All 11 entries tagged Piuss

View all 376 entries tagged Piuss on Warwick Blogs | View entries tagged Piuss at Technorati | There are no images tagged Piuss on this blog

November 26, 2009

DOE & missions

It's a long day today. Unlike working, need to be more focused and concentrated. Most of all, need to learn soooo many  things. Yesterday a very wise man told me that he learned amazingly amount of knowledge when he did his master, school education totally opened his eyes. And now I'm feeling the same. Needless to say, there's always a lot to learn and discover in our lifetime, but sometimes you just have to work hard to get the knowledge and be sensitive enough to even recognize it's around you.

You can easily be inspired if you notice some little details or knowledge, or even just random words that people say. You never know what you're gaining. Life's full of surprise. And I love surprises!

This morning, thanks to Jarrett helping me untie the knots in my head of all the unfamiliar statistics analysis. I learned a lot in the morning. Am very proud of myself. In the afternoon, thanks to my teammates, we had a fun practice and analysis on the data, I'm really glad that I learned something totally new. This is actually one of my goal that I'm looking to achieve from the courses -- challenges. I believe challenges bring out the potential of a person to achieve something greater than one ever expects.

Now I know Taguchi (basic concepts), and how to analyse or create a product using the arrays and the statistic methods. I also enjoyed as much when running the experiments and actually doing all the practices and exercises. That helps the whole learning process. Coz you gotta use it if you're learning it - like languages~

Learning how to make a robust products also rang a bell on me as well. If I want the business to be sustainable, I need to have strong products or services. People never say no to innovative products. Anything that's new, funky, practical, interesting, convenient, delicate, fancy..etc, the market is always wide open for it. And making a robust product not only interests the customer, satisfies them, and also surprises them. As I said before, I love surprises! So the product will at least have 1 customer -- me. :-)

So today, my favorite part is practicing the real data analysis plus product manufacturing process. Learning by doing is the best, what else you ask for? :-) And being able to realize what we can discover after analysing the data after the experiment also taught me a lot that making sure I do the same in the future. It's a very objective and practical method to apply to.

Anyways, time to jot down few things to remind myself for today -- ANOVA, DoE, Mean value analysis, Interaction analysis, Signal to Noise Ratio Analysis (difficult, but very useful) -- details see paper notes.

In statistics, analysis of variance (ANOVA) is a collection of statistical models, and their associated procedures, in which the observed variance is partitioned into components due to different explanatory variables.

Design of experiments, or experimental design, (DoE) is the design of all information-gathering exercises where variation is present, whether under the full control of the experimenter or not.

*** Thanks to the team and Ivan, attached pls find our data for the helicopter exercise!


November 25, 2009

Taguchi method – loss function in statistical theory

This afternoon I learned something totally new, which totally opened my knowledge and had a lot of info processing in my head. I would say I still have many question, but I really wanna figure all this out in order to be able to actually apply the method in real business world.

While I was listening to the lecture, I jot down the key points that I think is totally fresh to me and that I want to keep reminding myself of the Taguchi concept afterwards.

And I think the idea is really important and useful in terms of trying to make a robust product that never goes out of style or the market. This is something that all the company should try to achieve in manufacturing industry. Therefore, this idea is very precious, and the method should be studied deeper and implemented widely to organisations.

Taguchi Philosophy

-         Focus on the economic pursuit of quality

-         Minimise loss to society

-         Easily controllable to robust products

How to experiment

-         Poke and hope

One factor at a time (factors and levels)

Interaction (ex. difference, similarity,)

2 factors can affect each other and create a different result.

2 levels, 7 factors = 2^7=128 experiments (can investigate almost every possible factors)

Professional statisticians have welcomed the goals and improvements brought about by Taguchi methods, particularly by Taguchi's development of designs for studying variationTaguchi's work includes three principal contributions to statistics:

  1. Taguchi loss function

Tolerance              l Product 1 and 2 are identical, but 3 is entirely different than 1 and 2

Loss function: This is the case for nominal is best

L= K(y-m) ^2

K=constant of proportionality


M= Average

Loss function:

1.     Effectively a one-sided tolerance => L=KY^2 (small is better)

2.     L=K(l/y^2) (bigger is better)

  1. 2 types of quality control: on-line & off-line

Off-line – (design) : limited ability to improve the process

DOE: design of experiment ; FMECA

On-line – (6 sigma focus): Conformance (SPC….

