By Greg McMillan
and Stan Weiner, PE
Stan: As usual Hunter Vegas nailed the answer to the Puzzler. In the August 2005 issue, we tried to stir things up with the question about why the standard mixing design practice of inserting a dip tube half way down into the mixture is a deadly practice for neutralization control.
Hunter: There are two reasons why I would avoid running a reagent dip tube half way into the tank liquid level:
- The typical flow rates of reagent in a neutralization to 7 pH can be very low. If the dip tube extends below the level of the liquid, the dip leg effectively becomes a reservoir that delays the addition of reagent to the vessel. (In other words, as the reagent is added it takes a long time to push the process liquid out of the tube and actually reach the vessel. Also, even after the reagent flow stops it still continues to be added as the concentrated reagent is slowly diluted by the process fluid from the vessel.) These lag times will make an already difficult to control process nearly impossible.
- The second problem is one of material compatibility. Low concentrations of acids can be very corrosive, and it will attack the dip tube as the acid becomes diluted near the end of the pipe. (If the pipe material is poorly chosen, it'll just corrode away until the pipe stops just above the liquid surface and all of your process dead time problems would be solved!)
Greg: When the reagent control valve closes, the reagent in the dip tube leaks into the vessel and process fluid migrates back into the dip tube. Eventually, the dip tube is back filled with process fluid that must be flushed out when the control valve reopens. The result is a leak that can persist for hours and a transportation delay on the restart of reagent flow that can approach an hour. For example, if the back filled volume of the dip tube is just 1 gallon and the reagent flow is 1 gph, the delay before you see the change in pH from reagent addition is more than an hour. The first time I ran into this was when we started sulfuric acid addition during water batching of a “well mixed” vessel, and there was no drop in pH after 30 minutes. We thought maybe the electrodes still had on their protective caps.
Stan: Electrodes last a lot longer if you keep on the protective caps.
Greg: The result of a lab sample was 5 pH. After I verified that the three electrodes, which indicated the vessel was above 7 pH, were actually seeing the process, I asked for three more samples. The new lab results all came in above 7 pH. The previous lone sample had been put in a beaker with residue of a 4 pH buffer. The vessel pH finally did drop after an hour. This convinced me to find better ways of reagent delivery, such as the injection of reagent into a feed or recirculation line that has a high flow rate, just before it enters the vessel.
Stan: We had a creative answer from Henree Pernash Locknester PhD, DD, MS, GED, a certifiable automation engineer.
Henree: When I first heard about this problem, the first thing that came into my mind was to connect a bicycle air pump to the top of the dip pipe. When the reagent control valve shut, a flashing light would tell the field operator to begin pumping. This would blow the reagent out of the tube. However, the more I thought about it, problems began to surface. The pump would only go to about 75 pounds, and suppose the operator was in the control room cooking dinner. Then you would have to automate the blow down.
Stan: Another situation that can put you over the edge is the use of pre-Beta software.
Greg: In pushing the limits of new advanced control tools I have the privilege of using software hot off the developers desktop. Sometimes while riding the leading edge, you find yourself in a freefall off a cliff. You just hope the landing is soft, or there is a trampoline at the bottom, or a bungee cord to bounce you back. Here are the Top 10 Signs.
Top 10 Signs You Are Over the Edge with Your Leading-Edge Software
- The number of undocumented features exceeds the number of new features.
- The software “talks back.”
- Your computer is experiencing “flashbacks.”
- Your computer says “it is tired now” and takes a nap.
- The software does unspeakable things in unthinkable ways.
- The software has an “X rating” (restricted to hard-core computer enthusiasts only).
- When you tell the support person the software version number, he/she says “say what?”
- When you tell the developer the software version number, he/she just laughs.
- Your computer’s hiccups scare you out of your hiccups.
- You have entered a new dimension—“The Twilight Zone.”
Caption on screen: I am tired now...excuse me while I take a nap.
Stan: We also are happy to say that there is happy outcome to Henree’s request for “A Job for an Unemployed Nice Guy.”
Henree: I really want to thank you guys helping me find a job. It will be designing complex control systems for an air-conditioning contractor in Southwest Florida. As soon as I get a preliminary design, it will be passed on to you guys for review. My friends would personally like to thank CONTROL magazine for showing my letter explaining the difficulties in finding a job after getting out of prison. I was kind of thinking about starting a contracting company like the one Stan worked for in St. Louis.
Editor’s Note: You can see the letter from Henree in “Other Voices” on ControlGlobal.com.
Greg: Keep those letters coming. We just got similar correct replies to the July 2005 puzzler on dissolved oxygen measurement from David Pierce and Patrick Robeck. So to complete this column, we include David’s answer.
David: The D.O. probes read higher in the lower nozzles because the water is deeper and the pressure is higher. Therefore the total pressure in the air bubbles is higher, the partial pressure of the oxygen in the bubbles is higher, and there is more driving force for oxygen to dissolve and more mass transfer.
This Month’s Puzzler:
The Case of the Batch Unicycle?
What are the biggest opportunities to reduce batch cycle time?
Send an e-mail with your answer to The Puzzler, CONTROL questions, or comments to ControlTalk@Putman.net.