Problem in crack gas compressor start up

[Mohammad Ka]

Hi everyone
During start up of an ethylene plant, discharge temperature of the crack gas compressor (fifth stage) was drastically increased, meanwhile very high amount noises like fast opening and closing of a valve was observed. Moreover the compressor faced with insufficient flow. In this case, opening of the antisurge valve was 100%.It should be noted that there is no FE on antisurge loop Of CGC stage 1 and no piping pockets in the suction. However, a pumping trap is installed in upstream of suction line CGC stage 1. Antisurge valve with a difuser is specified to be a low noise
Other trends could be listed as below:
1- Outlet temperature of crack gas from quench tower was increased about 4-5 deg.C.
2-Level of discharge and suction drums of the crack gas compressor (stage 4) were increased during starup.
3-Discharge temperature of 5th stage of the CG compressor was increased to 130 C.
4- Fluctuation of flow was observed in all five stages of the compressor.
5-Alongside the flow fluctuation some small fluctuinations of pressure drop was observed in caustic tower.
6- No vibration in CG compressor was detected by DCS.

The following activities to remedy have been done:
* Antisurge valve was opened and fully inspected . No failure was detected.
* All suction and discharge drums, check valves, orifices, cooling water exchangers and caustic tower in the line were inspected and no chocking or unusual thing was seen. 
* About 1000kg mixed yellow oily water was observed in cooling water exchanger of 5th stage of the crack gas compressor. Antisurge valve is branched from outlet of this exchanger. The water was completely drained. 
*Stage 5 of the compressor was inspected by a borescope camera and no failure was detected.

After these remediations and start the plant for the next three times, the noise, insufficient flow and increased temperature were again repeated.
I was wondering if anyone could propose a solution? 

 

[rotw]

As far as I remember, 130 deg C is borderline as for polymerization temperature of cracked gas.
Is the train equipped with some washing system (e.g. Naphta injection nozzles)?
I could speculate the 4th stage may be causing the problem to the stages forward.
Other question?
Which stage anti-surge valve is opening ?
If there is no FE on anti-surge loop, how does the control system controls the valve?

Is this steam turbine driven as I suspect? is the gas composition sampled and molecular weight within specs. ?




Life is not about waiting for the storm to pass. It's about learning dance in the rain.

 

[rotw]

Also, are the impellers Teflon coated?

Also have a look here:

https://www.researchgate.net/public...racked_Gas_Compressor_Case_Studies_on_Fouling

 

[Mohammad Ka]

thank you rotw
when the compressor was started and passed the min. governor speed to reach the normal speed, the discharge pressure and temperature of stage 5 increased while discharge flow did not rise and began to fluctuate. Reaching to about 130 C at discharge, the speed was reduced and the train was stopped.
A wash oil injection system is provided in the CGC but it is normally used in normal operation and not in start up.
Anti surge valve of stage five is fully opened and other antisurge valve located between stage one and four is about 60% opening.
The first and second antisurge valves are controlled by discharge temperature and pressure of stage four and five respectively. However FEs are installed on discharge line of CGC both stages four and five.
The turbine as you mentioned is steam driven.
Normally during each start up the gas composition is not sampled and the outlet temperature of furnace is the only criteria to be in spec, however it is analysed via the lab and the results are in range including gas composition and MW.

 

[rotw]

First impression, is that problem on stage 5 is probably an induced problem and some stage(s) upstream stage 5 is at fault. But it's hard to say, the least we can say is that it is not easy to diagnose machinery issues let alone such a complicated beast. It gets even more complicated if there dynamic effects on top (start-up).

Some speculations:
- Stage 1 to 4 have a common anti-surge valve, this means the 'weakest' stage from 1 to 4 is going to trigger recycle for the complete arrangement 1-4. Having a combined anti-surge valve decreases capital costs but squeezes operating range. Cracked gas compressors have typically modest operating range depending on the gas heaviness. So, when the stage at fault has recovered from surge it has done so at the expense of other stages operating off design and with possible efficiency drop because each stage is not equipped with a dedicated valve.

- If anti-surge valve opens, there may be a (dynamic) effect of drop of forward flow to the successive stage (five) which has triggered the 2nd anti-surge valve to fully open, this has positioned the operating point of that stage onto the left of the speed curve (i.e. control line flow) which corresponds to higher delivery pressure and higher discharge temperature.

Is this a start-up for new unit? If so, maybe you could check if the design is referenced especially in terms of anti-surge valve arrangement? Is the process open art or licensed and check if the anti-surge valve and control system manufacturer vendors have also references. If the set up is referenced for this capacity or equivalent, fine, if not you could start to wonder what has been done.

- Is there some sort of boiler feed water injection to the impellers in order reduce the discharge temperature (cooling) in order to prevent polymerization temperature?


Life is not about waiting for the storm to pass. It's about learning dance in the rain.

