Discussion Central

  • 1.  Process improvement and six sigma; is it for chemical engineers?

    Posted 07-23-2019 08:00
    Hey there!
    I saw a job posting in a company I want to join requiring six sigma green belt certification. I was not aware that such certification and it took me a while to learn more about the subject. My question is do you think it is the job of chemical engineer to be six sigma certified and applying it in his/her daily job. Isn't that an industrial engineer's function. If so why is not implemented in any ChE curriculum I saw and how do I learn these topics and start using them in my job as a chemical engineer. 
    Thanks

    ------------------------------
    Omar Abdelaty
    Ankara
    ------------------------------


  • 2.  RE: Process improvement and six sigma; is it for chemical engineers?

    Posted 07-23-2019 09:29
    Good morning Omar.  I will attempt to answer your question but will also undoubtedly leave some gaps in my response that others may better fill.

    There are several different types of Six Simga programs.  The Lean Six Sigma (and Toyota is often mentioned) is what you are thinking of it seems on industrial processes and manufacturing techniques.  Terms like "Spaghetti Charts" are useful to understand not only in the industry, but also in how you arrange your kitchen on the cooking and cleaning cycles.  Helpful, but not really relevant to chemical engineers.  It can be, but it's not a perfect match in my mind.

    There's also "MAIC" (also sometimes called DMAIC, LMAICL, LDMAICL, etc...) but at the root of it is Measure, Analyze, Improve, & Control.  L is for Leveraging (both into your project and afterwards to other projects).  The D is for "Define"...define the problem.

    In these stages are processes like determining whether your measurement system is adequate and able to define the level of accuracy required to determine whether a defect exists, analyzing the root causes of the defects, implementing improvements to the system, and then installing a control system which will sustain the gains of improvements and not "fall back to the way we used to be".  

    This is very relevant to chemical engineers.  I'm a chem. E (obviously), I got my Black Belt at a major chemical company, and am always very appreciative of that training.  I have used it many times across my consulting / engineering company days even after leaving the chemical company world.  Where I used it (2 examples only) within the chemical company are:
    A). Specialty chemical injection for a mining water application in salt domes--saved $80k / yr, and continues to operate today (> 10 years later)
    B). Pump discharge pressure on VFD's vs. required pressure.  Saved over $100k/yr and continues to operate today (>11 years later).

    Both were interesting but from a chemical engineer's standpoint, neither was terribly complex.  The pump system kind of was, but in my mind was simple hydraulics issues and valve positioning feedback.

    I've seen some Six Sigma storyboards where the chemical engineers dive into the details and savings exceed $1MM/month by defining defects in the process and using statistics to identify the root cause, which may or may not include operations procedural issues.

    I encourage taking courses to LEARN about it, but the best learning is doing it and being mentored.  I don't like when companies say they "require" a certification, as the culture and practices at each company may not directly leverage to other places.  But if that's the door you are facing, then taking an online course is likely best to learn the material, and then apply it first chance you get.  

    By the way--I do NOT have recommendations as to which courses to take...I hope others will reply.

    Hope this is helpful.

    ------------------------------
    Eric Parvin PE
    Owner / Process Manager
    Highlands Ranch CO
    ------------------------------



  • 3.  RE: Process improvement and six sigma; is it for chemical engineers?

    Posted 07-25-2019 00:14
    This sounds a lot like what we used to call Statistical Process Control, with control charts and the like…  Am I seeing this correctly?

    ------------------------------
    Steve Cutchen
    Investigator
    US Chemical Safety Board
    Houston TX
    ------------------------------



  • 4.  RE: Process improvement and six sigma; is it for chemical engineers?

    Posted 07-25-2019 09:52
    Steve, you are absolutely, 100% correct.  Statistical control charts show the random variance of a process measurement.  Random variance is caused by many factors including the lag of instrumentation and even the weather.  That variance can be increased if one "chases the needle" instead of waiting until the change of the measurement value exceeds the random variation.  This is particularly true when the instrumentation response is slow.

    Previously, the AIChE Academy offered a course, ELA-128,  Statistical Control Charts.  It showed process variation by the use of a Quincunx Bead Machine to generate random data.  A background in statistics was included as it was not required.  The appropriateness of different types of control charts were explained.  The course included basic statistics, Run Charts, process capability, the Shewart type charts for Gaussian distributions including the X-Bar and R Charts, X-Bar and S Charts, Individuals Charts, Zone Charts, Exponentially Weighted Moving Average Charts and Cumulative Sum Charts, along with Attribute Charts for processes with a binomial distribution.

    ------------------------------
    T. David Griffith, Ph.D.
    Blessing (Bay City), TX
    ------------------------------



  • 5.  RE: Process improvement and six sigma; is it for chemical engineers?