Taguchi's rule for manufacturing (elements of design)

Taguchi realized that the best opportunity to eliminate variation is during the design of a product and its manufacturing process. Consequently, he developed a strategy for quality engineering that can be used in both contexts. The process has three stages:

  1. System design
  2. Parameter design
  3. Tolerance design

System design

This is design at the conceptual level, involving creativity and innovation(jap focus: 40%).

-         Getting into the ‘design space’

-         Creating a feasible design

-         Involves innovation

Parameter design (Focus)

Once the concept is established, the nominal values of the various dimensions and design parameters need to be set, the detail design phase of conventional engineering. Taguchi's radical insight was that the exact choice of values required is under-specified by the performance requirements of the system. In many circumstances, this allows the parameters to be chosen so as to minimise the effects on performance arising from variation in manufacture, environment and cumulative damage. This is sometimes called robustification.

-         optimizing within the ‘design space’ (not changing anything fundamentally) (jap focus 40% ; euro 20%)

-         settings for the factors identified in systems design

Tolerance design

With a successfully completed parameter design, and an understanding of the effect that the various parameters have on performance, resources can be focused on reducing and controlling variation in the critical few dimensions

-         Tightening tolerances on important factors

-         Better materials/new equipment may be needed

-         Has significant cost implications

Factors affecting the design or process

Control: level can be set and controlled during both the experiment and the final design of the product or operation of the process

Noise: affect the product or process but cannot be ECONOMICALLY set and controlled in the final design of the product or operation of the process

Sources of Noise

-         Inner : deterioration, wearing out of components

-         Between product: manufacturing imperfections, variation in products manufactured by the same process

-         Outer: variation in the operating environment, human factors

-         à design a robust product that drive out all the noise factors!

Full Factorial Experimentation

-most rigorous approach: every level is investigated, preferred by scientific community, investigate all possible interactions and conclusions are highly producible

-disadvantages: time consuming&costly


Orthogonal Arrays: ABOUT BALANCE! (particular type of fractional factorial design)

-         All factors will average out!

-         From Average A and B, can tell the difference. (established by Fisher)


-         Give good reproducibility (be accurate)

-         Very efficient


-         Interactions must be specifically included

-         Can’t deal with third level interactions and above

-         Must apply engineering judgement to establish which interactions may be relevant.

Where does the tolerance come from?

-         Strategy

-         Standards

-         Customer requirement (customer change their mind with the tolerance level; better/worse quality)

-         Process

-         Designer

-         Cost

-         Previous experience

How to define tolerance?

-         Limits acceptable variation

-         A recognition that we can’t really have

** How bad can we get away with being? People don’t want tolerance, but on target.


Inside the tolerance, a potential of cost—may move away our position from target

**mentality of 6 sigma

Old: in tolerance is good enough

New: on tolerance with minimum variation

Airplane exercise evaluation/feedback

I very much liked our evaluation. I think we pointed out all the problems, and everybody was trying our best to do better next time. By looking at the points below, I think we can always remind ourselves that how we could do better and how well we can actually be.



-         Chaos at first (Not well planned/organised)

-         Leader too involved in process (better work as an observer)

-         Jump to conclusions/fixes

-         Poor quality definition


-         Issue identified and solutions trailed

-         Good logistic arrangement (separated our parts into 2 sides along the assemble line, and distributed the parts into 2 sides that the operators in each side know where their parts are and can collect them with ease)

-         Communication

-         Identified skill (some of were better at tasks than others – specialist tasks)

Learning point

-         Importance of following process (not jumping to conclusion)

-         Upfront communication important to keep team engaged

-         Timing and lack of planning (run out the parts)

-         Small changes in process can have large impact

-         Don’t try to fix everything at once

-         Leader should specifically address the whole process and goal

-         Don’t wait for the leader to make team decision (if problems are discovered)

November 24, 2009

Improvement exercise – Airplane

Today on my way home, I gave a lot of reflection on our exercise earlier today.

