Read Spring 07 text version

www.asq.org/auto SPRING 2007

2006-2007 DIVISION COUNCIL ROSTER

Chairperson (Voting Officer) CHERYL FRANKS DENMAN Tel: 313/919-3294 E-mail: [email protected] Past Chair (Voting Officer) LOU ANN LATHROP GMPT Tel: 248/685-4875 E-mail: [email protected] Chair Elect (2007-2008) (Voting Officer) JACKIE PARKHURST Work: 313/758-4570 E-mail: [email protected] Treasurer (Voting Officer) FRANK BYKAYLO RDA Group Work: 248/836-6048 E-mail: [email protected] Secretary (Voting Officer) DENISE TISO GMPT Work 248/431-9852 E-mail: [email protected] Automotive Excellence Editor/Exhibits Volunteer AMY LICHONCZAK Tel: 586/435-6290 E-mail: [email protected] Historian/MQC Liaison/AIAG Liaison/QFD Liaison (Voting Member) LLOYD D. BRUMFIELD Work: 248/364-0196 x170 E-mail: [email protected] Health Care Liaison DAN REID GMPT Work: 248/857-1166 E-mail: [email protected] Standards (Voting Member) DOUGLAS BERG Work: 248/348-2765 E-mail: [email protected] Examining/Chair (Voting Member) JERRY BOWEN Home: 810/694-1586 E-mail: [email protected] Regional Councilor, WCQI/Annual Dinner Chair FRANCIS W. CURTISS Home: 763/425-3724 E-mail: [email protected] Membership Chair/Professional Development Detroit Section CLEM GOEBEL (Voting Member) Goebel Consulting Group Work: 810/599-6188 E-mail: [email protected] McDermond Award Chair TERESA PRATT Delphia Automotive E-mail: [email protected] ASQ, BOD President RON ATKINSON TEL: 248/821-4806 E-mail: [email protected] Awards Chair (Voting Member) CHARLES L. TOMLINSON Home: 248/652-1844 E-mail: [email protected] Exhibits Chair DENNIS C. SEEGER Detroit Edison Work: 313/235-9025 E-mail: [email protected] Jarvis Award Chair open position Quality Leader Award Chair CAROLE MALONE E-mail: [email protected] Craig Award Chair LARRY R. SMITH Tel: 313/623-7724 E-mail: [email protected] Koth Award Chair ALLY HAMOOD General Motors Corporation Work: 586/575-2838 E-mail: [email protected] AICE Conference Chairs Co Chair & Speaker Chair CAROL ANN BLANCHARD John Deere Waterloo Works Tel: 319/292-8170 E-mail: [email protected]eere.com Co Chair & Website Registration CAROLYN HAASE Pearson Gevernment Solutions Tel: 319/665-7720 E-mail: [email protected] Quality Professional Award Chair/Assistant Awards Chair/ Recognition Awards Banquet JAYNIE L. VIZE Tel: 248/371-2413 E-mail: [email protected] Ann Arbor Liaison ERIC ZINK Eaton Corporation Work: 734/741-5399 E-mail: [email protected] Detroit Liaison/Coordinator Team India ABHIJIT SENGUPTA Work: 313/595-5310 E-mail: [email protected] Northwestern Illinois Section 1212 FRANCES BLOSSER E-mail: [email protected] Scholarships HIRA FOTEDAR Tel: 440/933-3626 E-mail: [email protected] Assistant Scholarships/Saginaw Liaison KEN ZIMMER Tel: 989/868-4811 E-mail: [email protected] Certification Chair open position Vice Chair Programs DAS NARAYAN GM Warranty Group Work: 586/492-4671 E-mail: [email protected] Chair-Paper Symposium ERIC HAYLER BMW Mnaufacturing Co., LLC Work: 864/989-5577 Coordinator Team India Project BHARATH VIJAYENDRA RDA Group Work: 248/836-6008 E-mail: [email protected] Team Thailand Leader MARIA TOLETOVA E-mail: [email protected] Job Opening Coordinator MARTY DOBSON Enmark Tool & Gage Co. Work: 586/293-2797 E-mail: [email protected] ASQ Headquarters SHIRL FURGER American Society for Quality Work: 800/248-1946 E-mail: [email protected] Online/Internet open position

In this issue: Solving Modern Problems with Tried and True Quality Analysis Techniques

4 7 FMEA Thinking Applied to Program/Project Management A Standards World Viewed from Dr. Deming's Teachings

11 Monte Carlo Simulation in Quality Engeineering 14 Y to X Problem Solving using Shainin Strategies

ASQ Automotive Division Members & Friends Annual Dinner Cruise

We will board the Rivership Romance to cruise the historic St. Johns River and be immersed in an adventure of dining and dancing while we enjoy Florida's natural splendor. If you're looking for an attraction that allows you to experience Florida the natural way, look no further. This is the true Florida!

Monday, April 30, 2006 5:30PM to 10:30PM

Advanced Registration Required: $35 til April 15, 2007, $40 after April 15, 2007 Ticket price includes bus transportation from World Congress, cold appetizers, dinner, and entertainment. Cash bar

ºINSERT ALL INFORMATION REQUESTED INTO THE FORM BELOW FOR EACH PERSON THEN: Mail with a Check in the Appropriate Amount, Payable to "AUTOMOTIVE DIVISION ASQ" to:

Francis W. Curtiss, 6209 Creekview Lane N. Brooklyn Park, Minnesota 55443

Use your credit card by sending an e-mail containing all of the information requested below to [email protected]

Name: Phone: Company: Address: City Type of Credit Card Expiration Date: Name of Credit Card Holder State: Credit Card # Zip:

ASQ Automotive Division

VISION To be the worldwide automotive industry's leader on issues related to quality MISSION To facilitate continuous improvement and customer satisfaction by identifying, communicating and promoting: · Quality knowledge · Management's leadership role · Industry Cooperation · Professional development · Recognition · Opportunities to network CUSTOMERS PRIMARY · Automotive Division members · Automotive suppliers ­ all tiers · ASQ sections · Division sustaining members · Potential Automotive Division members SECONDARY · Automotive original equipment manufacturers (OEMs) · Other ASQ Divisions · Strategic alliances -- SAE, AIAG, SME, ESD, ASI, organized labor · Community colleges/universities · ASQ headquarters/Board of Directors/General Technical Council TERTIARY· Quality award initiatives (federal/state/local) · Standards activities ·Automotive dealerships · International global markets · Aftermarkets/ service parts · Third party registrars · Recruiters/ consultants

2 Automotive Excellence

Spring 2007

CHAIRPERSON AND EDITOR'S COMMENTS

Spring is in the air and with it the landscapes come alive again in may hues of green. The new growth we observe as the temperatures increase can be quite invigorating and motivating. Just seeing that snow disappear is an instant high for me. If you are taking some deep breaths of fresh spring air right now and are inspired to seek out new opportunities for yourself as well, you have come to the right place for professional growth. Immediately as you read between the covers of this season's Automotive Excellence, you will have some material to ponder, review and even possibly share with an associate or friend. Remember ASQ Headquarters goal of "Each Reach One" as successful recruiting pays off in free ASQ bucks to spend on reading material at the ASQ store. Forwarding Automotive Excellence along with a short note about the $29 forum membership may be just what a colleague needs this Spring to join or renew ASQ membership. Before you send this issue along though, check out the Roster of the Automotive Division Council members and reach out to any of us ot learn more about upcoming opportunities. If you would like to join us on a committee working remotely or as an event volunteer on-site, please let us know so we can make the connection. Talk about warmer temperatures, ASQ members near and far are meeting in Orlando soon and the Automotive Division has planned some great events to network, learn and even have some fun. My goal is to learn more about your volunteer and member experiences during these events, so please seek me out to let me know about your current and upcoming needs and expectations. If you are unable to attend the World Quality Conference in Orlando, please email me at [email protected] so that we can meet on-line or by phone and dialogue about your experiences so that Automotive Division can continue to deliver outstanding member service and development. Cheryl Denman Chair 2006-2007

