., additional need) of measures from the view point of control technology.

    The optimal RI
    Since the optimal score of each element is 2, the optimal RI is also 2. When each element obtains the score which does not exceed 2, current control passes reliability assessment promptly. For example, the failure mode that obtains the score shown below passes evaluation smoothly.
    a=2, b=2, c=2

    Necessity for Special Index and Formula
    However, it is difficult for the case where the combination of a score is
    a= 3, b= 3, c= 1
    to make an overall judgment based on expertise of each field.

    The following formula will be obtained when this problem is dealt with using control technology.

    You can substitute the element scores into this expression and get a synthetic evaluation score, based on the next reasons.

    The Reason for the Dimensional Return
    The elements a, b, and c are one-dimensional values which show the necessity for measures. In order similarly for RI to serve as a one-dimensional value which shows the necessity for measures, the three-dimensional value needs to be returned to a one-dimensional value.

    The Reason for the one-dimensional value
    The calculation of RI should make not an arithmetic average but a cubic average. It is because each score a, b, and c has a mutually different meaning.

    (3) Case Study of the Product FMEA
    All failure modes have come to pass the evaluation as shown in the worksheet below. However, a study of optimization is necessary because
    IR = 1.6 < 2.0
    implies surplus measures.

    Line 1:
    " a= 2"
    is the optimal point and the evaluation is a "Pass."

    Line 2 to 5:
    " RI=1.6"
    is not the optimal value and some optimization is needed (refer to the step 7).

    Calculating RI of the Product FMEA
    ( a:Severity, b:Occurrence, c:Detection, RI: Risk Index )
    Product:Gas kitchen range Product FMEA Review
    Line Article or
    Interface
    Failure Mode (Breakage) Effect Mechanism Current Control (Grounds) a b c RI Action Added a b c RI
    1 Introduction hose Cracks Gradual gas leakage Aging Annual check

    Stainless steel cover

    2 2 2 2.0          
    2 Retainer screw Slack Introduction hose coming-off Sudden gas leakage Collision with cooking devices

    Spilling salty broth

    Child's mischief

    4 1 1 1.6          
    3 Hose band Ruswt          
    4 Cracks          
    5 Hose connector Rust,
    Collapse of shape
             


    (4) Case Study of the Process FMEA

    All Lines except Line 3:
    The failure modes have come to pass the evaluation as shown below. However, because of IR=1.6, some optimization should be considered (refer to the step 7).

    Line 3:
    IR=2.5 means that measures are not sufficient, and additional measures should be contrived (refer to the step 7).

    Calculating RI of the Process FMEA
    ( a:Severity, b:Occurrence, c:Detection, RI: Risk Index )
    Process: Medicine Bottle Ass'y Process FMEA Review
    Line Step No.
    Name
    Failure Mode (Violation) Effect Mechanism Current Control (Grounds) a b c RI Action Added a b c RI
    1 Step 3
    Cap twisting
    Breach of 60 rpm Slack of screw cap Human error Fixation of the gearshift lever 4 1 1 1.6          
    2 Change of 5 kg-cm          
    3 Wrong oil Color marking 4 2 2 2.5          
    4 Using up oil Automatic oil dropper 4 1 1 1.6          
    5 Omitting work Feeder starts with the processing completion signal 4 1 1 1.6          





    Chapter 6. Seven Steps / Case Study

    Step 7. Optimization


    6-7-1 Optimization for Product FMEA

    (1) Significance of optimization

    Adjust the Current Controls
    Optimization is to adjust so that the Current Controls may approach RI=2. It is required to adjust Current Controls in consideration of all the elements about a result including quality, a cost price, delivery (time and quantity), security, and environmental protection, so that the best may be gained. At this stage, the advantage of the Absolute Evaluation Method appears with a very clear form. You can judge the surplus and lack of measures about every failure by an Absolute Evaluation Method. And you should recognize that this is totally different from the traditional Relative Evaluation.

    Know the Situation of Current Controls by RI
    By value of RI, the designer sees the situation of the Current Controls in real time. Then necessary measures are taken early, and Concurrent Engineering is enabled.

    (2) What Does RI Indicate
    By numerical values of RI, Current Controls are judged as follows.

