This Month’s Puzzler
We would like to increase the cooling water flow to a tower condenser (heat exchanger A in Figure 1) from 2,900 gpm to the exchanger nameplate rating of 4,500 gpm. A review of the files shows the heat exchangers and control valves have plenty of capacity. In fact, we are concerned that so many of our heat exchangers are oversized. We’re also looking at ways to reduce the recirculation flow and pressure losses caused by the balancing orifices — we may be able to re-use the pump if we can eliminate some of this flow and the pressure losses caused by the balancing orifices. Do you think we can increase heat exchanger A to its nominal flow rate? What will this entail? Will this affect the best operating point of the pump? Do we really need the balancing orifices?
Take A Broader View
The idea of increasing the flow for one exchanger seems flawed: 1) if raising the flow to one exchanger is part of a tower expansion, other nearby towers will be affected, boosting cooling water flow for these condensers; and 2) these clearly are nameplate pressure drops — scaling increases the drops and reduces the cooling water rates to exchangers.
I modeled the process from the drawing (see Figure 2). Although the butterfly valve on HX-A will accommodate a large increase in flow to that exchanger, it could starve the other exchangers on the leg from the header. I could increase the flow to HX-A up to 4,500 gpm without affecting HX-B and HX-C and most of the other exchangers in the network by adjusting the restrictive orifices (ROs). Normally, I’d suggest moving the tie-in for HX-A to location closer to the header but this would disrupt other heat exchangers in the network. Also, notice that the discharge head for the pump jumped from about 93 psig to 97 psig. You may need to modify your pump to achieve 4,500 gpm.
If you can live with less than 4,500 gpm to HX-A, you will need to reduce the recirculation. Decrease the bore-hole size for RO-1 and RO-2, and adjust RO-3 to bring down pump head, to steal from the recirculation flow. If you don’t, the pump head drifts up with the higher flow rate to HX-A — this would require another stage to the pump or a larger impeller to handle the greater flow with the resulting higher system pressure drop. Fortunately, there is plenty of net positive suction head available for a larger pump.
A downside to stealing water from recirculation is variances from steady state. How often will the pump operate away from its best efficiency point (BEP) when the higher flow to HX-A is not necessary? This would change the required recirculation flow from steady state. An alternative to ROs might be back-pressure regulator valves. You adjust them according to when you have a wide-open valve at HX-F1 (the valve with the highest % open). As flow to HX-A goes up, you want more pressure loss at the back-pressure valves — the % open should decrease. This must be done carefully because some downstream exchanger control valves may run at choked flow for some pressure settings. Fortunately, this is a chronic problem, not an acute one. Still, valve damage will occur over time.
You also may want to reduce the size of the valves on HX-H through HX-M. These valves are in the 20–40% open range.
So, to answer your question: no — you can’t get rid of the ROs unless you replace them with valves. If you do remove or see wear in these orifices, you will experience starved valves, cavitation, choked valves and wear in pump basins and impellers.
If you must increase the rates for other nearby exchangers to HX-A and you still can maintain flows to other exchangers in the network, try the following: 1) add stages to the pump — but check the motor hp; or 2) install a booster pump on the leg feeding HX-A through HX-F1,2. Whatever the change, adjust the ROs or backpressure valves to keep the pump at its BEP and reduce the recirculation rate to minimize energy costs. Increasing the size of the control valves, including the one for starved HX-F1,2, won’t have that much of an effect; in addition, butterfly valves and ball valves are not the best control valves — globe valves are superior. In the long run, replace your oversized heat exchangers with ones within 10% of your nominal operating rate; make some allowance for startups and shutdowns.
One last thought: Don’t do projects involving multi-million dollar processes without a complete material balance and hydraulic study. Collect data on valves and pressure drops for a period of two years or, better, five years. Get the full story.
Dirk Willard, consultant
We manufacture 97%-pure ethylnylcyclopropane (ECP, also known as cyclopropyl acetate — CAS No. 6746-94-7). The batch ECP process involves these primary steps: reaction with an alkyl lithium compound; blowing in NH4Cl with high purity N2 to form the ECP; pumping the product solution through a bank of filters; quenching the alkyl lithium with cold heptane and 2-propanol; washing with chilled deionized (DI) water, followed by filtering; decanting in a bed filled with random packing to remove the salts, separating the organic and aqueous phases; and, finally, distilling the organic phase to separate the ECP (distillate) from the 2-propanol. The wastewater, which still contains a trace organic phase, undergoes batch vacuum distillation to remove the trace. The wastewater then passes through our new trickle-bed air biofilter before discharge to the city sewer water plant.
Unfortunately, the new manager made some changes to reduce costs: 1) using air instead of N2; 2) replacing deionized water with plant (well) water; 3) bypassing the chilled water cooler for the quench; and 4) cutting back the 2-propanol to almost zero. He was following the advice of an experienced operator. The first campaign didn’t go well.
We had organic carryover from the biofilter into the city sewer; the stream also contains n-hexyl lithium. In addition, our product quality suffered; we only achieved about 88% purity after distillation. What did we do wrong? And can we get rid of the 2-propanol without substituting something worse?
Send us your comments, suggestions or solutions for this question by Sept. 11, 2015. We’ll include as many of them as possible in the October 2015 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, 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.