Welcome to the Spring 2007 edition of Automotive Excellence. Most of the articles that you will be reading in this issue are from the 5th Annual Quality Symposium held on February 19, 2007 in Detroit, Michigan. The Summer 2007 edition will also contain articles from the Annual Quality Symposium as well as information regarding our annual Awards Ceremony. Look for the Summer 2007 edition to be delivered in May. There are many events scheduled in the up and coming months around the United States. Check out the calendar in the newsletter and on-line to see what is happening in your area. From Quality Expositions to Conferences to Event Speakers, there is a lot of activity in the automotive industry, and the Automotive Division plans many events to keep you current and looking ahead to the future. Automotive Excellence is a newsletter designed to communicate information to the Automotive Division membership. Without your articles, we would not have a newsletter. Articles from 1000 to over 3000 words are accepted, with or without graphics. If you or your team has a practice to share, we are always looking for content to spotlight the diversity of our membership. If you have any questions on topics, content or ideas, I would be glad to answer any questions you may have. Please feel free to contact me at [email protected] Amy Lichonczak Publications Chair 2006-2007

EVENTS

APRIL

April 16 Full Council Meeting April 16 SAE 2007 World Congress Sponsored by Toyota Motor Corp Detroit, MI Apr 28-30 DAC, WQC April 30 Annual Dinner Cruise aboard Rivership Romance 5:30-10:30 pm For information: www.asq.org/auto/event s/ pdf/2007-dinnercruise.pdf

MAY

April 30 - May 2 WQC-"Fueling Innovation", Orlando, FL May 8-9 Global Automotive Aftermarket Symposium Chicago, IL May 21 Council Meeting JUNE June 14 Council Meeting Annual Awards Banquet @ GM

Published by BCG, LLC and ASQ Automotive Division Editor-in-Chief: Amy Lichonczak Publication Committee: J e rr y B o w e n , L l oyd D . B ru m f i e l d, L arr y S m i t h, D e nn i s G e ntr y Publisher: J e ff r e y C ook Direct all editorial submissions to: Amy Lichonczak @: [email protected] For advertising information, eMail Jeffrey Cook @: [email protected]

When reordering request document number B0623

NOTE: Please forward all change of address requests to ASQ headquarters at www.asq.org

Automotive Excellence (ISSN) is published by BCG, LLC., for ASQ Automotive Division. All Editorial submissions should be directed to Cheryl Denman, Editor-in-Chief, or Jeffrey Cook, Publisher at 44468 Sussex, Novi, Michigan 48375. Advertising inquiries should be directed to Jeffrey Cook, BCG, Inc. Copyright © 2005, ASQ Automotive Division. No information may be copied without the express written permission of the Editor-in-Chief, ASQ Automotive Division. Neither the ASQ, ASQ Automotive Division, nor the Publisher's office is responsible for the accuracy of information in editorial articles or advertising in this publication. Readers should independently evaluate the accuracy of any statements in the editorial or advertising of this publication which are important to them and rely on their own independent evalualuation.

Spring 2007

Automotive Excellence

3

QUALITY ANALYSIS

FMEA Thinking Applied to Program/Project Management

Leslie Carey, Rand E. Winters Group, Deborah Gordon, Honda of America

ABSTRACT Project Management has the necessary ingredients to use PFMEA thinking to improve attainment of goals and plans; past problem history and experiences with projects that failed to meet their goals, Gantt chart in place of a flow chart, many potential customers with multiple expectations (specifications), and opportunities to be proactive and to detect potential problems. Honda of America had 11 plants with a need to further "commonize" their interactions with suppliers. Today the Supplier Quality Manual common to all sites continues to be improved on a planned, periodic basis by the users. The project itself was completed on time and budget four years ago. BACKGROUND Murphy's Law states that if it can go wrong, it eventually will go wrong. This law appears to live in the world of program and project management. Failure can occur on many levels without being captured in the financial records under following categories: scrap, rework, or regrade . Scrapped projects that those that are started and abandoned, Reworked projects are those where phases of the project are repeated several times. Regrade projects are those that are diminished in their scope while in process or not maintained once the gain is realized. Opportunity for this type of failure appears to be common to all projects/program changes. The information required to conduct an FMEA is available. FMEA thinking requires knowledge of Critical to Quality (CTQ) factors which can be identified from the Voice of the Customer (customer needs that have been converted into requirements). Evidence of past problems exist and analysis of their cause(s) can be conducted. People with experience with management of these events or those that experienced the result of projects can assist in identifying potential problems that have yet to surface. With all of this information we are ready to conduct a FMEA with a cross-functional team. FMEA methodology and thinking applied to program and project management will improve the success rate and the ability to maintain the gain. PFMEA was used as a tactic for the two issues identified in the above paragraphs. One tactic resulted from root cause analysis of why projects/programs can fail regardless of the business focus. The root causes tended to group themselves: ineffective communication plans/methods, human relationship issues, and ineffective management of improvement/problem solving activities. These issues can be captured on a single PFMEA that is applied to all projects/pro4 Automotive Excellence

checklist of questions. The following list is not all inclusive but serves as an example of the type of questions that should be included. VOC · Who is the customer? · Does it stay the same over time? · How do we identify our customer's needs? · How do we ensure consistent feedback? · How do we evaluate them-ease and practicality of use? · What are their needs today and over time? · What are their expectations? · How do we know? VOB · What are the requirements to meet customer needs and expectations? · What are the critical to quality (CTQ) characteristics? · What has been our history with these requirements: successes, failures, resources available? · How do the customer needs/expectations fit into our current and future business plans? · Our delivery ability? · Our opportunity costs? · Risk of NOT doing this? · Any required changes in roles & responsibilities? · Our change management ability? · Any new skills we need to learn? PROJECT/PROGRAM MANAGEMENT · Type of structure of this event (goal, process based)? · Review and reporting structure? · Alignment to business goals? · What are the related, functional requirements to meet the CTQs? · What experience do we have with these historically? · Our evaluation plan · What objectives show our level of progress along the way? · Roles and responsibilities? · Interpersonal conflicts? · Time lines, milestones agreed upon? · Risk involved in this project/program? PREPARATION FOR FMEA METHODOLOGY · Gather CTQ (VOC) · Gather functional requirements needed to meet the CTQs · Gather evidence of past problems · Gather information from similar events · Gather successes of past · Gather complaining statements and survey responses about potential barriers and pitfalls and conduct root cause analysis · Identify cross-functional team

grams, which are then updated and revised in the normal method. Most of the information for this PFMEA was discovered by analysis of failure and listening to people "complain" about other people, recognizing those complaints as predictors of behavior. Anything we can predict we can build into our planning to influence it for the good. Comments such as: "Well, you know he will just get an attitude!" and "We have always done it that way and it has worked!" were the basis of root cause analysis. The other issue involves the CTQ's and may require a new FMEA for each project. How Project Management and FMEA fit into the "Big" Picture of Improvement. Improvement methodologies involve moving from divergent information to convergence through the use of "tools" that improve our ability to be effective and efficient. The methodology repeats, diverge with collection of information; converge on appropriate details, over and over...

TOOLS Many tools exist for Program/Project Management but this discussion is limited to a simple

HONDA OF AMERICA Approximately 6 years ago Honda of America formed a new function they entitled the

Spring 2007

QUALITY TECHNIQUES

North America Technical Group (NAT). It includes quality, supplier development, supplier training, colors, tooling with the purpose of "commonizing" activities and best practices. Soon after the formation of this group, Deb Gordon of the Quality section, and Lestie Carey were asked if they could facilitate the development of an on-line, single Supplier Quality Manual (SQM) for all North American Honda sites. At that time, the sites in Ohio shared one manual and the other sites had their own; none were on the Honda portal. STRUCTURE OF PROGRAM CHANGE The first issue we faced was a definition of this activity. It was agreed that this change constituted a program change rather than a project. Projects imply short lived and compete for resources with other life threatening appearing situations so the resources required can be a moving target. Program changes can be more stable in part because they align to business level changes and the resources are more likely to be available throughout the process. Collection of information would involve surveys, interviews, observations, focus groups, and case studies. We would be looking for 3 types of evaluation information: · The reactions of the customers · Changed attitudes and perceptions · Changed behaviors It was determined to handle this program as a process with defined inputs, process steps, and outputs in the Plan, Do Check, Act cycle. 3 Circles to Define Customers, Roles and Responsibilities and Communication Plans Control Circle are the process owners. Communication purpose: APPLICATION and IMPLEMENTATION (must know why, what, where, when, how-open ended questions/ draft level for input) · NAT Quality as mirror for all North American sites (not corporate) · Quality site Managers and associates for input and as customers if SQM maintained by NAT Influence Circle are those who influence or are influenced by potential changes. Communication purpose: UNDERSTANDING for input at beginning milestones, and on-going feedback (must know sufficient why's and what's to ask questions and give feedbackopen ended questions/draft level for input) · Suppliers

