Ethylene Plant Turboexpander-Recompressor Problem

[sshep]

I have posted this troubleshooting problem about a year ago without much response. As result of a recent unscheduled hurricane related outage we were able to warm-up the system, make inspections, and restart. Unfortunately the problem still remains. This problem will be a bit lengthly but easy to understand I think.

This is how it is supposed to work: Our 3 stage expander system lets down demethanizer distillate vapors (H2,CH4) to produce a cold liquid CH4 stream (after vaporized we call this MOG) and a hydrogen rich off-gas stream (HOG). The refrigeration value of these streams is used for condensing demethanizer reflux. On the other side of the expander shafts, recompressors raise the pressure of the HOG above fuel gas pressure to provide motive force for using the HOG for drier regeneration. The pressure profile through the expander-recompressors is anchored by the demethanizer pressure (controlling offgas to the expanders) and a back pressure valve on the last recompressor. All of the HOG must be recompressed and the last expander outlet pressure (and recompressor inlet) simply float on the result.

This is the problem: Our expander speeds and efficiency are low (22000rpm vs 28000rpm, isentropic effi 70%ish vs 86% design); consequently our MOG make is 1/3 of design, our HOG make is high and purity low, our demethanizer is losing ethylene to MOG due to lack of cooling, our recompressor spillbacks are closed due to the high HOG flow.

This is my question: Because all of the HOG must be recompressed, is there a feasible path start-up issue with the system as designed? There is no way to unload the compressors to account for low expander efficiency. When the expander efficiency is low the amount of HOG (and average MW) which must be recompressed is significantly higher. The work balance constraint across the shafts seems like it could result in an off-design stable operating point. I think that our system differs from other systems in that we have no means by which to reduce the load on the expander shaft and get our speed up to design.

Any help is appreciated.

best wishes, sshep

 

[Zoobie]

Who is the manufacturer of your expander-compressor? We have 30 yr old Mafi-Trench unit and I have found there support to be absolutely excellent. They can run off-design curves for you to help figure stuff out and I have had a lot of luck with just being able to call them up to assist with some troubleshooting. I have learned more about expanders talking to them for 15 minutes than from any textbook or operating manual. Of course if your's is of different make you may not have the same success.

Something that you might not want to hear but if you are way off your design point you may need new wheel(s).

 

[sshep]

Hey Zoobie,

Thanks. Our equipment is made by GE Rotoflow. I am pretty sure that we have a design feed within the envelope I specified in a 2000 redesign, but Rotoflow is sending me their field data sheet to complete and return for analysis.

Incidently in this 2000 project I convinced our managment that we should include at least a tie-in point at the inlet to the first compressor to allow unloading the compressors via a drag of HOG off to the flare or fuel gas. I asked for a pipe to do this because I believed our old system sufficiently matched design conditions and should work. I feared our redesign would result in exactly the same situation again, which is exactly where we are now.

When Rotoflow analyzed my "drag" case, a drag of only 2000pph would change the force balance to significantly raise the speed of the expanders. A 5000pph drag was into a significant overspeed that would have allowed use of our compressor spillbacks- instead we are compressing about 5000pph extra gas (relative to design) and our speed is consequently slow. Unfortunately, even now I cannot get agreement to run the needed pipe based on my "theory". I would like to cite at least one other instance of a start-up path problem or corroberating opinion to support my idea, or otherwise get advice to solve this problem.

best wishes as always,
sshep

 

[Zoobie]

I understand the problem better now but I can't really offer much in the way of help. Your 'theory' though does seem to make sense from a physics point of view. I would think that if your losses are adding up do to poor efficiency that the piping job, even if it didn't work, would be a worthwhile try for what it would cost compared to other solutions.

I'm actually looking right now at the opposite problem. We are doing some work here that will see us actually not having enough compressor mass flow to prevent overspeeding the unit. My first idea was to pipe in gas from our demethanizer overheads to the suction of the compressor to slow down the machine but I haven't followed through with the whole exercise. (my application is a bit different as it is for a deep-cut ethane extraction plant

Where I always get stuck (and perhaps where you may have some resistance from colleagues) is that when your expander and compressor are also coupled process-wise quite intimately you get some catch-22 phenomena. Its like you're damned if you do and you're damned if you don't.

I won't be looking closely at our unit until the spring but if in the meantime I get some sort of brainwave regarding manipulating 'drag' or 'braking' I'll be sure to pass it on.