 

[Mohammad Ka]

Thank you so much rotw
The train has been started for many times and this problem has not occurred up to now.
BFW is injected to impellers in stages 1-4 and wash oil is injected intermitently. Recently wash oil and BFW was injected in stage 5 (before the last stoppage).It seems the impellers are not coated.
It should be noted that a caustic tower is located between stage 4 and 5 so the discharge of stage 4 is directed to the tower and the CG from overhead of the tower is sent to suction of stage 5.
One more thing, a pressure valve (PV) is located in discharge of stage 5 to flare and when the valve is limited the discharge temperature and noise is drastically increased.

 

[rotw]

If there is failure of after cooler of stage 5 then temperature of gas going to anti-surge valve will increase. There might also be obstruction of flow passage on heat exchanger (fouled or altered after BFW injection and naphta oil washing?). The hot gas goes back to suction of stage 5 and while it is mixed with main feed flow it has drastically increased temperature and pressure at the suction and incidentally at discharge of stage 5. This has triggered the anti-surge valve of this stage to fully open. Possibly this has also induced effect on stage 1 to 4 which started also to recirculate facing higher system resistance from stage 5.

Highly speculative in any case.

Life is not about waiting for the storm to pass. It's about learning dance in the rain.

 

[Mohammad Ka]

The after cooler of stage 5 is NEN type exchanger which CG passes through tubes and cw in shell. The exchanger was inspected and no plugging was observed in tube side. When wash oil was used during one of unsuccessful start up, about 1 ton wash oil and polymer was observed in the outlet and tube side of the aftercooler. The mixture was completely drained and the train was started. The noise, heightened temp. and insufficient flow were repeated.
What about the plugging of the stage 5 of CGC ?

 

[georgeverghese]

You have a parallel post on this topic - we can continue with this one to keep correspondence together. E-Tips policy discourages double posting on any one topic.
High discharge temp on stage 5 tells me maybe you've got low pressure on stage 5 suction. Is stage 5S pressure within normal range? - Any blockage on the demister in the caustic tower?
Am trying to understand compressor antisurge scheme:
One ASV on Stage 4D returns to stage 1S and is operated by either stage 4D high temp or high pressure
One ASV on stage 5D returns to stage 5S (downstream of caustic tower and upstream of stage 5 suction drum, I presume) and is operated by also by high temp or high pressure on stage 5D.
Stage 4D and 5D flow elements have no input to antisurge control.

Presume there is no input from turbine speed to the antisurge algorithms, so the surge line is most likely assumed conservatively to be the one at max permissible operating speed.
Is there a capacity control recycle valve ? What happens when you have low pressure on stage 1S, especially during startup when cracked gas flow is low? A capacity recycle valve would then recycle on stage 1S low pressure also.
Another reason may be foaming in the caustic tower due to some kind of contamination of the recirculating caustic solution. This may be the reason for high level in stage 5S drum.

 

[Mohammad Ka]

Dear rotw
Trends of discharge and suction temperature of stage 5 show that when the discharge temperature increases, the suction temperature decreases or almost constant. So, the failure of after cooler stage 5 is debunked.

 

[rotw]

What about the plugging of the stage 5 of CGC ?

Did you look at historian / data trends to understand the behavior vs. time?

Another question is, why has the oil washing and BFW injection been applied to stages 1 to 4th only and not to stage 5th?
What was OEM/operator instruction in this regard and have the prescriptions (if any) been followed in the past?



Life is not about waiting for the storm to pass. It's about learning dance in the rain.

 

[Mohammad Ka]

georgeverghese
Thank you for your hint.
At discharge of both stages 4 and 5 the flow is cooled through cooling water after coolers to about 45 C. So, the flow in antisurge lines are cooled enough and have lower pressure than discharge line. FEs are installed on discharge of stage 4 and 5, however there is no FE on antisurge line. The opening of first and second antisurge valve is adjusted based on the suction temperature and pressure of stage 1 and 5, respectively.
I supposed there is no capacity recycle valve in the line.  
When the discharge pressure stage 5 is increased, the discharge temperature and suction pressure of stage 5 are rised. The suction pressure of stage 1 is fluctuate, first increased next reduced then rised.
In case of low pressure on 1S the antisurge valve will open.
The level in suction drum stage 5 was remained constant and the demister of caustic tower was inspected, it was clean.

 

[Mohammad Ka]

rotw
The available trends show that there is no such significant deviation in performance of CGC.
Injecting wash oil and BFW in stage 5 may increase the level of discharge drum stage which may increase the possibility of liquid carryover and send much more liquid to the dryer. Recently discharge temperature of stage 5 has been rised approximately 4-5 C however the common discharge temperature was deviated from design value about 10 C. So it was decided to start the injection of BFW and wash oil in stage 5. The vendor was also recommended to inject BFW in all stages of CGC.

 

[rotw]

Has the operator taken the decision/responsibility to deviate from vendor recommendation that BFW injection and naphta oil injection shall be applied to all stages? I suggest to maybe check this point; also because you did mention the impellers do not have special coating.

