4 replies to this topic

[Jorge_RZ]

Hi,

 

I'm working on a project for a self-refrigerated LNG plant. If you're not familiar with it, please see attached a simplified process diagram.

 

Due to cost considerations, the cycle is provided with necessary shut-off valves (SDV-1, SDV-2) to isolate two distinct pressure sections during an emergency shutdown. One section reaches a settle-out pressure of 9 barg (designed for at least 12 barg) and the other reaches a settle-out pressure of 125 barg (designed for at least 140 barg), according to my calculations.

 

Both compressors K-1A/B and K-2A/B are two-stage reciprocating compressors.

 

The suction drum (V-1) of the main gas compressor falls on the high-pressure section (HP). It is designed for 140 barg and the relief valve protecting it (PSV-1) is set at 140 barg.

 

Looking at different relief scenarios (fire case is excluded by client), one realizes that is not possible for the system to reach 140 barg (other that at settle-out conditions) because there's no pressure source that high (natural gas feed available at 36 barg, and recycle compressor has a discharge relief valve, PSV-2, set at 40 barg).

 

Therefore, the only way PSV-1 would ever open is during a shutdown if the settle-out pressure is higher than expected.

 

What would you say should be the sizing basis for PSV-1 in this case? It's not obvious to me.

 

To give you an idea of the size of the equipment, the total gas in the HP section when it reaches settle-out conditions is 35 kg @ 125 barg. A typical ASME relief valve with a C orifice (0.068 in2) at this pressure has a relieving capacity of approx 3,862 kg/h of gas. I consider it is plenty to protect this drum but I'm not sure about it because I don't know how fast the system could overpressurize.

 

Hope I explained myself correctly.

 

Thanks,

 

 

 

 

Attached Files

•  SKM_C30818080210480.pdf   118.71KB   95 downloads

[Bobby Strain]

Only 35 kg? There is not enough information on your sketch. One needs a P&ID to give advice.

 

Bobby

[Jorge_RZ]

Thanks for your reply Bobby,

 

Yes, it's a small-scale LNG liquefaction unit for use in gas fields.

 

Unfortunately, this is proprietary material and I cannot share P&IDs. What kind of information do you find missing that could have an impact on the sizing basis of this valve? I'll be happy to provide it if possible.

 

Regards,

[Art Montemayor]

Jorge:

 

I spent a lot of my career working with, around, and applying gas compressors.  I also spent some time in the LNG industry.  Your post interests me because I’ve already looked at small gas field-produced LNG production in the past and I can relate to what you are facing.

 

I can appreciate that you have a proprietary design and can’t reveal many things about it, but I think I know what is involved and some of the challenges you are up against.  Firstly, let me ask some definition of what you call “settle-out” pressure conditions.  I’ve never had the use of this term when dealing with reciprocating compressors.  It is a term that has been applied, in my experience, only to dynamic machines - centrifugal compressors that are devoid of valves and are shut down with eventual equalization of the discharge conditions with the suction conditions.  In the case of a reciprocating machine this can’t happen because of the positive displacement features of the machine.  So I’m having trouble visualizing what is meant by your use of the term.  (Your sketch shows the conventional symbol for a centrifugal compressor - not a reciprocating one - so that threw me at first)

 

All the reciprocating compressors I’ve processed designed, installed and operated always started up or shut down in the unloaded condition - or at least they were designed to do so.  I was never able to justify starting up a recip in the loaded state - especially out in the oil patch.  We inevitably blew the compressors down or into the flare system to unload them.

 

But even if you are able to keep the compressors’ system totally pressurized with natural gas during startup and shut down, you still have other potential problems you have to resolve in your sketched process.  You are essentially using an external refrigerant (your description of this being a self-refrigerated process is not correct) to pre-chill your compressed natural gas to get it down to a temperature level where the Joule-Thomson expansion will be such that you can create sufficient post-expansion LNG and return the associated cold expansion vapors back to recompression.  Since your feed gas pressure is entering at a relatively high suction pressure into your main, 2-stage compressor, you are forced to employ a 2-stage booster to get the return vapors up to the main compressor suction conditions.  This presents several control and pressure relief problems.

 

Your compressor cylinder sizing is going to be tricky and a challenge in controlling the capacities of each stage.  Your initial start up will always demand a large capacity due to the fact that your initial gas feed into your J-T expansion valve is not going to be cool enough to give you the ideal flow of return vapors back to recompression.  This flow rate will initially be inflated until the system reaches some sort of process equilibrium with respect to temperatures.  You may go into total recycle, but that means you will need to have what you don’t show: a blocked feed gas inlet with total recycle of gas slowly being reduced until the feed gas can be regulated into the main compressor as needed to maintain the design conditions.  Your compressor controls as well as instrumentation will be taxed.  The temperature, as well as the flow rates, in the return vapor stream will be differentially changing with time.  This is a tough specification on a compressor - especially a recip.

In my opinion, you cannot justify or rely on a relief valve (PSV2) to operate as a flow or pressure regulator on your booster compressor discharge, flowing back to the booster suction.  How do you regulate the capacity of the main compressor while you are trying to regulate the capacity of the booster?  This type of “control” or pressure relief system would never fly in any of my Hazops.  I also know it would not fly in any of the major oil companies’ projects that I have worked on or led.

 

In the meantime, while you have blocked off the return expansion vapor to the suction of the main compressor, how is the same main compressor supposed to continue taking feed gas and compressing it with its discharge also blocked off?  --- and it doesn’t show any pressure relief valve on either of the two discharge stages.

 

As Bobby Strain has indicated, it is very difficult - if not impossible - to comment and render advice when there is no detailed P&ID.  To try to do so would leave you with the impression that any well-intentioned comments could be safely applied - in spite of the fact that not all required basic data was available and any advice could lead to a tragic mistake or error.

I see a lot of more potential concerns, but I won’t proceed because they are based on my past experience with this type of process and they may not apply due to the scarcity of details.  However, I think I’ve expressed my concerns on this process and what you propose with respect to pressure relief scenarios.

[Jorge_RZ]

Art,

 

I was wondering where I got the term "settle-out" pressure for reciprocating compressors. Now I remember. API Standard 618 (Reciprocating Compressors for Petroleum, Chemical, and Gas Industry Services) actually uses the term "settling-out" pressure when it discusses required motor torque at start-up conditions.

 

I imagine it refers to cases like mine where you have a passage from high to low pressure side (by-pass valve in fail-open positions, or other recycles in the process). I still agree with you that backflow through compressor head is not possible if functioning properly.

 

Regards,

 

Jorge