This Month’s Puzzler
The 5-ft-diameter caustic scrubber at our refinery is supposed to capture H2S. We sell the sodium bisulfide produced; this decreases the disposal costs of caustic soda. In our last turnaround, we replaced 30 sieve trays with two beds of high-efficiency glass-filled random packing. This design should improve efficiency because the caustic now goes through a plate-and-frame chiller that allows us a much larger reservoir in the tower. However, since the turnaround, our bisulfide product has been contaminated with sodium carbonate, making it a struggle to meet our bisulfide sales obligations.
Inspection of the liquid distributors — which were retained from the old internals at the insistence of operations — shows they are level and not too badly plugged. A model of the tower indicates that the packing is underperforming. The H2S going to our vent system exceeds our permit.
The sales staff blames the current problem on upstream CO2 membranes, which were replaced during routine turnaround maintenance. Concerns have been voiced about sodium carbonate eventually fouling the plate-and-frame chiller. Unfortunately, I’m not sure about fouling in the exchanger because it lacks taps for pressure gauges and thermowells.
Operations personnel want their sieve trays back. They never were fans of packing and were against the change, asserting “if it isn’t broken, don’t mess with it.” They repeatedly argued for a larger pump to move the bisulfide; the plant control engineer always hated the level control in the scrubber.
I think something must be wrong with the packing. I’ve seen this problem following three or four years of operation but never just a few weeks after addition of new packing.
Do you think sales is right? If so, how do we prove it? Is there something wrong with the packing? Did we miss something with the liquid or vapor distribution? Should we go back to the sieve trays?
Set Up An Internal Team
On reviewing the description of the issue, it is interesting that the operation is telling me where the problem lies. From my perspective statement “since the turnaround, our bisulfide product has been contaminated with sodium carbonate, making it a struggle to meet our bisulfide sales obligations” tells it all. Changes made are the cause of the issues. Basis and rationale for the changes made and their basis is not elaborated. It seems there is significant conflict within the organization. A stepwise process to solve the problem is needed.
Operators who have to put up with issues resulting from the change would like to go back to an operation what produced a quality product. This is normal. I would too.
Asking outsiders the question “Do you think sales is right? If so, how do we prove it?” suggests that the company has internal partnership/trust issues. They are one team and need to solve the problem internally rather than say who is right or wrong.
Whatever solution and pathway to solve the issue is selected, it has to be clearly and systematically explained to all to make sure everyone is on board. Unless this happens, finger pointing will continue. Based on the problem description, I would consier the following path:
The company should compare its current operation to before the changes took place because this will give clues of before and after the change of operation. Because the carbonate level in bisulfide has increased, it suggests that CO2 level entering the caustic scrubber is much higher than the previous level thereby contaminating bisulfide. Carbon dioxide is being scrubbed along with H2S and contaminating the bisulfide.
Based on carbonate contamination, it seems that CO2 membranes are not working at the before-change level. The described problem does not mention anything about this potential issue. As stated earlier, the company needs to re-visit the operation records before and after change to pinpoint what has changed and needs to fix the change.
The company should also review the solubilities of carbonate and bisulfide at the concentration and operating temperatures as mutual solubility will influence operation. Current operating conditions should be similar to prior to equipment change conditions unless design changes were made.
I am not sure of heat exchanger placement. Colder liquid will lower solubility and would result in dropping solids out of solution and may cause fouling.
As mentioned earlier, the company has to review and investigate cause-and-effect relationships. Just the feeling that spiral exchanger is better than plate and frame or spiral exchanger is better can all be resolved if the design criterion is properly defined and used for the design.
There are options and the company has to regroup internally to consider how the current issues can be resolved.
Girish Malhotra, president,
EPCOT International,
Pepper Pike, Ohio
Closely Examine The Scrubber
Consider the following points:
1. The scrubber working for a short time (“few weeks after addition of new packing”) suggests several possibilities. So, review the trend charts from the plant historian (database). For example:
• Check the feed rates and compositions of gas (e.g., CO2 content from the CO2 membranes) and caustic — then and now. Look for the possibility of high flows and possible flooding in the packing. Try adjusting flow rates to match the design.
• Is caustic flowrate adequate? Too low a flowrate could cause poor distribution in the packing and would lead to poor vapor/liquid contact.
• Did any abrupt increases in vapor flow occur? Such an increase could have disturbed the packing or support tray, leading to potential maldistribution of gas flow (and caustic flow as well) through the packing.
• Have caustic deposits plugged any sections of the packing?
2. To assess if the new CO2 membranes are working, check flows and compositions of the incoming and exit streams. If data show the replaced membranes never worked, the problem could stem from, for example, improper installation or use of the wrong type of membranes.
3. If operating data show the scrubber never worked at design conditions, then there may be a flaw in sizing or packing design — e.g., packing hydraulics, the height equivalent of a theoretical plate (i.e., mass transfer considerations), improper choice of type or size of packing, poor installation, inadequate packing or wrong support or hold-down trays.
