Compressor process control

[rotw]

All,

Is it okay to design a moto compressor centrifugal type handling clean natural gas without any control method provided (No by-pass, NO IGV nor throttling device, NO speed variation). The decision to not implement any control method is dictated by process.
Whatever can be the arguments to justify such approach, I beleive this is not a sound practice because there is no flexiblity.

In your opinion what can be the implications to go this way ?

Thank you

 

[zdas04]

The biggest reason not to do what you describe is that things change. Even in the most stable of operations, sooner or later something upstream or downstream will change (maybe just for a short time) and require an adjustment in compressor performance. Assuming that that will never happen is simply a mistake of epic proportions.

David Simpson, PE
MuleShoe Engineering

"Belief" is the acceptance of an hypotheses in the absence of data.
"Prejudice" is having an opinion not supported by the preponderance of the data.
"Knowledge" is only found through the accumulation and analysis of data.
The plural of anecdote is not "data"

 

[rotw]

Zdas,

I agree. The application is pure downstream, does it change something?

The argument I received is that downstream compressor somewhere in the process there is a control valve that could adjust the system curve. I have never faced such situation before for compressors that's why I am a bit puzzled.

In the upstream, nobody can predict the behavior of a reservoir and we might end up with huge variations, especially at the compressor inlet, MW changes and are never as expected. In the downstream, the process variables are generally better under control, but still I believe process upset may occur: fouling can happen in equipment, equipment performance might be altered etc.

In your opinion, what can go wrong if the design I have mentioned will be finally adopted.
Could you please illustrate with some example or scenario to figure out ?

Thanks

 

[zdas04]

What can go wrong? EVERYTHING. A large end user has an unscheduled outage and stops taking gas and capacity has to drop 20% within 15 minutes without raising pressure upstream of the compressor suction controller. An upstream facility loses a train and your suction volume drops 30% over a half hour. Suction controllers/recycle controllers can do a lot, but dang they are wasteful of energy.

David Simpson, PE
MuleShoe Engineering

"Belief" is the acceptance of an hypotheses in the absence of data.
"Prejudice" is having an opinion not supported by the preponderance of the data.
"Knowledge" is only found through the accumulation and analysis of data.
The plural of anecdote is not "data"

 

[JJPellin]

There are a number of possible bad outcomes. My first concern would be that the flow would drop and the compressor would go into surge. Unless it was monitored for vibration and armed to trip automatically, the compressor could run to destruction in a few minutes. Another risk would be if the differential pressure dropped too low and the compressor ran to stonewall (to the right on the curve). Thrust compensating mechanisms in the compressor rely on differential pressure to function properly. Running with low DP could result in high thrust with possible thrust bearing failure. Once again, unless the monitoring system will detect a thrust bearing failure and trip off automatically, the compressor could be destroyed. A catastrophic failure of the compressor could result in a loss of containment. This could result in fire, explosion, personnel injury or death.

Johnny Pellin

 

[rotw]

Process assume that the operation is very stable (as I already indicated this is for a downstream application). It is argued that the duty consist only to cover one operating point with fixed flow and Delta P.

Are you aware of any reference (code, engineering practice, book) where I can find a table that links between SIL level of the application and the minimum safeguard protection required to be implemented? I know some compressors manufacturers have minimum requirements on Discharge Temperature, Discharge pressure and Vibrations such as Voting rule for trip (2oo3) to match a given SIL level (as a minimum for OEM). If something is said on controllers it would be good.

In addition, with regard to the point of JJPellin related to automatic trip. I suppose that if the compressor is shutdown on emergency trip without damage or failure it does not mean that the rest of the process is not at risk. I deem this may (or may not) lead to unexpected behavior somewhere else in the plant. Which incidentally would lead to a risk of loss of containment elsewhere despite the compressor has been tripped without failure or damage. Is such assumption correct ?

 

[LittleInch]

Is it "okay"? Possibly providing there are no isolation valves and no significant changes in flow or pressure ratio and the flow cannot be isolated or controlled to behind the surge line or the "stonewall" line of the compressor.

Is it a good idea - definetly not. One key qestion to ask "process" is how do they intend to start and stop the whole system?? Many times people seem to think a system just magically runs at steady state without working out how to start and stop this thing, even if is a continuous process it needs to start at least once in tis lifetime and probably more during commissioning.

You may not use the bypass line very often, but if there is any possibility of a no flow situation occuring, no matter how remote or needing manual intervation (an isolating valve for instance) then your compresor will only last a very short time before suffering some major damage. Often these scenarios are not properly considered until the final HAZOP when you then need to go and add a reycle line into the design at the last minute.

My motto: Learn something new every day

Also: There's usually a good reason why everyone does it that way

 

[JJPellin]

You are correct regarding automatic compressor trips. There are processes where the trip of machine could result in a run-away reaction, unit de-pressure or other extremely dangerous outcome. We do not want to cause a larger risk than the one we are trying to mitigate. This is why we do Process Hazard Analyses on these systems to understand all of the interactions and implications.

But, if a compressor is running to destruction, it is going to shut down anyway. I can trip it off and bring it down safely. Or, I can let it wreck and wait for the driver to trip off on overload after the rotating parts start to weld themselves to the stationary parts. In some scenarios it comes down to a question of timing. Do you want to trip sooner or trip later?


Johnny Pellin

 

[rotw]

JJPellin,

Your point is clear. I would not design anything not in accordance to API617 and therefore stick to the trip setting on high vibrations aligned with API. By the way if the trip is initiated on high vibrations or on axial thrust, the trip sequence will be automatic. How much can you do to influence the timing then, no idea. Otherwise are we talking about operator action that will override the automatic sequence ?

LittleInch,

I noted (on the other parallel thread) that in some cases (reformer) with no closed valve in the loop, anti-surge protection via by-pass line can be omitted. Consider the type of gas being compressed. Hydrogen gas leads to very gentle, flat and wide performance curves. Something going wrong in a context where the turndown to surge is typically as high as 40% may allow you to buy some time. It is not like having steep curves with high molecular weight. With narrow curves, especially if continuous operation were already on a point close to surge at 100% speed, chances exist that the compressor will surge during trip, even a by pass protection may not suffice. So I can hardly imagine the situation where, on top of that, nothing has been envisaged to safeguard the machine. Not sure it will get better because the application involves in particular a reformer, or not?

Regards

 

[dcasto]

In a closed loop system I could see a none controlled compressor EXCEPT, hold it, Start up. How do you bootstrap up a none controlled compressor?

 

[BigInch]

Oh man! Kick the tires and light the fires ... on the Express Train to H2ell.

Independent events are seldomly independent.