    Numerical Value of RI Judgment
    4.0 3.6 3.3 3.2
    3.0 2.9 2.6 2.5
    Rejection More effective measures are needed. The closer to 4 RI gets, the more effective measures you should add to Current Controsl.
    2.3 Reservation Stand by until the result of a test comes out.
    If the test result is a failure, a small improvement has to be continued until it passes. However, when a suitable means is not found, it may be treated as temporary success for the time being. Or it may be regarded as a success, according to the contents of the test result.
    2.0 Pass. Current controls are at the optimal lebel.
    1.6 1.4 1.3 1.0Pass If RI is within this scope, the current controls also pass the evaluation. However, measures might be excessive, and it is necessary to consider actions as below are needed in order to regain the cost to pay.
    • A device to decrease cost, or/and
    • A device to increase sales charm.
    ( The borders are not strict. For example, 2.5 may be treated like 2.3 depending on contents of the potential influence. )

    When RI = 2.3
    Actions are decided by examining the numerical values of the elements "a, b, and c" that compose RI (refer to the table below). If RI is approximately 2.3, the judgment is postponed for a while. In many cases, the result of the examination about this failure mode is waited for. And in other cases, actions are postponed until a better idea appears.

    A number of actions are considerable as shown in the table below.

    The Possible Combination of a, b, and c for RI=2.3
    Line Combination Product FMEA Process FMEA
    a b c
    1 1 3 4
  • Exclude from consideration, because Severity score is 1.
  • 2 1 4 3
    3 2 3 2
  • Wait for the result of the reliability inspections (trial product tests, simulations, strength calculations, etc. may be included), because "c= 2 or c= 1" means that actions according to the result of reliability inspections are appropriate.
    1. If the reliability test is rejected, you have to continue to add the changes to the way until it succeeds.
      If solution suitable in the case of a= 3 is not found, you may determine as success for the time being. And action for solution is postponed until a good idea is found. However, in the case of a=4, you should take an effective measure even if costly.

    2. Additional measures are not required when the Current Controls passes the examination. Moreover, both b= 4 and b= 3 should be rewritten b= 2.
  • You should postpone detrmination until an effective idea of low cost appears. You have to continue to try the changes to the way until it succeeds.

    If solution suitable in the case of a= 3 is not found, you may determine as success for the time being. And action for solution is postponed until a good idea is found.

    However, in the case of a=4, you should take an effective measure even if costly.

  • 4 3 2 2
    5 3 4 1
    6 4 3 1
    7 4 1 3
  • Exclude from consideration, because b=1 leads to c=1, and these combinations never happen being based on the exceptional criteria,
  • 8 3 1 4
  • If Current Controls for Severity are very short (that is, a=4), and this may result in a serious result, you should apply FMECA, which is mentioned at the explanation (4). → Click here.

  • (3) The Case Study of the Product FMEA
    The above-mentioned ways of thinking applies to the case study of the following product FMEA.

    Optimizing Reliability of the Product FMEA
    ( a:Severity, b:Occurrence, c:Detection, RI: Risk Index )
    Product:Gas kitchen range Product FMEA Review
    Line Article or
    Interface
    Failure Mode (Breakage) Effect Mechanism Current Control (Grounds) a b c RI Action Added a b c RI
    1 Introduction hose Cracks Gradual gas leakage Aging Annual check

    Stainless steel cover

    2 2 2 2.0 Pass        
    2 Retainer screw Slack Introduction hose coming-off Sudden gas leakage Collision with cooking devices

    Spilling salty broth

    Child's mischief

    4 1 1 1.6 Pass        
    3 Hose band Ruswt        
    4 Cracks        
    5 Hose connector Rust,
    Collapse of shape
           

    Line 1:
    RI = 2.0 indicates that the Current Controls are probably optimal and no more actions are necessary.

    Lines 2 to 5:
    The protective effect of the stainless steel cover seems a little excessive, because of RI=1.6. However, you cannot find substitute means to protect from the things shown below;

    The whole including the introduction hose which was not planned at first was covered with a stainless steel cover.

    On the other hand, the device was designed to have sale charm such as beautiful appearance, simple cleaning, convenient use and safe improvement with the additional cost.

    (4) Failure Mode, Effect and Criticality Analysis (FMECA)
    FMECA mentions FMEA of the kind to attach great importance to Severity in particular. In the absolute rating method, the score a=4 means that the measures are very short, however does not always show seriousness of the effect. For example, as for the crack not only of the paper cup, but also of the passenger plane, Severity will get a=4 if measures are very shourt.