    Posted 07-24-2019 00:33
    Omar, Six Sigma certification is well worth the effort if you are involved with process operations, research and a number of other areas.  The main thing is if you need to evaluate operations which have variance.  In graduate school, I took a course called "Statistics for Experimenters" which was  taught by William G. (Bill) Hunter who had a B.S. and M.S. in Chemical Engineering and was the first graduate student of George E.P. Box at the University of Wisconsin-Madison who developed the course while he was at Princeton before going to Wisconsin in 1959 to start the Statistics Department.  Finally, George and Bill joined by J. Stuart Hunter, a Civil Engineer at Princeton, published Statistics for Experimenters.

    The only criticism I have about taking the Six Sigma Green Belt course was that I ended up teaching the instructors about the use of "aliases" and "generators" when specifying fractional factorial experimental designs.  They had never heard of them before.  For me, it was a waste of time.  If one does not have the course in school, then being certified would be of benefit for the employer.

    For a practical example, consider a chemical engineer who monitors the production of widgets.  There will most likely be variance which causes rejects.  The question is whether the frequency of rejects is changing.  A simple "t-test" will answer that question.  That analysis was developed in 1908 by William S. Gossett who was a chemist for the Guinness brewery in Dublin, Ireland.  He published his technique under the name of "student" so that Guinness would not be upset.  Brewing beer or ale is a chemical process.  And quality control requires statistics.  Undoubtedly, there are many other processes where chemical engineers can apply statistics.

    ------------------------------
    T. David Griffith, Ph.D.
    Blessing (Bay City), TX
    ------------------------------



  • 6.  RE: Process improvement and six sigma; is it for chemical engineers?

    Posted 07-24-2019 07:36
    I feel that a minor PS is important here.
    If you are using statistical analyses, do not get exclusively hung up on Normal nor Student's T distributions.  They are perfectly acceptable and applicable in most (maybe 80%) of the situations you may encounter.  Because I'm in the environmental business, I find that for things like strength of materials, and environmental discharge and other "natural" processes, the Weibull distribution is a better fit for the following reasons: 1) With a Weibull distribution it is impossible to get "negative" numbers -- those below zero on the process, 2) there is a statistically robust method for calculating using Excel and other web resources; and 3) the Weibull has a fatter right tail which can help account for process upsets.
    Lest I get too far into the weeds, consider  Nassim Taleb's book: The Black Swan and while it is about financial dealings and uncertainty, it has definite implications for process and plant activities.

    ------------------------------
    [David] [Russell] [PE]
    [President]
    [Global Environmental Operations, Inc.]
    [Liburn] [GA]
    ------------------------------



  • 7.  RE: Process improvement and six sigma; is it for chemical engineers?

    Posted 07-28-2019 10:20
    Hi

    I am a chemical engineer and I also went through a certification course on Black Belt by ASQ. It is good to know all the quality tools as they help in any situation, be it engineering problem or personal problem. But don't stretch it too much. I remember, one young chemical engineer tried to solve a quality problem in a distillation column with six sigma tools and after long hours or work, found that increasing the reflux is the best solution.

    R. Muthukrishnan
    Vadodara, India.

    ------------------------------
    Muthukrishnan Rengarjulu
    Baroda
    ------------------------------



  • 8.  RE: Process improvement and six sigma; is it for chemical engineers?

    Posted 07-28-2019 20:57
    Muthukrishnan, the overkill use of Six Sigma (statistical) tools is comparable to using Transport Phenomena to design a heat exchanger.  (That's an old story I heard decades ago from Joel Hougen who was Olaf Hougen's little brother.)  Both indicate that the engineer does not understand the applicability of the technology.

    Statistical analysis is absolutely required when a process has a lot of random variation.  A "run chart" can substitute for an analysis of variance (ANOVA) to determine if something like a t-Test would be needed.  If the process has minimal variance, then just make a change and go.  If you don't know the variance which Six Sigma should teach one to determine, then such a change could be disastrous.

    ------------------------------
    T. David Griffith, Ph.D.
    Blessing (Bay City), TX
    ------------------------------



  • 9.  RE: Process improvement and six sigma; is it for chemical engineers?

    Posted 07-29-2019 06:57
    Hi Omar, having a certification is always an advantage. Six sigma certification helps in any field from manufacturing, service, finance. You can refer to ASQ for the details of six sigma certification. There is lot of research and literature available on Six Sigma.

    ------------------------------
    Sachin Mahajan ING
    Senior Analyst Production Planning
    ADNOC Onshore
    Chorniche, Abu Dhabi
    ------------------------------