Two Lead and Lag Reciprocating compressors control 1 stage each

[guevara74]

Hi,
I am not so experienced in this field and want to share some thoughts regarding some standard skids of 2 reciprocating compressors that are suppose to work as lead and lag: First, the lead is getting ramped up till getting 100% load and if the inlet pressure still raises then the second (lag) compressor is starting and keeping ramping till getting a stable pressure on suction. If the Lag goes under 50% load when pressure getting stabilized, them the lead is ramping down so both compressors are reaching the same load, let say around 60-70%, that actually fits the economic regime of the gas driven engine and energy saving. Attached is the sketch of the skid/system. From I&C point of view I am wandering if such a skid can be controlled by a single control panel, single PLC based, that also will be able to handle some of the utilities feedbacks. Compressors speed (actually capacity control) is controlled based on suction pressure PT by SC.
Thanks,
Alex
The Drawing

 

[georgeverghese]

Should be possible to run both units from one panel, but of course, the panel should receive feedback of actual engine speed from both machines.i.e.
PT to PIC to master SIC to master load signal splitter then to individual engine SIC.
Feedback signal from each engine ST to individual SIC
Feedback from each engine ST to master load signal splitter also.
Master load signal splitter to contain these additional logics you've described - sounds workable.

 

[guevara74]

Thanks George,
I have been about in the same wave of having the control done. Actually was thinking to have the PIC as master for both SIC, but more dedicated approach will just make more reliable the control as you have proposed. Have attached some sketches. On the bottom side are the 2 versions of controls in two different layout. The feedback though can be picked by master via bus from individual controllers of speed. On the uppers side of the sketch there are 2 options of the PLC architecture. 1st with one single PLC along with a Panel View (of course) and the 2nd with 3 PLCs with a Panel View or Working Station as well. I think the 1st option could be considered moneywise at the most convenient than the option 2. Taking in counts that for option 2 all 3 PLC shall be SIL rated.
Thanks for your input and time.
Alex

 

[EmmanuelTop]

If the compressors are located a bit away from each other and one of them being much further down the suction header, it is recommended to consider separate pressure transmitters located at the suction of each compressor - otherwise one of the two units will always be operating at higher load. The load should be split so that the pressure loss in each suction line is equal. This applies if the two compressors are similar/identical.

In case the two units are different, you need to work out the load-efficiency curves and see what is the best load sharing concept for this particular application.
Your sketch does not show everything, so I guess you have covered other important subjects, i.e.: how the lead/lag operation is configured (set points for pressure switches), overspeed protection etc.

Get a copy of Bela Liptak Instrument Engineers' Handbook, Volume 2, Paragraph 8.15 "Compressor control and optimization". It contains some great deal of information.

Dejan IVANOVIC
Process Engineer, MSChE

 

[guevara74]

Yes, both compressors are of the same configuration and are located on the same base plate/skid. The distances from suctions of each compressor up to the pipe split is equal. The main incoming stream of the gas shall be constant in flow with some small deviations from time to time. When the system is set to run it will start first the Lead compressor ramping up. The feed will be provided in steps by adding volume form different souses connected to the inlet manifold. Upon reaching the max speed limit and still the suction pressure will increase, the second - lag compressor will start ramping up till the suction pressure will not get stabilized at the desire SP. Based on preliminary calcs, the load-efficiency curve for both shall be obtained at a certain load per compressor for the forecasted feed. So, for efficiency the Lead compressor will be set for getting slower and the lag to get up till their speed will not match at the load-efficient speed (close to), keeping the suction pressure still stable at the desired SP. In case if the load for Lead compressor will be less than 10% of them max speed limit and the pressure gets stabilized, then the Lag will not ran. In case of turn down the sequence will go vice-versa, ramping down first the Lag and then the Lead, keeping suction pressure within the positive SPs. In case of the compressor skid emergency trip or malfunction, the inlet stream will be redirected via an ESD valve to the by-pass line and will just avoid compression and the skid isolated and vented to the vent via BDVs.
Thanks,
Alex

 

[georgeverghese]

Alex,

Error in my previous post on first line

PT to suction PIC to master SIC to master load signal splitter then to individual engine SIC.

