Butterflies Are Free

Aug. 22, 2003

Why boosters spread the butterfly's wings, more Purple Reign, and a new puzzler: The Tripping Furnace

Stan: The June Puzzler "An Open and Shut Case"--where a butterfly for surge control slammed shut when a booster was used instead of a positioner--left most people waiting in the wings.

Greg: Before these readers take flight to do e-mail, how could they leave unquestioned a theory so widely stated and accepted that boosters instead of positioners should be used on fast loops?

Stan: I guess most people ignored it and never really get into the use of boosters. But bottom line is, we didn't get a single correct answer. No lavish gifts this month!

Greg: I should have been suspicious when the recommendation was based on a mathematical model with the usual abundance of simplifications to provide an exact solution.

Stan: Sounds like the mathematical models of retirement presented by financial advisors, and based more on hype than reality and optimized for the success of the advisor.

Greg: It has made me more sensitive to the need for testing on different combinations and applications before a theory is stated as an irrefutable rule. It turns out most boosters are designed to have an extremely high sensitivity in the outlet port where a change of a small fraction of an inch of water column will cause the booster to change its flow rate. When this outlet port is connected to a diaphragm actuator, a small amount of force up or down on the shaft causes enough of a compression or expansion of the volume to trigger the booster and create positive feedback. Believe me, I know it is a small force because I can grab the shaft of a 24-in. butterfly connected to a booster and stroke the valve. I have done this not only in the field on compressor surge control applications but also in a shop when I tested butterfly valves for furnace pressure control. The actuators were properly sized in both cases. If I can stroke the valve, a small amount of turbulence and unbalance on the disc can slam it shut, a very undesirable state for a surge valve protecting a $10 million dollar compressor from destruction. One of readers, Mark Chatterton, also brought up a good point that the reversal of pressure differential during surge can make the problem worse. It is a moot situation for piston actuators, because a positioner must be used anyway to load and unload the opposing sides of the cylinder.

Stan: If a booster must be used to speed up a large valve, it should be on the output of a positioner with a bypass valve adjusted just far enough open to eliminate the inevitable limit cycle. However, the improvement in stroking time will not be as dramatic as the increase in flow coefficient (Cv) of the booster compared to the positioner unless the size of the actuator connection, solenoid, and air piping is increased so that these restrictions are negligible. While the exhaust and supply Cv are not equal in a booster they are not as disproportionate as in a "dump or quick-exhaust" valve, which are notorious for fast but discontinuous action.

Greg: In case you also want a boost in your monthly general education process that started in the July issue, we offer our next installment of choice questions. The first four questions were in the previous issue.

More Questions from the Purple Reign

5. For a transmitter, the most important consideration is:

A. Cost

B. Smart features

C. Color

D. An undistorted view of small changes in the process

6. For a control valve, the most important consideration is:

A. Cost

B. Smart features

C. Color

D. Ability to make small changes to the process

7. The control valve performance spec that has the worst effect on loop performance is:

A. 10% nonlinearity

B. 3% dead band

C. 2% stick-slip

D. 3% offset or bias

8. A control valve without a positioner is:

A. A good cost savings

B. A good way to simplify the installation

C. Stupid

D. More fun than engineers should be allowed to have

9. The control valve with the best linearity is:

A. Sliding stem valve

B. Butterfly valve

C. Ball valve

D. Gate valve

Stan: Last month's Purple Passion explanation of advanced control reminds me that when I was going to engineering school, the school sent me to "The Experimental Lab." At this wonderful federal government facility, I had the opportunity to test petroleum products that were being used in our nation's defense. One of the assignments was the testing of the chemical components in oils. In order to do this work, we needed pure ethyl alcohol, which required a great deal of red tape to acquire. The chief chemist had developed a method of removing the nasty chemicals that make denatured alcohol into the good stuff. This feed stock was pumped into a series of filters, reactors, and distillation columns that ran 24/7. At the end, pure 190-proof ethyl alcohol dripped into a container. Unfortunately, this process produced far more alcohol than was needed for our work. In order not to waste this valuable product, we bottled it and gave it to deserving people. The story goes that a certain co-op college student took the stuff to the fraternity house where the brilliant engineers converted it into a variety of punches...but that is another story.

Greg: To further stimulate our readers we pose the following puzzler. A furnace pressure control loop uses a variable-speed drive on an induced draft fan instead of a control valve or damper for pressure control. The pressure can ramp off scale in 0.25 seconds if the loop is placed in manual. Why did the furnace keep tripping on startup after the analog electronic controllers were upgraded to a distributed control system?

Stan: Now for this month's disclaimer. We once actually expected raises based on performance.

Quiz answers: D, D, C, C, A

This Month's Puzzler:

The Tripping Furnace

A furnace pressure control loop uses a variable-speed drive on an induced draft fan instead of a control valve or damper for pressure control. The pressure can ramp off scale in 0.25 seconds if the loop is placed in manual. Why did the furnace keep tripping on startup after the analog electronic controllers were upgraded to a distributed control system?