4. For the long term, consider improving instrumentation and controls.
5. Plate-and-frame exchangers are prone to fouling/plugging in services that contain particulates. Because there are no taps for pressure gages or thermowells, consider using “skin” temperatures (i.e., pipe wall temperatures) to get some crude indication of fouling.
Maybe the project was implemented hastily and without buy-in from key stakeholders. Along with economic/market justification, you need to get general agreement to implement a project. Time invested upfront will help encourage team work and sense of participation by all. This is project diplomacy.
GC Shah, senior advisor
Wood Group, Houston
Analyze The Changes
The best way to look at this problem is via change analysis: what changed besides the packing? Turnarounds at refineries are complex affairs. Make a list of everything upstream of the scrubber that was altered. Look for different compositions of streams, upstream and downstream. Check temperatures and pressures, too.
Next, look at how the carbonate could form from H2S and caustic. Talk to research and do some of your own. Find out if the refinery has seen this before. If so, do a timeline, change-analysis and another stream review. Assess commonalities between this and past occurrences. Include a discussion with sales, production and maintenance.
While you’re looking at histories, find out why a sieve tray tower was chosen in the first place. How often were the trays fouled and by what?
Plate-and-frame exchangers have pluses and minuses. They are compact, which may have led to the choice. They also are easy to clean. A spiral heat exchanger might have been a better option if fouling was anticipated. Even a shell-and-tube exchanger with the fouling fluid put in the tubing, as it should be, will foul. Disassembly of shell-and-tube exchangers is more complicated.
Clearly, the scrubber packing depths were chosen based on adiabatic cooling of the gas stream. I’ve used this with HCl and the bed can shrink by at least four times. You may want to improve your scrubber design. Just because a scrubber worked for a few weeks doesn’t mean all is well.
In a pinch, I’ve installed high efficiency packing in place of more robust packing. Also, consider putting in a mesh pad scrubber at the vapor discharge. Change your caustic feed to add the caustic to the mesh pad on top of the scrubber. In addition, dilute your caustic if that doesn’t raise problems downstream. This will improve the mesh pad wash and perhaps delay the inevitable fouling.
The more immediate concern is the CO2. Look at the membranes upstream to capture CO2. Separating sodium carbonate from sodium bisulfide downstream is difficult. Carbonate acts as a buffer in water solutions. Typically, 3% NaS and 4% carbonate are maximums for quality NaHS. High pH is favored for good product, which may mean splitting your caustic makeup between different stages. You may have to go back to the drawing board on your scrubber design.
I suggest actual pilot testing not simulations because the former often reveal problems pure mathematics won’t show. A common challenge in scrubber pilot testing is unavailability of industrial packing. Packing makes all the difference. Liquid and, often ignored, gas distribution also are important.
Dirk Willard, consultant
Wooster, Ohio
April’s Puzzler
We want to expand our existing storage capacity for biphenyl by adding another, larger tank. We now use a shell-and-tube exchanger to keep the biphenyl in a molten state in the two current tanks. We maintain the temperature between 169°F and 205°F. Properties at the two temperatures are: density, 0.99 g/cm3 and 0.97 g/cm3, respectively; viscosity 1.42 cP and 1.03 cP; heat capacity, 0.382 BTU/lb-°F and 0.397 BTU/lb-°F; and thermal conductivity, 0.077 BTU/hr-ft-°F and 0.076 BTU/hr-ft-°F. The biphenyl’s flash point is 237°F. We use 30-psig steam with a 30°F superheat to heat the biphenyl and don’t allow tank temperature to get within 30°F of the flash point because of past incidents.
Our existing tanks are 28-ft high and 32-ft in diameter with 4 in. of fiberglass insulation except at the top and bottom. The new tank is 32-ft high and 42-ft in diameter with 6 in. of insulation, including on the top. Typically, we run the tanks at about 75% level after a production run. The winter design temperature is 0°F with a 10-mph wind.
Our shell-and-tube exchanger doesn’t exactly operate within specs: the flow rate is about 140 gal/min instead of 180 gal/min, and the heat transfer coefficient U is only about 56 BTU/hr-ft2. We derived this information from trends and calculations; we lack other data on the exchanger.
I’m concerned about whether the exchanger can cope with a third tank. In the winter, the heater already runs about 17 hr/d to heat the two existing tanks. Moreover, the piping is over 300-ft back and forth to the new tank — and it’s overhead.
Corporate managers want to rely on the current heat exchanger and put in a new pump and piping (at $22,000 for the new pump and $348/linear ft for piping — including engineering). I instead suggested installing a dedicated heat exchanger for the tank and using a pump we can borrow from another plant. The new heat exchanger would cost $23,000 and require only about 150 feet of new pipe. How do I sell my idea to corporate? Am I missing some reason why they want to only use the existing heat exchanger? Can you suggest any way to improve the system?
Send us your comments, suggestions or solutions for this question by March 15, 2019. We’ll include as many of them as possible in the April 2019 issue and all on ChemicalProcessing.com. Send visuals — a sketch is fine. E-mail us at [email protected] or mail to Process Puzzler, Chemical Processing, 1501 E. Woodfield Rd., Suite 400N, Schaumburg, IL 60173. 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.