Also, you also need a low suction pressure controller for the case when both compressors are at min speed and suction pressure continues to drop - this should open the overall recycle valve to keep suction pressure at the low suction PC setpoint.

Further, there should be a discharge PIC which overrides the compressor load controls through an auto sginal select which selects either the suction PC output signal or the discharge PC output signal. So revised first line will be

PT to suction PIC to signal select to master load signal splitter then to individual engine SIC.
Signal select to autoselect between suction PIC or discharge PIC.

Hope you know how to set up these 2 PICs in order to enable the auto select

When discharge PIC is running the compressors, and compressors are at min speed by signal from this PIC, then the discharge pressure continues to rise, then the recycle valve should be opened. Hence the recycle PCV receives its command signal from either low suction pressure PIC or high discharge pressure PIC via an auto signal select.

Also this means the discharge PIC has to be a split range controller - one range to operate the load signal for the compressors, and the other range for operating the recycle PCV.

Also, a high suction PIC may be required to cut back the source feeds automatically for the case when source supply flow is higher than permitted compressor capacity at that time.

Process controls never fail to bring up suprises, so keep working on this scheme and watch out for complications during programming due to misinterpretation and/or poor definition of the process control narratives or code language limitations.

 

[georgeverghese]

Sorry, just had a look at your second diagram:

a) The master SIC shown on your sketch is the master load signal splitter in my posting
b) Dont have working knowledge on actual wiring of these signals - whether to use serial links or FFB etc. etc - this is the working responsibility of the plant control systems instrumentation engineer - there may be others in this forum who can help.

As Dejan says, add on for other controls functions and keep building up a controls and safeguarding sketch that gives you a birds eye view of everything you need to know from a controls / safeguarding perpective.

 

[guevara74]

the situation is a bit different than the usual one. Basically the suction is providing a stable pressure of 15barg as a discharge of an another compressor train which is self acting and out of this scope. Now, the skid (both compressors working at max load) can handle about 140% of the provided debit from upstream train. the oversize was done based on increasing in production for next years for about 10-25% forecasts. Now, the discharge of our skid will be a common delivery manifold to a open gas pipeline. It can handle whatever amount of gas only if the pressure is above 40barg. It will not allow to have the pressure increase above this figure. Our skid is about 12 km away from the pipeline delivery point. In case, but very rare situation if the discharge line gets blocked somehow, the skid will trip via SIS and the ESD will occur with all inventory blowing to vent after the inlet and outlet of the skid getting isolated by ESVs. The routine of this event will be handled properly and the compressor vendor will implement the logic accordingly. So my concern and reference would be the suction pressure as I have shown in the sketches. Yes, as I have mentioned above I was thinking to get the PIC and control the speeds of each compressor by him. But than is not too suggestive to have the Pressure dictating the speed in special as controller block, so an intermediate Speed controller as a spliter and master speed controller will handle the SIC of each compressor. Of course, my sketches are just schematics in executing the control, the main narrative and functions will come at the later project progress execution and will be the clear support for the compressor vendor to do the logic and control. Also, will be selector switch available to make selection which is Lead and which is Lag and some other local facilities to help on achieving better and optimized control.

 

[georgeverghese]

So we have another compression train that provides feed to this unit.
To trip this compressor skid when pressure reaches 40barg at pipeline inlet may be one way to run this unit. Another way would be to use a discharge PIC which could be used to reduce the speed of the upstream compression unit. Then we wont need to trip this new compression train / existing compression unit. If this is of interest, you could use this discharge PIC low and high range signals to operate the speed and recycle valve of the upstream compressor.

 

[guevara74]

Rare, but very Rare the downstream will get blocked so the new compressor skid to have discharge raising quick high up. It is open discharge to a pipeline. But, in case, if, somehow, the pipeline gets blocked by an misoperating of something, then the shot down in a domino sequences will occur to upstream feed up to the well shut-down. In case the new compressor skid is getting off, then the inlet will switch the stream, by a SDV, to a bypass directly from discharge of the old upstream compressor station directly to the pipeline. Bypassing the new compressor skid. Discharge pressure monitoring is there, in place.