Wes Fulton is founder and CEO of Fulton Findings (TM) which he started after a long career in aircraft engineering, touching on many "wow" projects like the Indigenous Defensive Fighter (IDF) leading edge flap actuation system (LEFAS) development and production, the Rockwell/MBB X-31A LEFAS flight test program, and the F-16 Fighting Falcon LEFAS production and deployment support.
He also co-patented a multi-fuseable shaft (high performance drive train device).
As if that wasn't enough he has 20 years of programming experience as a private programmer and developed the first widely-used Weibull software. An impressive track record and an obvious choice for our Profiles in Leadership series of interviews.
Q1. Weibull analysis seems to be going from strength to strength at the moment globally. Companies seem more willing and capable of implementing a Weibull and other probabilistic approaches today than (say) five years ago. Does this agree with what you are finding, and why do you think it is?
Answer: The technical community was reluctant at first to get past using 100% deterministic analysis (DA). An example of DA is where you calculate or estimate a single value (point estimate) for burst strength of a pressure vessel. But variability concerns make that single number suspect for use in newer applications. The old way of addressing this was to assign a "safety factor" (increase) to the expected load or to "derate" (lower) the strength by some arbitrary amount ... or both.
While this method worked reasonably well when re-using established designs, it turns out to be too risky for implementing new designs that utilize innovative configurations and state-of-the-art materials. Plus it can be overly conservative leading to excessively heavy or excessively pricey products. We are seeing an increasing amount of new designs appearing in service today due to business competition. So, the Weibull engineering way of modeling variability is gaining in popularity as safety factor use and derating become obsolete.
Q2. The war for talent is a double edged sword right now. A difficult time for companies, and a wealth of opportunities for talented individuals. With all the training and consulting you do, what tends to separate the leading professionals from the rest in your area? And, do you find difficulties in locating good talent for your own consultancy?
Answer: For all technical analysis, especially Weibull engineering, useful results are based on some inputs and usually some simplifying assumptions. The people that carefully evaluate the inputs and scrutinize the assumptions are the best. You don't want someone that plugs-in questionable data entries and blindly trusts the output.
Q2. The war for talent is a double edged sword right now. A difficult time for companies, and a wealth of opportunities for talented individuals. With all the training and consulting you do, what tends to separate the leading professionals from the rest in your area? And, do you find difficulties in locating good talent for your own consultancy?
Answer: For all technical analysis, especially Weibull engineering, useful results are based on some inputs and usually some simplifying assumptions. The people that carefully evaluate the inputs and scrutinize the assumptions are the best. You don't want someone that plugs-in questionable data entries and blindly trusts the output.
I personally work with great people. I consider some people I work with brilliant (Dr. Bob Abernethy, Paul Barringer, Carl Tarum, Dr. Allen Liu, Dr. Todd Marquart, Jorge Granada, Dennis Keisic, and others). So I have not had difficulty locating good talent. However, in most cases people need second and third opinions to get it right. If you are searching for help with Weibull engineering, you first ought to look for someone with practical hands-on equipment operation experience. Those people know variability exists. Their Weibull engineering talent then comes easily with training.
Q3. What is the single largest impediment to implementing a robust Weibull approach and what advice could you give others to tackle this?
Answer: The #1 hardest thing to do is to acquire the appropriate data from equipment operation. Theoretically, this is not hard. But in reality, many data-recording systems are not properly established in the beginning, and also errors occur in the data-recording process. This can be easily fixed at the design stage, fixed with significant effort during start-up operation, and downright very costly and difficult to change if attempted later than that. What data do you need for Weibull engineering? You need the age on each relevant piece of equipment sampled, whether it is still operating successfully (called a suspension) or it has failed (called an occurrence).
Q3. What is the single largest impediment to implementing a robust Weibull approach and what advice could you give others to tackle this?
Answer: The #1 hardest thing to do is to acquire the appropriate data from equipment operation. Theoretically, this is not hard. But in reality, many data-recording systems are not properly established in the beginning, and also errors occur in the data-recording process. This can be easily fixed at the design stage, fixed with significant effort during start-up operation, and downright very costly and difficult to change if attempted later than that. What data do you need for Weibull engineering? You need the age on each relevant piece of equipment sampled, whether it is still operating successfully (called a suspension) or it has failed (called an occurrence).
Other information like the root cause of failure for occurrences is also extremely useful so you can analyze only one root cause at a time. If you have to analyze multiple root causes with one analysis, then either the analysis gets overly complex or your results are not as useful for determining corrective action if needed.
Q4. As I read it, Dr Bob and you have advanced the method in quite a few areas. Are there remaining areas that you think could still be advanced? And on a personal level (As a research junkie myself) how do you manage to continue research, manage a business and have a personal life to boot?
Answer: I am very fortunate to have worked with Dr. Bob Abernethy for over 20 years now. The more we discover, the more future work seems to appear. Some future work is needed to look at the uncertainty (potential error) associated with popular complex methods like estimating confidence bounds on accelerated-testing projections. You first gather age-to-failure data at two or more different load levels. Then you can then estimate the effect of load on lifetime (higher load generally reduces lifetime).
Many applications use a power function or an Arrhenius function for the effect of load on lifetime. Then there is the extra step of adding confidence bounds on estimates of lifetime capability at untested load levels. In my opinion most analytical software now on the market grossly underestimates potential error in such results. On the very different subject of continued research, business management, and personal life ... I invoke the 5th amendment. My wife would prefer otherwise, but I spend a lot of time on the business. I am hoping that personal cloning becomes affordable soon.
Q5. What can we expect from Fulton Findings over the next (say) 12 - 24 months in terms of new markets, new areas, or other business areas?
Answer: Jorge Granada and I have launched the first (to my knowledge) web-based Weibull engineering application at http://www.riscape.com called "Rapacci". Access to Rapacci is currently free so we can introduce the concept and improve the service, before going commercial. Please feel free to check it out. You just have to set up a User Account first. I'm positive that all analysis will be done this way, over the internet, eventually.
Q4. As I read it, Dr Bob and you have advanced the method in quite a few areas. Are there remaining areas that you think could still be advanced? And on a personal level (As a research junkie myself) how do you manage to continue research, manage a business and have a personal life to boot?
Answer: I am very fortunate to have worked with Dr. Bob Abernethy for over 20 years now. The more we discover, the more future work seems to appear. Some future work is needed to look at the uncertainty (potential error) associated with popular complex methods like estimating confidence bounds on accelerated-testing projections. You first gather age-to-failure data at two or more different load levels. Then you can then estimate the effect of load on lifetime (higher load generally reduces lifetime).
Many applications use a power function or an Arrhenius function for the effect of load on lifetime. Then there is the extra step of adding confidence bounds on estimates of lifetime capability at untested load levels. In my opinion most analytical software now on the market grossly underestimates potential error in such results. On the very different subject of continued research, business management, and personal life ... I invoke the 5th amendment. My wife would prefer otherwise, but I spend a lot of time on the business. I am hoping that personal cloning becomes affordable soon.
Q5. What can we expect from Fulton Findings over the next (say) 12 - 24 months in terms of new markets, new areas, or other business areas?
Answer: Jorge Granada and I have launched the first (to my knowledge) web-based Weibull engineering application at http://www.riscape.com called "Rapacci". Access to Rapacci is currently free so we can introduce the concept and improve the service, before going commercial. Please feel free to check it out. You just have to set up a User Account first. I'm positive that all analysis will be done this way, over the internet, eventually.
No comments:
Post a Comment