Spring 2007

· · · · ·

Honda purchasing / procurement Honda delivery / logistics NAT Supplier Development NAT Training Honda training

· Computer issues · Communication plan and issues · Human conflict issues 4. Analysis of costs and benefits PFMEA BASICS Assume Deign or Goal Statement is correct Not need everyone to agree but to use diversity to arrive at effective consensus Phase 1 · Select appropriate team members ­ use the 3 circle information · Identify primary & secondary input data ­ again from the 3 circle · Develop or obtain process flow chart and validate ­ for this assignment we use the Gantt chart list of actions Phase 2 · Identify process functions and list · Identify potential failure modes · Identify adverse effects for each failure mode · Determine the severity rating for each effect · Classify the critical characteristics based on these effects · Identify potential causes for the failure modes · Determine the occurrence rating for each potential cause · Identify current process controls · Determine the detection rating of each control as related to each cause · Calculate the risk priority number (RPN) for each failure mode scenario (priority of the failure)

Concern are those further removed from the consequences but have need to know and to give feedback. Communication purpose: AWARENESS and available for feedback, review (must know why and big picture fit ­ close to final phase/details available) · Top Management

· Design Function · Production departments Program Management Partial List of Activities (Gantt chart material) 1. Define and agree on goal and scope · One SQM on line w/in 2 years with appropriate levels of access available to Honda and Supplier associates · Continual improvement a controlled process to avoid need to ever develop another, new SQM 2. Continual Improvement If this business change is worthwhile; then begin now to identify how, who, resources necessary to maintain and improve once it is in place. 3. Development of the SQM · Identify those in the 3 circles and bring on board at appropriate levels · Current situation · Concept of initial SQM of North America

Phase 3 · Determine recommended actions to reduce all the higher RPNs · Determine responsibility and target completion date · Document actions taken and recalculate the RPN Phase 4 · During the product life, revisit the process FMEA as appropriate Honda's PFMEA for Development of single, North American Supplier Quality Manual online within 1.5 years of start date. Material brought to FMEA: CTQ and requirements, Gantt chart of activities, and information from surveys, interviews, and focus groups Suggestion: Develop one PFMEA for all projects/program development. The activities in a Gantt chart for project management are similar s once a PFMEA is developed, it could be reused, updated, and revised. The problems involved in change and that contribute to project failure are mostly predictable. Anything we can predict in this arena we can influence and on a FMEA that is called recommended actions.

Automotive Excellence 5

QUALITY ANALYSIS

The CTQ/requirements would be different for every project and might require a new PFMEA. RPN = 216 Recommended Actions: · From on-set emphasize WHY the change is occurring, what it is and is not, and risk of not making any changes · Explore how it can/will affect their operations and include their feedback in the plans for change · Get information to site associates as early as possible for their consideration and no surprises at meetings · When adding or changing procedures, get feedback on length of trial periods desired by site management · Review procedures via email and not meetings so personality and size of operations not have strong impact on results RPN = 216 One example from Honda program change: · Meeting purpose carefully identified and consistent feedback from Honda associates at published ahead of time, people notified quarterly meetings and from suppliers through about WHY they are needed at the meeting email, focus groups, and training sessions. (3 circles), and agenda is labeled "Probable Training to support this method of doing busiAgenda" to allow room for "learning" durness continues to improve and change to ing the meeting. address improvements and gaps that are iden· Meet with focus groups to get their input tified. prior to first meeting about the strengths and weaknesses of the goal, what advanThe SQM supports the functions of other tages and disadvantages it might bring to their jobs, etc. · First meeting start by One example of PFMEA Basic to ALL Projects and Program Changes: getting input about what can go wrong, Potential Failure Potential Effects of Sev. Project what it can do for Modes Failure Function them, their perRequirement ceived risks. · Have outsider facilCapture resistance Not create the Failure to achieve itate first series of to learn real truths space and safety goals meetings required to capture and gain buy-in · For everyone in the and use open Results are less 3 circles, identify Use communication resistance than optium thoroughly WHY plane effectively becausenot learn this change needs Not ask right intimate details questions at good about failures from to be made, the risk

times (example: when asking for feedback, ask when document or plan is 80-90% done) Communication not following the 3 circles plan other projects and changes Timeline not meet because facilitators will be "dragging" some key people through the process, information withheld,delivered late Increase diversity of practices rather than reduce the variations Potential Cause or Mechanism of Failure Fear of change being done to them · Not believe in the value of the project, passive resistance · Territorial issues to protect operations from change O 6 Current Process Controls (D/P) Feedback in meetings Feedback from document review process D 4 7 Potential Cause or Mechanism of Failure Not agree to follow common procedures due to: Fear of change being done to them Not believe in the value of the project, passive resistance Territorial issues to protect operations from change not being well thought out O 6 Current Process Controls (D/P) Feedback in meetings Feedback from document review process D 4 7

Recommended Actions: · Create time and space for resistance to be up front. Start w/email survey anonymous. Compile and return to all participants. No meeting until this is done and people have time to read. Questions include potential barriers, past failures, what success would look like. Do two iterations prior to first meeting to capture their own divergence and move toward convergence. · Agenda of first meeting set in part by survey results. · Prior to communication, determine awareness, understanding, or application level and deliver in appropriate manner.

6 Automotive Excellence

of not making the change (provide evidence), controls in place to maintain the integrity of current practices during the change. Conclusion The results of this approach to program/ business change are an on-line, single Supplier Quality Manual for all Honda manufacturing sites in North America. It was available online 10 months after the start of the assignment. The SQM is continually improved through

departments at Honda and includes these needs in their quarterly meeting reviews. Honda NAT has a method of change and improvement to apply to other projects.

Deborah Gordon, ASQ Automotive Division Member, email: [email protected] Leslie Carey, Quality Consultant, email: [email protected]

Spring 2007

QUALITY TECHNIQUES

A Standards World Viewed from Dr. Deming's Teachings

Douglas L. Berg and William M. Harral Dr. Deming had a dismal opinion of quality standards and requirements and auditing, based on his experiences from World War II through his passing in 1993. However, the world of auditing and standards has matured to recognize many of Dr. Deming's teachings as the current foundation for the ISO 9000:2000 series. In the last years of his life Dr. W. Edwards Deming was asked on many occasions what he thought of the ISO 9000 series of quality systems standards and the registration process that was just getting started in the United States. He made rather acerbic and disparaging comments about both at the time in both public forums and in more private settings. Echoes of those comments have lingered among some who knew and worked with him. The authors of this paper were active participants of a Deming Study Group in Detroit, formed around those identified with and interested in the teachings of Dr. Deming. These were people from industry, service organizations, education, government and other public sector entities. They would meet with Dr. Deming on a regular basis when he would come to the area and during the intervening time would meet to pursue particular topics and assignments. The topic of standards came up repeatedly during these meetings and there were several presentations and discussions. Dr. Deming could at times be quite terse and abrupt in his responses to questions. He did not appreciate being baited and often felt some questions were really attempts at some sort of self-serving validation on the part of the person posing them without putting them into the context of his writings and current thinking. He could often appear to be somewhat enigmatic or Delphic in response, but that would be to those who had not read, let alone studied his writings, digested them sufficiently and ruminated over their contents. Often, people tended to extract words or phrases from his writings or comments and use them to substantiate or bolster their own pet perceptions or positions which in a broader sense would be quite at variance with Deming's overall perspectives. DR. DEMING AND STANDARDS IN GENERAL Dr. Deming wrote and often spoke about the benefits of standards particularly product standards. He would hold up a device,