    About the crack of the paper cup, you can push forward a normal FMEA to go in principle, but should apply FMECA about the crack of the plane, the result of which is particularly serious. FMECA is applied when the result of the failure is as important as a death accident. When the result is estimated to be extremely serious, the need of measures is judged in severity alone regardless of RI. If point of severity is 4 and the result turns out critical, it is necessary to apply the FMECA. In many cases, the special measures taken besides usual measures are as follows.

    (4) The Case Study of the Process FMEA
    Similarly, the above-mentioned ways of thinking applies to the case study of the process FMEA.

    As for RI = 1.6:
    Three measures to be shown below were adopted by the reason of low cost and the effectiveness.
    Line 1 to 2: Fixation of the gearshift lever with screw or a knock pin.
    Line 4: An automatic oil dropper unit is to be attached to the twisting machine.
    Line 5:The automatic start of the feeder by the processing completion signal of the cap screw bundle machine.

    Calculating RI of the Process FMEA
    ( a:Severity, b:Occurrence, c:Detection, RI: Risk Index )
    Process: Medicine Bottle Ass'y Process FMEA Review
    Line Step No.
    Name
    Failure Mode (Violation) Effect Mechanism Current Control (Grounds) a b c RI Action Added a b c RI
    1 No.3

    Cap twisting

    Breach of 60 rpm Slack of screw cap Human error Fixation of the gearshift lever 4 1 1 1.6 Pass        
    2 Breach of
    5 kg-cm
    3 Wrong oil Color making 4 2 2 2.5 Monthly check of oil 4 3 3 3.3
    4 4 4 4
          2.2
    4 Using up oil Automatic dripping,
    Automatic alarm
    4 1 1 1.6 Pass        
    5 Omitting work Feeder starts with the completion signal 4 1 1 1.6 Pass        

    As for RI = 2.5:
    The color marking system is one of the methods of identifying grades of lubricant and preventing the mistakes in grades by color painting of the containers or instruments. However, this was estimated to be a little insufficient, and the addition of the following measures was determined.

    The Method of Adding RI
    It was decided that in addition to the color marking system (the current measure A), a new measure (B) of a monthly check of oil grade by the chief of the manufacturing section was performed.

    Calculation of the new RI
    In this case, how should the RI calculation be treated?
    The numerical value of each element of Severity, Occurrence, Detection, Production of these and RI for the countermeasures A and B are shown respectively in the table below. So far, it is clear by the already described explanation.

    Here, a virtual countermeasure delta Δ is introduced. The numerical values comes to be shown in the following table. It is because the virtual countermeasure is a countermeasure which is not helpful at all in fact.

    Naturally, the production is 64 and RIΔ=4.0.
    Substitute those numerical values to the formula, and get the improved result RIc= 1.9.


    (5) Design Review and Concurrent Engineering

    Completion of the FMEA
    You perform functional design first about the suitable small range which constitutes a product or a process. Next, you begin the reliability design of the portion. Then, you begin FMEA, you add a countermeasure if necessary, and you complete the design of that range, and advance the similar steps till the whole range is covered.

    Eventually, as a whole, a design and an FMEA are performed simultaneously. Moreover, it can be said that it is carried out in order of a functional design, reliability design and its FMEA.

    Design Review
    The design review (DR) team should assess the design plan including the FMEA executed by the designer.

    Concurrent Engineering
    It becomes very easy to carry out concurrent engineering as a result of the design management in which designers perform design and FMEA simultaneously. Since it takes time too much, it is very difficult to carry out concurrent engineering at the factory which carries out traditional FMEA.

    Objection Theories

    Robin E. McDermott p.39:
    "The resulting RPNs can be organized on a Pareto diagram and compared with the original RPNs. In addition, the total RPNs of the before-and-after product or process can be compared and contrasted. You should expect at least a 50 percent or grater reduction in the total RPNs after an FMEA."
    Commentary:
    The explanation above is totally meaningless about the matters shown below. Since RPNs do not indicate necessity of measures but mere priority, reduction in the total RPNs does not indicate that it is an appropriate improvement. Countermeasures may be insufficient or maybe excessive. Hence, the optimization is quite impossible in the above-mentioned method.

    Furthermore, it is a mistake to compare only RPNs without an overall judgment considering of all elements including cost.

    Robin E. McDermott p.39 :

    "There is no target RPN for FMEAs. It is up to the FMEA team and the company to decide on how far the team should go with improvements."
    Commentary:
    The upper explanation confesses that the relative evaluation method cannot distinguish the insufficiency of measures from the surplus. And this explanation is an evidence that the relative way of FMEA should be abolished immediately.

    (The End)