I think in our team, we worked motivatedly and happily together. I had a great time especially cheering for the operators and brought everybody's workforce together. :-) Everybody was working hard to achieve the best result we could get -- stay within the control limit, lower the variation (or eliminate it if possible). From this point of view, we did well. We stayed in the control limit, the whole process was fine. But we had at least 25% defective rate out of the 10 planes (3 times losen tail, 4 times wheel problem, 1 time misplacing a bolt).

During the process, we kept our time always ahead of customer's required time(195-225secs), yet, at the beginning of the process the QC pointed out the defect part and suggested the immediate change, the team decided to stay within the control limit - keep up the same pace without changing our speed. Was this the best decision? Did we have other better choice?

From a process improvement view, it's hard to balance if we should slower the process, make sure the quality and still maintain the production in the same time range.  On the other hand, if our goal was to keep in the same time pace, were we subconciously changed or ignored our original goal for the project assigned?--> make zero defect products. Did we start another project while we were making a change or improvement for the process? I'm not sure, but when I was on the bus way home, I think maybe we did. Maybe we were executing another project which was not 'making what our customer wants', but intstead we were trying to achieve 'what we wanted to achieve' -- stay in control.

For a plant manager, maybe this is what he/she should do. Yet, for the customer view, the process means NOTHING. The RESULT is EVERYTHIING. So were we supposed to focus on our own performance stability or on customer and product result?

I had also learnt a lot from the other team while we were sharing our results together. Even though they exceeded the time for few of the planes, the whole process was still in control, and they made zero defect products. From this point of view, we could think about that when we knew that we were ahead of our schedule almost 50 sec faster from the very beginning, we could have adjusted our process right away when we finished the first plane and we had a great chance to achieve zero defect and maintain the whole process within time limits(195-225secs). I believe that we could achieve this.

The question is, is that good or wise to adjust right away after each problem in the process, or should we evaluate after the whole process was finished? Because making changes while the process was still going may disrupted operators' pace of working or concentration if the new instructions were not handled properly. Yet, from an optimistic point of view, immediate change could also bring efficiency and effectiveness on an improved process.

I think the best part we did was we all stayed together as a team and tried to work out the best result. We also organised a logistic area and seperated all the parts for operators according to their needs in order to speed up the process.

In conclusion, from a customer's point of view, the faster is always the better as long as the quality is good. I don't mind receiving my products earlier. Earlier is better than later :-) And if we could spend 10 more secs to ensure the quality at the last step of the process, we could increase our non-defective rate of the products as well.  

Either way, I had a great time today. And analysing the data was really fun, I hope we could do more practice on that.

November 23, 2009

Six sigma successful factors

I just wrote probably the longest blog I've ever written and then the blog server crashed....

Yet, I think I'll just focus more on what I wanna emphasize on :-)

After our presentation today, I learned some new things. sometimes we might forget easily, but if we keep reminding ourselves, we can do much better time after time. From today's team exercise, I was just thinking that we could actually try to create a team bound in the whole process to speed up our production as well as reducing the defective rate. As we saw the last person had many models piling up on the table, but if he can't finish, that means all of us won't make a dime. So why not helping him?? There's nothing we can lose. On the other hand, we might have more non-defect products at the end. This is another way around to do the project, instead of individual tasks, it becomes a team based task. But I think the goal we were asked to achieve was more like a team project, coz the result is team based.

Therefore again I guess we talk about the People factor. People in every way are the most crucial role. Things are done by human. even machines need to be designed and activated by human. so human is the one who's making a difference. If an organisation can create a healthy environment for the employees or whoever that's associated with the company, then the whole business process/project could have more chance to succeed.