Spring 2007

typically a small hand held tape recorder, and tell how regardless of where in the world he was, from Singapore to Switzerland, he could get cells ("batteries"), in this case AAA cells, and they would fit and work fine in his device. This, he would point out, was due to standardization. An "AAA" cell, he would say, might not be the ideal one from the perspective of an engineer at one company developing a particular product, nor that of another elsewhere. But in each case, the engineer would accommodate his or her design to utilize the more ubiquitous and less expensive common cell. Those who would have a more specialized, idiosyncratic battery or cell, could do so, but that consumers, being able to freely chose, would more likely select those products that use the more easily obtained replacements. Standardization in this case, opened the market to more competition and facilitated free and open trade. In a sense, it lowered the overall cost to the consumer and enabled more companies to enter a market and introduce competitive offerings. Standards, particularly product, process and testing standards, enable purchasers to pursue with greater confidence commercial relationships with suppliers who adhere or incorporate these standards in their products. They form a basis for establishing the product offering and allow for the comparison among suppliers on an equal, "apples to apples" basis. They are enablers to trade and commerce, benefiting the purchasers, suppliers and, eventually, the consumers. Dr. Deming would often point to standards for products, testing or processes as examples of selfless cooperation among competitors and others. Knowledgeable minds, supported by their companies or other organizations, would work together in a consensus process to develop these standards which, when followed, would form a basis of confidence, performance, and conformance. This would benefit consumers and allow companies to compete in the marketplace in terms of features and price. This type of cooperation, what some have termed "cooperatition" (sic), is a form of competition that is good for the consumer, good domestically, and good for international trade. In Deming's view, it works to the benefit of everyone, including the companies and organizations involved. However, ISO 9000 seemed to evoke a different response. William W. Scherkenbach recalled that Dr. Deming said you can not install a quality system like it was a refrigerator. When asked Dr. Deming would reply brusquely and gruffly to the effect that ISO 9000 and the registration or certification process were silly, nonsense, foolishness, that he had never heard of such things before and other disparaging terms. Strong words? Harsh words? A harsh assessment? To be sure, and more so in private than in

public. But why? Pressed further regarding the standards and especially the auditing or assessment process he became clearer as to what the key issues and problem areas were as he saw them. ORIGINS Partly this may have been due to the origins of the standards themselves, their "progenitors", so to speak. The development of the original version of the ISO 9000 series issued in 1987 was greatly influenced by BS 5750, a British quality standard. This, in turn, had a lineage that traced back to the US military MIL-Q9858. Both had broad usage respectively and there was quite a lot of accumulated experience with them and their implementation. Other influential requirements documents were employed elsewhere in the nuclear and defense industries. It was perhaps the anomalies and problematic aspects seen in the companies that had to comply with them that prompted some of Dr. Deming's concerns and issues. Associated with this was the role of the auditors or assessors who determined conformity or compliance and the purposes and motivations behind their noting of nonconformities or deficiencies. Dr. Deming was undoubtedly aware of some of the problems observable in how particular companies and organizations would approach these standards as imposed upon them as requirements by their customers. These standards reflected an approach to quality that had a strong emphasis not on process and process control but inspection and inspection technologies and the development of quality assurance measures surrounding them. This along with the structure of the documents themselves had the effect of compartmentalizing quality into specific functions or job descriptions and isolating it from others. There was little emphasis on the importance of management, management systems, the interaction and interdependencies among processes and operational units within a company. In short, lack of both an appreciation of a "system" and a systems focus plagued many implementations. With this went a misplaced focus on "downstream" activities in production, away from the customer and interaction with the customer on the part of engineering and sales. Also, it tended to be remote from the interaction with key suppliers in a functioning partnership that Deming believed was essential to the success of the organization and its products or services in the marketplace. Like an intoxicated man looking for the keys he lost elsewhere by crawling around underneath a streetlight, the concern of quality appeared to be in the convenient, what could be grasped easily by the quality functions or elements in the organizations. To Dr. Deming, the important things, the hard things were not dealt with directly.

Automotive Excellence 7

QUALITY ANALYSIS

The ISO 9000 series from its inception through its 1994 version was even issued with three requirements documents. ISO 9001 was the most comprehensive. ISO 9002 was like 9001 but did not include engineering and was intended for those organizations that build to a provided design. ISO 9003 was intended for those products whose quality was best ascertained through inspection processes, rather than quality assurance and process control. However, at the time of their introduction organizations that were design responsible would seek to become certified and registered as being in conformity with ISO 9002 since they had doubts about the readiness and willingness of their engineering organizations to cooperate with the effort. Others sought certification and registration to ISO 9003 not because it was appropriate, but that in their eagerness to be "registered" they thought it would be "quicker", "easier", and "cheaper". The auditing or assessment process for these standards also raised concerns. While with the best of intentions the compliance or conformance auditing approach ­ referencing the standards or requirements and noting deficiencies or nonconformances ­ could motivate organizations to address them and, ideally, become better in the process, it appeared to Dr. Deming that this was a fairly inefficient and ineffective means of doing so. It was also difficult to see how addressing an accumulation of negatives could result in some significant net positive, let alone a transforming result. This was especially the case when one considers what was not strongly addressed in the requirements documents themselves and the emphasis that Dr. Deming placed on more fundamental, transformative factors. Dr. Deming saw this carrying over into the certification and registration process, then coming into the fore. Much of this same mindset on compliance auditing or conformity assessment was there since the earlier model permeated the thinking at the time and shaped the experience and knowledge of the auditors or assessors themselves. Of course it had to be that way, was there any other way? Dr. Deming thought much of this was a waste of time, and worse, dangerous. Jumping from nonconformity to nonconformity, applying band-aids or even well intended corrective actions, runs a grave risk of missing the mark and sub-optimizing even as root causes are identified and addressed. Without a systems focus, without clear management direction and goals for the organization, the result was often a patchwork of "fixes" to remedy various nonconformities. The focus on deficiencies and nonconformities factored greatly in the certification and registration process and the

8 Automotive Excellence

training provided to the assessors. The noted author Alan Sayle has decried this as not being a positive, constructive process, that it is focused not on the future but on the past. Looking backward, and not forward, somewhat like driving a car, while looking in the rear view mirror. At this point we may have to ask. With over 15 years and more of certification and registration activities across a wide range of industries and products, including services, have they helped or hurt? Are the organizations and their suppliers better off? More competent? More competitive? Healthier? As that noted current "advice guru" Dr. Phil McGraw, "Dr. Phil" of TV and literary fame, asks, "How is that working for you"? EVOLUTION OF ISO 9000 SERIES Dr. Deming was open to discussion and consideration of the application of the ISO 9000 series standards and the certification and registration process. It was a topic of presentations, reasoned discussion, debate and reconsideration. Part of what was presented was the overall directions and plans for the evolutionary development of the series as envisioned at the time by the technical management of ISO and the strategic direction of the Technical Committee (TC) 176 charged with the series' development. At the time, these were being put in place through the drafting process of what became the 1994 versions. Most of this was to address immediate issues arising from the application of the 1987 versions, clarifying the concept of product to include services, expanding the scope in terms of organizations, and resolving some inconsistencies and incongruities. Also, the TC emphasized that the appropriate standards or requirements document needed to be meaningfully applied. Dr. Deming died before these documents were finalized and issued. They remained "descriptive" standards. They expressed the "what" but not the "how", "when", "how much" or "how often". What they spawned, however, were some industry specific requirements documents in the automotive and aerospace industries, in particular, which included more "prescriptive" requirements either in them explicitly or in associated documents. What Dr. Deming did not see and could not anticipate was how the series would evolve over time to its 2000 version and the impact, both explicit and implicit, his thoughts and perspectives would have on them. In their current version, there is only one requirements document, ISO 9001; one vocabulary document, ISO 9000; and one implementation guideline, ISO 9004. By strategic direction and technical management, and recognizing the impact changes would have worldwide, no significant changes are envisioned in these docu-

ments until 2008. DR. DEMING'S INFLUENCE IN THE ISO 9000 SERIES On the surface the current versions of the ISO 9000 series are notable documents in terms of the incorporation of the ideas and concepts of Dr. Deming. The most obvious is the inclusion of what Dr. Deming called the "Shewhart Cycle" but what others have termed the "Deming Cycle". It is the familiar "PDCA" or "Plan", "Do", "Check", "Act" cycle, which Dr. Deming later referred to as "PDSA" or "Plan", "Do", "Study", "Act". Permeating them are the concept of a management "system", the emphasis not on the organizational structure per se, but on processes, the interactions within the organization, and continual improvement ­ all key concepts in Dr. Deming's work. Top management in the organization receives more explicit concentration on its roles and responsibilities to ensure a focus on customer satisfaction, establish a quality policy set measurable but meaningful quality objectives, communicate these within the organization, conduct planned management reviews and ensure adequate resources, including a trained competent workforce. Would Dr. Deming have approved? We believe that he would have been gracious enough to acknowledge the work done. He also would have been pleased to the extent that people took what he said and wrote seriously enough to have incorporated them as extensively as they have in these documents. No other contemporary quality "guru" has similar impact. Like all such standards, he would have respected their efforts and appreciated their personal and professional commitment. ISO 9000 SERIES APPLICATION ­ ISSUES? But we think he would have looked at the application of these standards and raised some questions. If the approach to MIL-Q9858 and other requirements documents had problems due to "low leverage", how would an audit or assessment process, no matter how thorough, in and of itself, get an improvement? Through the management representative and his or her efforts? Through the management review? Through imposing requirements on top management? Have we seen it? Have we really seen it? He would have pointed to the gamesmanship on the part of organizations and their management in the certification and registration process. Examples: A) Minimal efforts to meet the letter of the requirements. B) "Vigorous" competition and aggressive low bidding among the registrars to acquire and maintain business. C) Registrars becoming mere surrogates for major

