Question from September's Chemical Processing
We operate an acid regeneration plant (ARP) for a steel mill. In the process, FeCl2 solution is oxidized to Fe2O3 and HCl in a spray roaster; the HCl is recycled to the mill and the oxide is agglomerated and pelletized for feeding back to the mill (Figure 1). Currently, operators do laboratory tests to monitor the quality of HCl (target: 18.5%). The absorber hits an azeotrope at 20.4% at which time the scrubbers become absorbers and we will be fined for an acid emission. Someone suggests monitoring the concentration of the acid in the absorber. e.g., by inferring concentration by measuring conductivity or density, or measuring directly by using an automatic titrator. Automating the absorber also is appealing because this loop requires constant attention, especially during start-up and shutdown. What should we do? Keep in mind that ARPs are water-restricted.
Can you support the titrator?
Given the choices named, I would go with the automatic titrator. Density cannot be easily correlated with a Coriolis meter because it depends on temperature and concentration. We found that the best you can do is about ±1% with a Coriolis meter and this just isn’t good enough for contracts with steel companies. They need ±0.5% concentration or better.
A few years ago, I looked at program requirements for using conductivity. The trouble is that the control loop would require constant attention, the set-up would be difficult and the curve relating conductivity to concentration is parabolic. The peak is near 20.4%, the azeotrope. I considered programming that would track the concentration but it became too complicated. So, the least troublesome choice would be the automatic titrator.
The newer models of titrators run about $40,000 installed and do self-diagnostics. They have configuration that permits the frequency of the testing to be run out if the process is stable. You can expect a few problems if you try to duplicate operator testing.
The standard isn’t absolutely stable. We used trietholamine (TEA) as a standard. TEA was chosen because it burns up in the roaster. After sitting around for a few weeks, the potency drops. The result: operators must still do the test ever day or so. If you go with caustic, the standard should be more stable but it can’t be dumped into the pickle liquor sumps, like the TEA, because sodium will contaminate the iron oxide produced in the roaster. The waste from a caustic test must go to the mill wastewater treatment, which is usually restricted for a process handling wastewater.
Another problem is the specific gravity. You could use the daily test to verify the density with a Baume bulb gauge or by weighing on a digital scale. The best approach might be a curve plot of HCl concentration; the titrator could be configured to report the second calculation. One of our engineers tested a curve and it worked well.
By far, the worst difficulty will be reliability. I am a fan of good technology. Used in the right environment it’s wonderful — but a lot of plants don’t spend enough on plant instrument maintenance to support this equipment. Complicated instrumentation becomes more of a liability than a help in the end. I remember — not so long ago — trying to resolve instrument problems overseas. Imagine trying to explain nodes, networks and redundant data paths to them. A 4-20 mA loop is easier to explain and fix — plus it affects only one instrument (unless the wires are shorted in which case it could be a whole card!). It’s often just as bad in the states.
A race car serviced by “shade tree mechanics” isn’t even going to finish let alone win. This philosophy is what concerns me with the automation of the absorber system.
John Brockwell, consulting instrument engineer,
APT technical services, Coraopolis , Pa.
We tried conductivity — not recommended
One of my primary functions is to test the chemistries in these various washers, and to recommend cleaning product additions or tank changes where applicable. To minimize the amount of time taken each day to titrating each tank, it was once proposed that we put several of these tanks on recirculating loops with conductivity meters. The idea was that the conductivities of the various solutions would, once calibrated, give us a daily tank concentration reading for each tank. We believed that this would save time testing tanks. However, this idea was quickly abandoned because of its complexity. Conductivity can be influenced by three solution characteristics, namely: temperature, pH, and dissolved solids concentration. With three potential contributors to conductivity, it becomes difficult to pin down exactly which one is fluctuating at any one time, even with other instruments.
Since the accurate control of the operation of your scrubbers is so critical, I would recommend auto-sampling and/or auto-titration as your preferred test method for measuring the HCl concentration in this stream. It may be a bit more expensive, but it’s also more reliable for testing streams at extreme pH, such as the HCl stream to which you refer.
David Todd site manager,
Cummins Diesel ReCon, Memphis, Tenn.
Automation can also boost morale
For efficient and economic operation of the ARP, control of the acid concentration is necessary. Thus, I would recommend that the possible control technology choices that you have outlined, and others that are available on the market should be compared and one should select the best operational and economic option. Savings during start up and shutdown operation along with need for constant attention will justify the expense. In addition, inclusion of an automatic control scheme would be a morale booster for the operational people. This has benefits that are difficult to dollarize in the justification.
Girish Malhotra, President
EPCOT International, Pepper Pike, Ohio
What about chromatography?
Perhaps high performance liquid chromatography (HPLC) could work. There are selective columns used to remove chlorides prior to isolating organics. If a gel selective to chlorides can be found, this method might be superior to those suggested like titration.
Furst Moore, retired engineer
Sun Oil, Hallsville, Texas
I’d make a home made densitometer. I’d take a small slip stream off the circulation pump to the top of the scrubber through a globe valve into a home made densitometer made from plastic lined pipe. The densitometer would be made from a vertical 3-in pipe, say, 50-in tall with the high and low taps for an electronic dP cell installed at 0 and 40 in. The bottom outlet of the densitometer would go to a loop seal back to the process. The loop seal would be designed to keep the densitometer level constant at 40 inches. The 4-20 mA output of the dP cell would be calibrated with hot 15% HCl and 20% HCl solutions. Hopefully the plant could use this densitometer to set up a trim controller on the water make up to hold 18.5% HCl. High alarms could be set to prevent operation above 20% HCl. This system should be insulated as temperature affects all the online properties that are being measured (pH, cond, titration, etc.).
If space isn’t a problem, the height of the column could be increased to improve the resolution. A capillary-filled dP cell with sensitive waffle heads should work in this application. I learned this trick from Catalytic Cracking units. It was used to get a “pseudo density” for control. I believe the “pseudo density” could be correlated to the actual density for Title V environmental compliance. Sometimes low tech is better.
Accumulation of junk in the column could be a problem. The answer to this is simple. Install a second pump-around wash zone scrubber downstream of the existing one. Fresh water could be feed to the top wash zone and totally drawn off for recirculation. This will eliminate the environmental problem. This recirculation water could be feed as make-up water to the original scrubber. You should draw off 18.5% HCl the original scrubber to a dual tank system where periodic average samples could be manually analyzed and HCl blend draw rates adjusted. While one tank is being transferred to the mill the second could be filling.
Timothy Goebel, engineer
Citgo, Nederland, Texas
During our plant shutdown a flow control valve was replaced (Figure 2). The globe valve charges a reactor used for making a volatile organic chemical, an intermediary to a polymer. The previous valve served faithfully for several years, though we had trouble with it once because the heat tracing failed. The board operator said that flow dropped to about half, which took longer to fill the reactor. Although the new control valve served well for a few weeks it is now sticking — and it is getting worse. For some reason, the flow problem has also returned. Obviously, we need to pull the valve. What should we be looking for?
Send us your comments, suggestions or solutions for this question by December 12, 2006. We’ll include as many of them as possible in the January 2007 issue and all on CP.com. Send visuals — a sketch is fine. E-mail us at ProcessPuzzler@putman.net or mail to ProcessPuzzler, Chemical Processing, 555 W. Pierce Rd., Suite 301, Itasca, IL 60143. Fax: (630) 467-1120. Please include your name, title, location and company affiliation in the response.
And, of course, if you have a process problem you’d like to pose to our readers, send it along and we’ll be pleased to consider it for publication.