However, the world is always changing as if the process and improvement is always revising to become better. And when people encounter change, they're likely to feel depressed, defensive, reluctant or panicked, etc.

So I think people should develop the ability or "habbit" to continuously re-examine ourselves, to put ourselves outside the frame in order to observe better of what we haven't noticed that we are doing or what we have done. We also need to have the courage to admit and face the problems, and further improve them or solve them for good. Only if we admit our mistake, then we can find a way to fix it. (this is what we know everyday, but we always forget to do so)

Today I learned when people encounter change, there're four stages of mindset:

Denial -- Resistence -- Engagement -- Acceptance  // and it's normal that when people find things different, they feel insecure. therefore the four stages are the common process that people would have to/ might go through.


The DATA ANALYSIS part of the day was definately interesting and useful. Coz we can do nothing about our data unless we know how to read them and understand them. Therefore, this is a very important skill to build up to.

November 14, 2009

SIPOC template

As I'm learning PMI online sources these days, I think it'd be nice and useful to make the SIPOC template and Project Charter for future or current possible usage.

It can also reminds me everyday that I should review and assess myself of my work and studies. :-) Share with you all if you're working already and would love to use the template now :-)



November 12, 2009


I think DMAIC, PDSA and the whole Profound Knowledge are very well connected and related. And this is the basic and essential process that everybody or every company would go through when discovering and solving a problem.

I wanna try to look beyond the picture...that's why i was thinking, PDSA and DMAIC is the same. can be related and combined together. they're not very different. and DMAIC is more focused on the measurement and statistic analysis, but PDSA is a tool and day to day cycle. but actually they're using the same idea...just different methods.

Do I make sense? 


Jack Welch's Video:

Define: [Select Priorities] Define is the first step in the process. In this step, it is important to define specific goals in achieving outcomes that are consistent with both your customer's demands and your own business's strategy. In essence, you are laying down a road map for accomplishment. à PLAN

Measure: [Learn about process]In order to determine whether or not defects have been reduced, you need a base measurement. In this step, accurate measurements must be made and relevant data must be collected so that future comparisons can be measured to determine whether or not defects have been reduced. à DO

Analyze: [Investigate sources of variation and waste] Analysis is extremely important to determine relationships and the factors of causality. If you are trying to understand how to fix a problem, cause and effect is extremely necessary and must be considered. à Study

Improve: [Test theories+Study Results+Implement] Making improvements or optimizing your processes based on measurements and analysis can ensure that defects are lowered and processes are streamlined. à Study

Connect Analyze and Improve stages to PDSA's Study-->

ð Study is the biggest part here that we need to spend time on, because we not only have to fully understand our problem, what are we dealing with, but also analyzing all the data we have in order to dig out the real cause, and solve the problem permanently without creating new problems.

Link the whole process to Profound Knowledge -->

ð This is also the emphasis on Profound knowledge that after we have the whole picture about the system, we can then try our best to eliminate all the waste to increase efficiency and improve the whole process. Not only we need to theories to help us analyze the situation, but also need the cooperation of the people to create an environment for improvement.

Control: [Standardise + Review] This is the last step in the DMAIC methodology. Control ensures that any variances stand out and are corrected before they can influence a process negatively causing defects. Controls can be in the form of pilot runs to determine if the processes are capable and then once data is collected, a process can transition into standard production. However, continued measurement and analysis must ensue to keep processes on track and free of defects below the Six Sigma limit. à Act



Six Sigma is truly changing the way that we think and act on a daily basis, on both a strategic and individual level. In order to success, we need to consider:

1. Eliminate waste, tie projects into business strategies, monitor and report the progress, improve the process and achieve highest efficiency.

2. Organisational learning environment: Get everyone engaged and speaking the language of Six Sigma.

3. Focus on improving the infrastructure to facilitate both measurement and control of processes (in the factory and in nonmanufacturing applications).

4. Psychological influence: Apply human elementcommunicate both the why and the how of Six Sigma as early as possible, and provide the opportunity for people to improve their comfort level through preliminary training classes.