Spring 2007

QUALITY TECHNIQUES

industrial organizations that are externalizing costs, while maintaining extensive control under some schemes. Dr. Deming would have cited the introduction of a third party into the customer/supplier relationship and the potential of this to distort this relationship which he felt was a key long term commitment. He would have looked at the state of American industry as a whole, and certain industries in particular, and asked if the attention paid to requirements, assessments and certification and registration schemes has had a net positive effect? If it has, has it masked the need to address broader, bigger, more substantive issues? TOP MANAGEMENT, CONTINUAL IMPROVEMENT Is top management taking the responsibilities and commitment seriously? Are they simply "involved"? Are the goals and objectives being set for the organizations meaningful and challenging? Or are they "low-balling", merely scrounging around coming up with "measurables" that can be met without challenge? Are they aware when objectives are set that they are truly value adding for the organization as a whole and not some misbegotten "management by objectives" that, as Deming pointed out, can lead to internal competition and sub optimization? Is there true continual improvement? Meaningful, effective improvement? Or is it continual improvement in some measure as evidenced by some obtainable "objective" data while the organization as a whole is not truly improving? Is the "customer focus" actually leading to customer delight and innovation or is it simply getting customer feedback in a passive sense? Or is the focus on meeting customer wants, needs and expectations? This by no means implies the latter are not important, but in the long term, Deming would point out, they are not sufficient. It is not a substitute for leadership in terms of the market, product or process. Is top management motivated by improvement? Or are they driven by "compliance" or simply "fear of non-compliance"? Dr. Deming would acknowledge that compliance to law, regulation, or requirements would or should be a given, but management needs to go beyond mere compliance. CORRECTIVE ACTIONS? There is a significant emphasis on corrective action in the standards and their application. Much of this is in response to a "problem" noted in the product or service, processes or systems and is essentially "reactive". Find a problem, fix it. Dr. Deming would say, according to William W.

Spring 2007

Scherkenbach, that all the examples in his books came from companies whose management thought they had no problems. But ask customers and experienced auditors and assessors and they will tell you that many corrective actions are problematic. Analysis tends to be superficial. Organizations confuse "remedial actions" or "corrections" for true "corrective actions". Part of this is poor analysis. Part of this is the misunderstanding and misplaced emphasis from a Deming perspective on finding a "root cause". Even if that is successful the solutions are often not effectively implemented or, when implemented, prove not to be truly effective. Many times fixes do not seem to stay fixed, and the ongoing lack of effectiveness is such that the problem reoccurs. Without proper analysis at the beginning, the organization can have a good fix for the wrong, or some other problem. Scherkenbach also observes that for every complex business problem, popular methodologies will come up with a simple solution...and it is wrong. Dr. Deming would note the need to start with a clear definition of what the problem is in operational terms. This would be followed by its classification as evidence of either a system that is fundamentally "out of control" or an occurrence of lack of control of what is fundamentally a stable process that is usually in a state of control (which would indicate a special or assignable cause). The former would frustrate any well intentioned problem solving approach regardless of the number or color of the belts anyone might have. The second would lend itself to the type of analyses popularized and promoted by various proponents and quality professionals. It could also be that the problem might just be from a stable process, in control, but producing an unacceptable result. This would require yet another set of approaches and tools. To truly determine a "root cause" requires knowledge of the system and the constituent processes, elements, and their interactions. Many of the popular problem solving approaches tends to dismiss the interaction effects, considering them of secondary importance. The reason for this in many cases is expediency, in that it tends to mess up the analyses and requires real, fundamental knowledge of the particular situation and its idiosyncrasies. It is messy work. However, if one looks at the more dramatic failures of manned space flight, building and structural collapses, major utility outages, and most recalls, few, if any, have a single traceable cause. It is usually a sequence of failures or conditions that set up the final cause enabling the failure. Our experience day to day tells us that. Even simple examples such as the fire triangle (Does the spark alone cause the fire?) or accidents demonstrate the need to consider conditions and interactions. Dr. Deming would remind those seeking

corrective action of the need to understand the process and how it interacts with other processes and the various parts of the organization, the customer or the environment. This complicates the search for the ephemeral "root cause" since the interactions, unless studied or controlled could make a root cause "conditional". Therefore knowledge of a "system" is essential in all but the simplest or most direct instances. PREVENTIVE ACTIONS Beyond the reactive corrective actions, are preventive actions which address potential problems, that is, those that have not yet occurred. These would implicitly assume that the fundamental processes are in a state of control and there are no indications of the need for corrective actions as indicated above. Otherwise, some action would have been taken in corrective response. These preventive actions might involve improvements in technology or processes, knowledge of problems encountered elsewhere in competitors or those with similar products or processes, or knowledge upcoming changes in regulations or requirements, or knowledge of the organizations particular products or processes. In short, preventive action takes knowledge. Again, talking to customers or experienced auditors or assessors and they will tell you that there is overall scant evidence of true preventive action. Part of this is the misunderstanding of the distinction between corrective and preventive action. The major reason we believe is that it requires critical examination of the status quo, when there are no or few apparent problems. Also, they require some creative thought as to what has the potential for causing a problem that has not yet occurred and putting in place measures to address them, and often not without some cost. The true cost of the preventive action, as studies and experience have shown, is usually much less than the costs due to the resultant failure that would occur eventually, but the current expenditure lacks the urgency driven by a pressing problem or failure and is easy to put off. Management looks at this as a current expense only, not as an expense which offsets greater expense and difficulties in the future. And if management has difficulty with preventive action, which is in its essence still focused on a "problem", they really have difficulty with "continual improvement" in products and processes and the disturbing of what appears to be a smooth running stable system. As a result, one tends to see corrective actions that are basically remedial actions, fewer truly effective and effectively implemented corrective actions, fewer still preventive actions, and even fewer instances of continual improvement.

Automotive Excellence 9

QUALITY ANALYSIS

MANAGEMENT REVIEWS? The reason for this state of affairs can be traced to a lack of knowledge and understanding of the role of top management, even where partially defined by these requirements standards. Experience shows that most "compliance driven" organizations seek a minimal level of effort and demonstration that is the absolutely least they can do to achieve compliance. This can be seen in infrequent, ineffective management reviews. Management has pushed addressing the requirements down into the organization or has placed them on the shoulders of the management representative. Often, this person is the only member of management who may meet or interact significantly with the auditors or assessors throughout the process. The management might make an appearance at the brief entrance or exit meetings, but that is it. It is viewed as the management representative's "job". Recently, one of us was party to a discussion at a certification/registration body regarding what minimal amount of information or documentation would be required to "meet" the requirements of ISO 9001 for management review. This same discussion has taken place in just about every certification/registration and accreditation body. It is not much in a legalistic sense. Yet experienced auditors or assessors can recount that such discussions are more than just frequent in dealing with organizations as clients. Training courses abound that teach what is the least that an organization can "get away with" and yet still be considered complaint. Or argue that it is compliant, when it is questioned. One would be hard put to see how such a minimalist approach would ensure a program of effective preventive action or true continual improvement, let alone the type of management commitment, not just involvement, Dr. Deming would view as the necessary next steps to ensure long term survival of the organization. PROFOUND KNOWLEDGE The technical management of ISO has directed that the various quality systems, environmental, health and safety, and other requirements documents be consistent with, or "harmonized" with the ISO 9000 series at least with regards to basic concepts and management practices. This is a laudable goal in that is intended to foster an integrated management approach and avoid duplicative effort. However, it still perpetuates the compartmentalization of quality, environmental, and health and safety and the tendency of top management to view them as being distinct and also to delegate responsibility downwards in the organization. This reflects a lack of systems thinking or the appreciation of the interactions within the organization.

