November 2008

Final Say feedback

I appreciate George Buckbee's article in InTech (August 2008 Final Say). I suggest "increases profit margin" instead of "reduces operating costs" as a better term for paragraph three describing improving control system performance. Yet even this doesn't capture the benefits from improved safety, health, and environmental performance.

I suggest that your first metric is often misused and therefore a bad one: "What percentage of control loops are running in MANUAL affecting both efficiency and quality."

I've seen the data used to lie. Quality, production rate, raw material efficiency, energy efficiency, you name it, show undesired behavior during these times.

It's too easy and misleading to use these data to show that when the loop is off control that performance suffers. Unfortunately most managers aren't savvy enough to recognize this.  The cause and effect relationship gets abused, and the manager gets told "All you have to do is improve loop utilization and you'll improve performance."  Nothing could be further from the truth. The process isn't upset because the loop's off control.  The loop is off control because the process is upset (sometimes because of what the loop did when it was on control).

Focus, as you say, should be on the bottom line, not on closed loop utilization.  Possibly eliminating the cause of, or reducing the effect of, the upset would be a more appropriate effort.   

Your real world examples are classics.  Basically reducing the magnitude of process variations so that the process can run closer to constraints.

Working in process control is a most often a conundrum. Either one works in a production department or in a computer/IT department. I've worked in both.

In a production department, nobody (OK, there might be some exceptions) has the luxury of implementing a process control solution, and besides, there's another department that does that work.  There are control standards and safety reviews to learn and adhere to. There's the control platform to learn, and an introductory level of knowledge might be enough to be dangerous. There's documentation and interfacing with and training users to be done. 

In a computer/IT department, the manager is relatively insensitive to the needs of the production department, because the computer/IT manager's evaluation is heavily weighted by other areas.

Your firm, ExperTune, Inc. sounds like a process control consulting firm, which is a good place for your expertise.  When your firm gets involved, production management is probably already involved. The problem is important enough for them to authorize spending money outside the corporation.  With a favorable result, your contribution will be recognized, and your business value will get demonstrated.  No working in a vacuum here. 

Robert H. Wilusz

Response:

Thanks for your reply.  You make some very good points that I was unable to address in the short one-page article.

"Time in Normal mode" is an excellent high-level indicator, and you are right on the mark in stating it does not tell the whole story. "Time in Normal" is more a symptom of many underlying problems. When an operator mistrusts the performance of a loop, because of an instrument, valve, process, or control problem, the loop is usually removed from its normal mode of operation. So "Time in Normal" is a good indicator of overall health, but it is not very specific about where the problems lie. There are many online metrics that can be used to provide more specific, actionable advice. These include metrics for oscillation significance, integral of error, average error, opportunity gap, valve at limit, and so on.  ExperTune's PlantTriage software manages over 70 such metrics for every control loop.  If engineers follow the right metrics, they can stop wasting time working on the wrong problems.

The link between process control metrics and business results is very real and has been demonstrated repeatedly.  Metrics that do not lead to business results should be abandoned, so that the focus can be placed in the right place.  The primary point I wanted to make was engineers must be aware of how to measure their results, in business terms wherever possible.

Thanks again for your interest and thoughtful response. 

George Buckbee

Outsourcing errors

Hans Baumann in his article "Outsourcing: What is it good for?" (July 2008 InTech) makes multiple economic errors.

He discusses the difference between domestic and foreign outsourcing. However, there is no difference. They are examples of a fundamental principle of comparative advantage. Others specialize in areas I am less qualified in, such as dentistry or tailoring for example. Therefore, I must, if you like, outsource. The resulting increase in wealth to both parties occurs independent of political boundaries. All voluntary transactions are undertaken only if both parties expect to benefit. 

With outsourcing, we see jobs lost and industries shut down. We know where and how many jobs have been lost. What is unseen and unknown is where the new jobs will arise from, in what fields, and when. Throughout history, lost jobs and factories have been replaced by better paying jobs in new areas, and in most cases were unforeseen at the time of the losses. One should read Frederic Bastiat's tale of the Broken Window (http://bastiat.org/en/twisatwins.html#broken_window) where he blames these myths on focusing only on the things "seen."

Baumann said foreign outsourcing increases the national debt. Perhaps he has in mind the mythical problems of Trade Deficits. There is no such thing as an over-all trade deficit. The balance of payments always balances.

Baumann goes on to say we are not beginning to see a reversal in outsourcing. The U.S. is simply enjoying continued economic growth. U.S. manufacturing revenues (adjusted for inflation) reached their all-time high in 2006. 2006 was also a peak year for inflation-adjusted manufacturing profits in the U.S. and for inflation-adjusted U.S. manufacturing exports. And the U.S. accounts for the largest share of the globe's manufacturing output; Americans, today, produce 2.5 times more manufactured goods than do the Chinese. (http://www.freetrade.org/AnswersToTheManufacturingQuiz)

Tod Martens