5. Amend the process or the system if needed: Restructure the organization to drive the culture change and make Six Sigma a part of everyday life. Keep improving.


Six Sigma Case study – GE

A small fraction of GE case study:
*Page 1*
Six Sigma is a measurement. A more illustrative explanation can be found in a 1997 letter Welch sent to GE stockholders: "The Six Sigma quality initiative, very briefly, means going from approximately 35,000 defects per million operations, which is average for most companies, including GE, to fewer than 4 defects per million in every element in every process that this company engages in every day."

*Page 2*
By mid 1996, Six Sigma efforts were underway in every process, product, or service. The 1997 result of this broadened focus included the completion of 1,000 projects in GE Appliances alone, with a savings of $43 million. The 1998 goal for GE Appliances is $100 million in benefits and over 3,000 completed projects.

Six Sigma training

The Six Sigma methodology is a significantly different process than had been used by operational managers in the appliance industry. Thus, most people have had to learn a new skill set for successful, data-based problem analysis. At GE Appliances, the process includes:

Measuring the process output.

Analyzing the process inputs for criticality.

Improving the process by modifying inputs.

Controlling the process by controlling the appropriate input.

Page6 - Wrap up ideas about how to implement and how to be successful

***Six Sigma philosophy. It is truly changing the way that we think and act on a daily basis, on both a strategic and individual level.

Tie defect reduction goals into cost reductions. And tie projects into business strategies. Monitor and report the progress.

Get everyone engaged and speaking the language of Six Sigma.

Focus on improving the infrastructure to facilitate both measurement and control of processes (in the factory and in nonmanufacturing applications).

Don't forget the human elementcommunicate both the why and the how of Six Sigma as early as possible, and provide the opportunity for people to improve their comfort level through preliminary training classes.

November 11, 2009

Some Books to Share

Found some books may be useful for you all. :-)

1. "The Six Sigma Way: How GE, Motorola, and Other Top Companies are Honing Their Performance"
    by Peter S. Pande

2. "What Is Six Sigma?" by Peter S. Pande

3. "The Toyota Way: 14 Management Principles from the World's Greatest Manufacturer"
    by Jeffrey Liker

4. "The Six Sigma Way Team Fieldbook: An Implementation Guide for Process Improvement Teams"
    by Peter S. Pande


Last time I used standard deviation was to prepare for my GMAT to apply for MBA!! It's really nice to review all these again. Afterall, it's a very basic and important concept to have especially when analyzing data.

Quickly jog down the notes I learned...

There are two important categories of variation: Special Cause & Common Cause

Special [typically large impact]- out of control

- Temporary

- Often avoidable

- change to a process

- usually a definable outside cause

ex. Accidents/ strike/ extreme weather/ traffic jam/illness.hangover


Common [small impact/cause]- in control (can be partly special cause too)

- inherant in process

- predictability

ex. traffic lights/ vol of traffic/weather(vs.extreme weather)/parking/ physical condition (vs.illness)/driver mood


As we can find there will be a natural limit, a degree of predictability in the causes

*Six sigma is designed to drive out the special causes, and be able to predict and control the process, lower the waste, etc.

six sigma ex

** When analyzing the data, we need to run 2 CONTROL CHARTS:

Control chart

June 2023

Mo Tu We Th Fr Sa Su
May |  Today  |
         1 2 3 4
5 6 7 8 9 10 11
12 13 14 15 16 17 18
19 20 21 22 23 24 25
26 27 28 29 30      

Search this blog



Most recent comments

  • You're certainly right Srisuda (Keng). It would be the ideal way to be able to achieve that. I was j… by Lavinia on this entry
  • Great lavinia, I quite agree with you that leadership is very important which impact whether bad or … by on this entry
  • True… But I believe and hopefully can create an environment where people are not only engrossed in… by on this entry
  • In any of the organization people always expect something in return from the organization if they ar… by annawoods04 on this entry
  • Great post Lavinia! A wise person once said, "Many people seek to figure out what things mean, what … by Michael on this entry
Not signed in
Sign in

Powered by BlogBuilder