10 Automotive Excellence

One direction that ISO could consider is to develop a single document, descriptive requirement or guidance, which would consolidate the management oriented concepts into an integrated, consistent whole. However, it could be argued that this alone would not be sufficient to accomplish the type of transformation Dr. Deming foresaw as being essential for organizations to survive let alone thrive. It may be that this would have to be introduced in a guidance document such as ISO 9004 in the next revisions. It will take time, to Dr. Deming, time was short and of the essence. Towards the end of his life, he advanced the idea that management must not only provide leadership but that a fundamental transformation must take place. This would be the case not only for organizations in the private sector, but also in the public sector, such as government and education. This transformation would require what he termed "Profound Knowledge". While this term has been misused and abused by many who use it flippantly or inappropriately, Deming had a specific meaning for this term. He understood it to be truly transformational on the organizational level, but required transformation as well at the individual level. It could not be approached piecemeal, it was an integrated whole and required an understanding of each component and how it impacts or interacts with the others. What did he mean by "Profound Knowledge"? From his The New Economics for Industry, Government, Education: · · · · Appreciation for a system Knowledge about variation Theory of knowledge Psychology

Would revision of the ISO 9000 series or the related management systems documents be necessary to accomplish this? Not hardly. But one can see where the movement of top management in this direction, to make the "transformation" would remedy many, if not all, of the issues raised earlier. Without it, improvement will be slow, if not stall out completely. An important step would be the acquaintance or re-acquaintance of the auditors or assessors with the key points of Dr. Deming's teachings. They need to have an appreciation for the true role of management in the organization. They need to internalize that all work in done by processes and understand the constituent elements and their interactions. They need to understand the two types of variation when they look at information, records, etc. and properly put an observation or the information they are obtaining in the proper perspective. They should properly write up deficiencies, nonconformities or corrective action requests, however termed, so as to point out the exact issues involved, and evaluate the resulting actions taken by the organization in terms of appropriateness and effectiveness. To be sure, these are small steps in light of the larger, more substantive transformational issues that confront modern management, but there are positive steps individuals can take to assist the process.

ACKNOWLEGEMENTS

The first two have been touched on in the earlier parts of this paper and presentation. The need for a theory of knowledge might not be clear, however. Dr. Deming stressed that leadership, the true role of management, was built on prediction. Prediction, on the other hand, required a "theory" and becomes knowledge through the systematic process of observation, revision, and extension based on comparing predicted results to what actually happens. Without the recognition of an underlying theory, there is no revision, no learning, and no improvement. All of this takes conscious thought and consideration. Much of what passes for management input is mere information, but information, in and of itself, is not knowledge without an underlying theory of the functioning of the system producing it. It requires understanding as to whether the system or processes are fundamentally stable, and thus capable of prediction and improvement, or unstable, needing a completely different set of immediate actions. Confusion regarding nature of the system or processes results in inappropriate behavior and ineffective effort.

This paper had its germination as had many others in the discussions the authors had over several years in bits and pieces as part of normal conversations. It was proposed by us together. Unfortunately, William Harral died unexpectedly in January 2006. This paper is dedicated to his memory as a quality professional, consultant, keen observer and friend. I would like to thank Dr. Roderick Munro and Mr. William W. Scherkenbach for their review and coments, contributing greatly to this article.

Douglas L. Berg, ASQ Fellow, email: [email protected]

LOOKING FOR A WAY TO STAY INFORMED WITH THE RAPIDLY CHANGING QUALITY INDUSTRY?

The Automotive Division has many opportunities to be come involved with the automotive quality industry. For more information contact: Jackie Parkhurst at: [email protected] or visit our web site at

GET INVOLVED

www.asq.org/auto

Spring 2007

QUALITY TECHNIQUES

Monte Carlo Simulation in Quality Engineering

Clement Goebel, President, Goebel Consulting Group Dr. W. Edwards Deming often said that a theory (a model) is very useful to help us learn about our world, even though it is never completely accurate. The Monte Carlo technique combines a model with probability to determine a `likely' answer. The method is named for the Monaco principality's (second smallest country in the world) world famous Monte Carlo casino because the element of chance is used in both. material. Credit for inventing the Monte Carlo method often goes to Stanislaw Ulam, a Polish born mathematician who worked for John von Neuman on the United States' Manhattan Project during World War II. It was only after electronic computers were first built (from 1945 on) that Monte Carlo methods began to be studied in depth. Ulam's contribution was to recognize the trol Assembly (see Figure 1) that his plant produced. The AC did not always come on when the levers were moved from the `Off' position into `AC' (->) or from the `Vent' position into `AC' (<-). His plant was performing 100% test and repairing 30-40% of the assemblies. The Assembly plant was replacing far too many assemblies when they checked the AC system performance. His

Figure 1. AC/Heater Control Assembly

In the second half of the nineteenth century, a number of people performed experiments, in which they threw a needle in a haphazard manner onto a board ruled with parallel straight lines and inferred the value of Pi to be 3.1416.... from observations of the number of intersections between needle and lines. An account of this playful diversion occurs in a paper by Hall (A. Hall 1873) "On an experimental determination of Pi". In 1908, W. S. Gosset (Student by `nom de plume') used experimental sampling to help him discover the distribution of the correlation coefficient (r). In this same year, Student also used sampling to bolster his faith in his so-called t-distribution, which he had derived by a somewhat shaky and incomplete theoretical analysis. The real use of Monte Carlo methods as a research tool stems from work performed on the atomic bomb during the Second World War. This work involved a direct simulation of the probabilistic problems concerned with random neutron diffusion in fissile

Spring 2007

potential of the electronic computer to automate such sampling. The Rand Corporation and the U. S. Air Force were two of the major organizations responsible for funding and disseminating information on Monte Carlo methods. The question was whether or not the Monte Carlo method could estimate the solution to a problem in an intractable class to within a specified statistical accuracy in time, bounded above by a polynomial in M. Dyer (1989) established it for estimating the volume of a convex body in M-dimensional Euclidean space. Chevrolet Motor Division included Monte Carlo simulation in its' Quality Engineering tools taught to its' Central Office Quality Control engineers and "high potential" manufacturing personnel in the late 1960's. Class was held in the mornings and real plant problems solved in the afternoons (this could be considered a pre-cursor to the Six Sigma certification training and project experience requirements of today). One of the students brought all the drawings for an Air-conditioning /Heater Con-

plant was also receiving many Air-conditioning Warranty returned parts (even after the 100% Assembly Plant AC system test and replacement). Many efforts were made to problem solve this non-linear problem at his plant, however none had been successful. A complex mathematical model (71 variables and 148 calculations) was made of the lever system that rotated the vacuum valve controlling the operation of the Airconditioning. The Monte Carlo results are shown in Figure 2, they clearly show that with all components to print specification, over 23% of the assemblies will not put the vacuum valve in the 37º -43º position when the lever is moved from the `Off' position into `AC' (->). Changing the vacuum valve casting molds resulted in correcting the problem and saving the company many dollars in testing, repair and warranty costs. As a simple example how to do a Monte Carlo analysis, let us consider the "opening height" (see Figure 3) between two blocks. The opening height is determined by the depth of the hole in the upper block, the depth of the hole in the lower block and

Automotive Excellence 11

QUALITY ANALYSIS

98.4 seconds to generate 500 samples of the 71 variables and the 148 calculations. Now you can see how to do a Monte Carlo analysis using a model and probability distribution. Let us compare the Monte Carlo method with the traditional Engineering Tolerance Stack-up method. For the MAXIMUM case we do the analysis as described in Figure 6. When there are only a few variables and a simple linear calculation, there is not that much difference between the Tolerance Stack-up method and the Monte Carlo method. However in the Air conditioning/Heater Control Assembly study there were seventy-one (71) input variables and 148 complex calculations including a nonlinear relationship. In that case the results look quite different:

Figure 2. Monte Carlo results for AC/Heater Control Assembly

the length of the rod in between the blocks. Our `mathematical model' is then: opening height = rod length ­ lower hole depth ­ upper hole depth To perform a Monte Carlo analysis, a random value is assigned to each of the variables in the right hand side of the equation. Once this is done, we can calculate a random opening height. This sequence can be repeated 50 times, 100 times, 500 times or

