THIS MONTH'S PUZZLER
Our pilot-plant batch reactor uses a new fiberglass packed-bed caustic scrubber for final polishing of vented gas before it goes to a thermal oxidizer. The design pressure is 150 psig for the reactor. The vessel is protected by a 4 in. × 3 in. pressure relief valve set at 35 psig, which we had sized assuming a heater coil valve-open scenario using steady state. The pressure in the reactor frequently spikes above 28 psig — which has blown the relief valve several times. We rely on the pressure drop through a throttling valve and knockout drum to prevent the vent gas from overpressurizing the scrubber, which typically operates with a discharge pressure of 1 to 2 psig. Now, management wants to use the reactor exclusively for heat-of-reaction measurements. We're unsure about how to isolate the fiberglass scrubber in the event of a runaway reaction and how to properly size the reactor's relief device. What's your advice?
The choice of fiberglass is unfortunate. At best, it can withstand a vacuum of 2–3 psia and a pressure of 3 psig. Also, fiberglass is a fire hazard. It's worth pointing out that ASME didn't even write an effective standard for fiberglass vessels until 1989: ASME RPT-1, "Reinforced Thermoset Plastic Corrosion-Resistant Equipment." API provided a standard a little earlier: the first edition of API 12P, "Specification for Fiberglass Reinforced Plastic Tanks," came out in 1986. This rates tanks for up to 15-psig pressure, but experienced engineers limit loading to 3 psig. NACE would support the use of carbon steel for typical scrubber caustic concentrations of 1–10 wt.% with heats of solution of about 212°F; steel should be stress-relieved at welds and bends. Glass lining also might be acceptable, especially where oxygen and acid meet.
This system is already in violation of the ASME code because it uses a throttling valve to protect a fiberglass vessel. Throttling valves aren't intrinsically safe and have a higher failure rate than pressure relief valves. In addition, the scenario used for a heater coil failure demands dynamic analysis; the resulting relief fluid could be two phase. The only acceptable correction is to resize the reliefs, decide if the knockout drum and throttling valve really are necessary, and replace the fiberglass vessel with a metal one.
How can you resize the reliefs if you don't know what the heat of reaction will produce? The last time I was asked to size a relief for a pilot vessel of this sort, the scientists told to be size for the "BDH" — biggest damn hole! A 4-in. rupture disc, if large enough, might be a practical replacement; otherwise the vent will be too small. One way to justify reducing the size of the relief is by dynamic modeling. This is allowed under the guidelines of DIERS [Design Institute for Emergency Relief Systems] and other acceptable standards.
There are other practical methods to avoid problems from an under-sized vent: 1) run the reaction at low and extremely low volume — this is commonly done with reactions involving acetylene; 2) upgrade the controls to detect and, if necessary, quench or kill the reaction; 3) calculate the maximum heat generation that can be tolerated by the relief and restrict pilot tests to those below this value — DIERS and vent sizing packages describe bench tests to define risks with smaller quantities of the reactants. The only effective way to catch a runaway reaction is by pressure; temperature is too slow to respond and easy to misinterpret.
Dirk Willard, lead process engineer
Fluor Global Services, Inver Grove Heights, Minn.
Our refinery uses a thermal oxidizer (TOX) to dispose of vacuum off-gas. Currently, a steam jet pulls this off-gas because we've had reliability trouble with liquid ring pumps. We've suffered severe corrosion in the TOX's burners and tiles. Startup problems with the TOX — particularly with the fire detectors and igniter system — have caused delays. What's causing the problems and how can we improve the unit's reliability? We also have a related concern: our production manager wants to send the off-gas to a flare while the TOX is out of service for repair, which will take a week. The typical composition of the off-gas by volume is 30% H2S, 5% CO, 2.75% H2, 2.25% N2, 1% CO2, and 0.5% O2, with the remainder C1–C3 hydrocarbons. The vacuum is 40 torr, absolute, and the temperature is 120°F. About 2,500 pounds per hour would go to flare while the TOX is offline. Do we risk damaging our flare? What other problems might arise? What temporary measures would you suggest to avoid environmental issues?
Send us your comments, suggestions or solutions for this question by December 14, 2012. We'll include as many of them as possible in the January 2013 issue and all on ChemicalProcessing.com. Send visuals — a sketch is fine. E-mail us at ProcessPuzzler@putman.net or mail to Process Puzzler, Chemical Processing, 555 W. Pierce Road, 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.