The Tolerance Stack-up Maximum valve angle when the lever was moved from the `Off' position to the `AC' position (->) was 53.09 degrees. The Minimum valve angle Thus, if we rolled two dice and got a 5, we when the lever was moved from the `Off' would assign the value 3.619 to the rod position to the `AC' position (->) was 26.19 length. If we rolled a second time and got degrees. This results in a total expected an 8, we would assign the value .791 to the variation of 26.89 degrees. The lower hole depth. Likewise, if we rolled a Monte Carlo Maximum valve third time and got a 3, we would assign the angle when the lever was value .743 to the upper hole depth. moved from the `Off' Then we could calculate the opening position to the `AC' posiheight to be 2.085 inches. opention (->) was ing height = rod length ­ 45.75 degrees. lower hole depth ­ upper The Monte hole depth Carlo Minimum valve angle when the Figure 4. Distribution of two dice Of course we could lever was moved from replicate this any number of times we the `Off' position to the `AC' position (->) desired. To perform this exercise 500 times was38.73 degrees. This results in a total would take quite a bit of time and effort, expected variation of 7.02 degrees. See the even for this simple model. The computer large discrepancy between the method automates this sampling and can perform results. Again this is due to the large num500 samples in seconds. In fact, the comber of variables in the Airplex AC/Heater Control model took only

Figure 3 Opening Height Graphic

even 1,000 times, as we choose. For clarity, we will use two dice to show how we could do it right here and now. First we have to know the "distribution" of the dice results as shown below in Figure 4. We can only get the value `2' by having a `1' on each die. Likewise, we can get a value `3' by getting a `1' on the first die and a `2' on the second die. Or we could get a value `2' on the first die and a value `1' on the second die. If the rod length specification is nominally 3.630 inches with a tolerance of +/- 0.30 inches, and the upper and lower hole depths are treated accordingly, we could develop Figure 5.

12 Automotive Excellence

Figure 5 Assigning random values to calculation variables

Spring 2007

QUALITY TECHNIQUES

Max. opening height = Max. rod length ­ Min. lower hole depth ­ Min. upper hole depthMax. opening height = 3.66 - .755 -.735 = 2.17 For the MINIMUM case we do the following analysis: Min. opening height = Min. rod length ­ Max. lower hole depth ­ Max. upper hole depth Min. opening height = 3.60 - .815 -.795 = 1.99

Figure 6. Example Maximum Analysis

conditioning/Heater mathematical model as well as the very large number of calculations in the Air-conditioning/Heater mathematical model. A short time after I was assigned to the Quality Engineering training activity, I was re-assigned to a new pick-up truck product development project. The intent was to have Manufacturing Quality Engineering work directly with Product Design Engineers through out the product development cycle to eliminate current product problems and prevent others from occurring at the production stage of the project. This was The `net hole' design avoided the need to keep tooling in use and in good repair at all times. The `net hole' design worked well in production. A simple Monte Carlo computer program would be structured as follow: 1. 2. 3. 4. 5. 6. Data Input Model Input Program generates `random' input values Program uses model to calculate results Program calculates statistics Program stores iteration results 7. Program repeats steps 3, 4, 5, and 6 until `iterations' are complete 8. Program provides outputs

4. "The Monte Carlo method" by Nicholas Metropolis and Stanislaw Ulam 1949 in the Journal of the American Statistical Association. 5. "History of Monte Carlo method" http://stud4.tuwien.ac.at/~e9527412history.h tml 6. "Monte Carlo Method" http://www.riskglossary.com/articles/monte _carlo_method.htm 7. "What is Monte Carlo Simulation" http://www.decisioneering.com/monte-carlosimulation.html 8. "Monte Carlo method" http://www.answers.com/topic/monte-carlomethod 9. "Introduction to Monte Carlo Simulation: http://physics.gac.edu/~huber/envision/instruct /montecar.htm 10. "Real Options with Monte Carlo Simulation" http://www.puc-rio/marco.ind/montecarlo-.html 11. "Monte Carlo Simulation in Excel: A Practical Guide" http://www.vertex42.com/ExcelArticles/mc/in dex.html 12. "Monte Carlo Simulation Basics" http://www.vertex42.com/ExcelArticles/mc/M onteCarloSimulation.html 13. "Monte Carlo Simulation" http://www.circle4.com/pww/mc/

Clement Goebel, ASQ Automotive Division Member, email: [email protected]

HELP WANTED?

Contact ASQ Automotive Division "Job Opening" Coordinator" For complete information and details contact: Marty Dobson directly at: [email protected]

Automotive Excellence 13

Figure 7 Ladder frame as an exaggerated parallelogram

not a common practice in the late 1960's at Chevrolet. One such current model problem that existed was that some pick-up trucks tended to `dog track' when driving down the road. This was due to the fact that the ladder frame was not rectangle, but more of a parallelogram (see Figure 7). A complex computer model was developed based on existing design specifications. It was validated by loosely assembling current parts on a surface plate with rivets in the proper holes and then `racking' the frame rails to maximum parallelism. Once validated, the model was used to evaluate several design alternatives including assembly fixturing and the incorporation of `net holes' in the design. This was done without the high cost of building physical prototypes to evaluate these various alternatives. The solution chosen was the `net hole' concept since there were six pick-up truck assembly plants with multiple shifts at each plant.

Spring 2007

The above discussion should give you an insight into how Monte Carlo was developed and how it might be useful to you. The examples were included to provide insight to those who learn best by example. The automotive industry is shifting to electronics and by 2010, the design is predicted to be 40% software. However, this still leaves 60% of the vehicle for mechanical designs like this article describes. Monte Carlo Simulation can be applied in many quality engineering problem solving situations.

Bibliography

1. "Monte Carlo Simulation" by Greg Kochanski March 10, 2005 2. "The Basics of Monte Carlo Simulations" by Joy Waller at the Physical Chemistry Lab University of Nebraska-Lincoln Spring 1996. 3. "Principles of Monte Carlo Simulation" by Clayton Deutsch, Oy Leuangthong, Hanh Nguyen, Karl Norrena, Julian Ortiz Bora Oz, Michael Pyrcz and Stefan Zanon

QUALITY ANALYSIS

Y to X Problem Solving using Shainin Strategies

Ha Dao, Delphi Corporation, Bill Maxson, Shainin, LLC ABSTRACT In any given process, a large number of causes (X's) are active at any point in time. When problems arise, measuring and comparing parts and processes to prints are the strategies used to attack said problems. For simple problems this approach may work, but for more complex problems a more rigorous, disciplined approach is needed. This paper discusses the alternative to X to Y problem solving with the Y to X approach, followed by a case study illustrating how using a progressive search which quickly and efficiently eliminated potential causes for leaky brakes on a drum brake application. BACKGROUND If you have ever had the pleasure of taking your vehicle to the repair shop several times for the same problem, you are most likely experiencing X to Y problem solving. The mechanic employed a trial and error approach to the problem based on experience, hunches, and guessing, creating NTF's (No Trouble Found), frustrated customers, and unnecessary warranty costs. As engineers, we use a more sophisticated approach with brainstorming, statistical analysis, redesigns, and sometimes a committee approach to solving problems, which often discover insignificant X's, miss interactions, or drive unnecessary engineering changes. When you compare the two approaches, no matter how different they may seem, they have one thing in common; they are X to Y. Identify the potential causes, or X's, and then test to see if they affect the response (Y) we are trying to control. The X to Y approach comes in many forms, which can be inefficient and rarely provides an insight into the physics or science behind the problem. The common strategy is to change one or more causes to see if they have an effect on the Y response. The X's are often derived by brainstorming, fishbone analysis, or "engineering judgment"; in other words, "guessing". The

14 Automotive Excellence

assumption here is you know the relationship between the cause and effect. Y TO X The Y to X approach starts with a welldefined Green Y®. The Green Y® is a response that has engineering leverage and reveals insight into the physics for the fail-

The fastest way to identify the Red X® is through a progressive search or the process of elimination. This technique eliminates the need to list out potential X's and then testing each to see if they have an effect on the Green Y®. During a progressive search investigation, each stage is divided into natural "buckets". By leveraging contrast with efficient strategies, these buckets are progressively eliminated and all further work is focused on what remains. By leveraging contrast between extremes, BOBs and WOWs, a multitude of X's can be quickly and efficiently eliminated leaving only a few suspects for the Red X® candidate. The candidate can then be tested with a well designed and simple confirmation test. The most effective Y to X approach is Red X® Strategies. Red X® Strategies gets its power from Pareto, Square Root of the Sum of the Squares and by talking to the parts. Combining the three aforementioned techniques created a powerful problem solving strategy. PARETO Many problems appear to be overly complex, with little hope of understanding them; leading to the conclusion that there are many significant causes or Red X's. This is simply not true. With Red X® Strategies, complex problems become understandable. It isn't necessary to control many X's to solve difficult problems; 50 years of practical experience and Pareto tell us this. The bottom line is that results speak for the experience, but Pareto may need further explanation.

ure. A good Y to X approach relies on looking for differences between the extremes of the effect to progressively close in upon the most significant cause, the Red X®. The only assumption you have to make is that you will be able to understand the physics by observing the extremes. If you have this understanding, options to control or eliminate the problem become obvious. Corrective actions can now be based on the economics associated with each.

Pareto is the 80/20 rule, which simply stated, says that for any given Green Y®, only 20% of the X's have any noticeable contribution. Vilifredo Pareto, whose name is most closely associated with this phenomenon, made this discovery after studyWorlds Wealthiest People in 2005 ing the distribution of wealth in nations under different political systems.

Spring 2007

QUALITY TECHNIQUES

He concluded that there is always a nonrandom distribution of wealth. This nonrandom distribution of variables makes itself felt in all aspects of life. Businessmen can tell you that 80% of all sales come from 20% of the customers. For problem solvers, it separates the trivial many from the vital few. When a problem can be expressed in terms of differences, the tallest bar on the Pareto is the Red X®. If we describe the problem as a change in Y, there are some cause-effect relationships that have more affect on the Y than all the others. This is the Red X®, which can be a main effect or an interaction. SQUARE ROOT OF THE SUM OF THE SQUARES The square root of the sum of the squares (SRSS) is a law of statistics and shows us that the power of the Red X® is exponentially more powerful than any other X! When two independent distributions are combined, their standard deviations do not add, but combine as squares. The standard deviation of the resulting distribution will be the square root of the sum of the squares of the source distribution standard deviations.. If the Red X is causing 5 units of variation in an output and a smaller cause is contributing 2 units, the relative effect ratio is not 5:2 but 25:4. The total amount of variation is not 7 (5+2) but the square root of 25+4, which equals 5.4. As you can see, squaring of the independent variables greatly magnifies the effects. (52+22) = 5.38 What happens when you attempt to control variation by addressing a true root cause other than the Red X®? In the example above if you remove one unit of variation from the smaller cause, the result will be a reduction of total variation from 5.4 to about 5.1 (the square root of 1 squared plus 5 squared is a little less than 5.1). (52+12) = 5.09 Had you instead removed one unit of variation from the Red X®, the total variation would have fallen to about 4.5 units. (42+22) = 4.47 Both the larger and smaller X's are true causes of variation, but the Red X® has much more power over the effect. This is often experienced when X's have been proven to be statistically significant but practically worthless; therefore adding cost when controlled with little return on investment.. TALKING TO THE PARTS The fastest route to finding the Red X is through a progressive search. This requires identifying contrast between extremes of the Green Y® distribution called BOBs and WOWs. By leveraging the BOBs and WOWs, insignificant X's are avoided, leaving only the Red X® as the only cause of such large variation in the Green Y®. This means only a few parts are needed to solve problems, sometimes only one. The statistics used to investigate the differences are simple and straight forward, letting the problem solver focus on strategy and tool selection and not statistical analysis. Once the Red X® candidate is identified, it is now ready for the trial. X TO Y Confirmation is the only time for an X to Y approach. With a well-structured progressive search many suspects will be eliminated, leaving only a few and often one Red X® candidate. A more efficient confirmation test can then be used to confirm that the Red X® candidate causes the largest change in the Green Y® response. Only after effective clue generation (screening experiments and DOEs are not effective clue generation) is X to Y appropriate. At this point we should have an understanding of the physics, allowing us to make a decision on what to control or eliminate. The confirmation test now demonstrates that you truly understand the problem and can predict the Green Y® response given a known Red X®. CONCLUSION For any output Y, there are a number of input factors X that affect variation. Given enough resources and time, we could discover all of these X's. As a practical matter, however, there is no need to do this. Pareto tells us there is only one Red X® and the SRSS says that the power of the Red X® is far greater than many of the other X's combined. Therefore, the only way to significantly improve the process is to find, control, or eliminate the Red X®. Controlling other X's, even though they are true causes, will lead to insignificant improvement. A progressive search is the quickest and most efficient route to find the Red X®. This is a combination of good strategy and decision making based upon physics and science. CASE STUDY Focus A vehicle assembly plant reported leaky brakes coming from drum brakes. The plants immediate corrective action was to increase the torque on the tube nut. This fixed some leakers, but others still leaked. Tear down analysis revealed that some of the leakers had rust, but some non-leakers also had rust. It was also known that concentricity of the bore was a cause for leaky cylinders. The frequency of the leakers were filling up the repair bays and delaying shipment of vehicles and causing customer dissatisfaction. Approach

Plain view of the drum brake and wheel cylinder

The Green Y is an air leak measured as decay from 55 PSI and converted to cubic centimeters per minute (cc/min). The initial evidence suggested that the difference between BOB and WOW vehicles was the most promising place to start. This difference was leveraged with Component SearchTM. The air leak detector passed Isoplot®. A profilometer was used later in the project to control the Red X®. It also passed Isoplot®.

Spring 2007

Automotive Excellence

15

QUALITY ANLYSIS

the part drawing. Other documents or updates made include: PFMEA, Control and monitor system implemented, Bill of Process, Control Plan and audit sheets. The project took one week to complete. LEVERAGE A wheel cylinder seat waviness specification was added to all other wheel cylinder drawings. DEFINITIONS, ACRONYMS, ABBREVIATIONS

Complete reversal occurred with wheel cylinder Isoplot® verified that the Green Y® measurement system was adequate for this investigation.

Converge Because we were dealing with an assembly of many components, a Component Search was used to leverage BOB and WOW vehicles. The first split to be made was parts vs. assembly. Component Search eliminated assembly of the components. What remained was the tube, tube nut and wheel cylinder. The tube and tube nut were eliminated leaving the wheel cylinder. At this point in the investigation, the Y to X approach has eliminated assembly and many of the components of the system by leveraging Vehicle to Vehicle with Component Search`. A new strategy was needed to eliminate features or other characteristics of the wheel cylinder. The strategy shifted to leveraging BOB and WOW wheel cylinders with a Group Comparison. This led the investigation from the component level to the name of the Red X® feature, identified as waviness of the wheel cylinder bore. TEST

80/20 Principle: An inbuilt imbalance between causes and results, inputs and outputs, and effort and reward2. BOB: Best of the Best are the extreme samples that can be readily taken from a distribution of the Green Y®1. BOB Seat Waviness WOW Seat Waviness Green Y®: is a response that has engineering leverage and reveals insight By building up new cylinders with BOB into the physics for the failure1. and WOW seat waviness, we were able to Red X®: Is the product feature or process confirm with 95% confidence that seat setting that drives the Green Y® from waviness was the Red X®. BOB to WOW1 WOW: Worst of the Worst are the UNDERSTAND extreme samples that can be readily taken from a distribution of the Green Y®1. Leak rate was confirmed as a function of

Confirmation test and Run Chart confirm Red X® found and eliminated.

seat waviness. Because seat waviness was not called out on the drawing an additional specification was required. A Tolerance ParallelogramTM was used to develop the specification for seat waviness. The of the wheel cylinder manufacturer confirmed that they were capable of meeting the new specification. This eliminated the need to continue the investigation at the manufacturer to discover the Red X® process parameter. APPLY A seat waviness specification was added to

Spring 2007

REFERENCES: 1. Shainin LLC, Problem Solving for Engineering, v1.1g 2006, Page: 1-2, 1-5, A-32, A-34 2. Richard Koch (1998) The 80/20 Principle: The Secret To Success By Achieving More With Less (Currency and Doubleday), pages: 4, 264 3. http://www.shainin.com

Ha Dao, 14 year member and ASQ Fellow email: [email protected] Bill Maxson, Consultant, email: [email protected]

The joint consists of a tube nut with a flared tube. The tube nut is torqued to the wheel cylinder, compressings the flared tube. The clamped load is maintained to create a sealed joint.

16 Automotive Excellence

Information

Spring 07

16 pages

Report File (DMCA)

Our content is added by our users. We aim to remove reported files within 1 working day. Please use this link to notify us:

Report this file as copyright or inappropriate

3229