Normally after washing stage #5, the performance should go back to normal; surge line for instance should match the parameters of the configuration of origin in the controller (assuming none has manipulated the anti-surge controller configuration).

However if the impellers have experienced irreversible alteration (as an induced effect to extensive fouling), means a degradation that would probably not 'disappear' upon washing, then the process stage map could shift (degradation case) and incidentally the anti-surge controller would no more be configured adequately. So, boroscopic inspections indicates no failure, Ok, but did they get to the impeller surface? I am myself ignorant on boroscopic investigation, so just wondering.

I also would like to say that 4-5 deg C deviation is not negligible (all other things being equal).

Life is not about waiting for the storm to pass. It's about learning dance in the rain.

 

[georgeverghese]

Again, your antisurge scheme is not clear. You now say
"In case of low pressure on 1S the antisurge valve will open."
Previously you said the ASV on stage 4 and 5 only works on high pressure or high temp of stage 4D or 5D respectively.
I think we need to have a complete picture on this control scheme and also a complete PFS or sketch.
The plant I worked on shows a large low point piping pocket in the feed to the intermediate caustic tower, which is steam traced and insulated, presumably to vaporise condensate that would have collected upon CGC shutdown /cooldown. Do you have such an arrangement in your plant, and is the condensate cleared out of this low point in the caustic feedline before CGC startup? - liquid slugging in this line may be the cause of these pressure fluctuations?

 

[Mohammad Ka]

rotw
It should be added to previous note that when the discharge temperature of stage 5 was increased they decided to inject BFW as per vendor recommendation.
By the way, the CGC stages 4 and 5 are opened and no fouling or degradation are seen. So the plugging of CGC stage 4 and 5 are rejected.
As far as I know, no body has manipulated the anti-surge controller configuration from the previous normal operation up to now.

 

[Mohammad Ka]

georgeverghese
A schematic of compression stages is attached.
The ASV controller uses suction temperature and pressure, discharge flow, temperature and pressure to adjust the valve opening. During start-up before reaching to minimum speed the ASV is fully opened. Above the minimum speed if the compressor was not in the surge,in case of low pressure on 1S the antisurge valve will open.
During the previous start up I checked the trends and I found that in stage 5:
Increasing the CGC speed >>> Increasing Discharge Pressure & Temperature >>> Increasing Suction Pressure & Temperature >>> Increasing the level of suction drums stage 1-4 and discharge drum stage 4 (no change in level suction drum stage 5) >>> DP of caustic tower was fluctuated (increased and decreased)
It worth noting that the level of discharge drum stage 4 (inlet to caustic tower) was as high as 98% during the first startup but in the next trials the level was controlled.
There is no low point from the discharge drum stage 4 to the caustic tower and the feed to the tower is preheated in a steam jacket via LP steam.

 

[georgeverghese]

There are some details missing on this sketch, but we can guess where/ what it is:
a) Speed control of turbine is presumed to be taken from a PIC on stage 1S
b) Stage 5D line to flare PIC is presumably upstream of propane chiller
c) Stage 5 antisurge line is shown to be taken from downstream of stage 5D cooler / upstream of stage 5D drum - but it should be from downstream of stage 5D drum.
The sketch does not show the inputs to the antisurge loops, so we will go by your description so far.

The plant drawings I have here also have this pumping trap on stage 1S and there is no CGC stage 1S drum. It also shows this LP steam jacket heating on caustic tower feed to be located on a piping low point. These drawings I have are for Assaluyeh naphtha cracker based ethylene plant.

Am running out of guesses, but a few more will be:
1)Any instability seen in the line to flare PCV on stage 5D?
2)Compressor casing drains all working well ? Where are these routed to?
3)Heat tracing on piping downstream of ASV working okay
4)Is there an input from steam turbine speed transmitter to stage 4 and 5 antisurge loops, and is there any irregularity in these feed signals from the turbine ST?
5)You said stage 4 antisurge loop also recycles gas to stage 1 in case of low pressure on stage 1S, which may happen when turbine speed is already decreased to min speed. Does the same concept apply to antisurge loop on stage 5 i.e. does it recycle gas to stage 5S on detection of low pressure on stage 5S?

 

[Mohammad Ka]

About the missed details:
a) Speed of turbine is controlled from a PIC on stage 1S.
b) Stage 5D line to flare PIC is upstream of stage 5D after cooler.
c) Stage 5 antisurge line is exactly from downstream of stage 5D after cooler / upstream of stage 5D drum and then go to inlet of suction drum stage 5.

This plant is ethane cracker based ethylene plant and the feed line to caustic tower has no piping low point. A sketch of isometric drawing from discharge drum stage 4 to caustic tower is attached.

About the questions:
1) The 14" line from discharge stage 5 to flare has no PCV and the flow is controlled via a hand gate valve. However, no instability could be seen in the gate valve, the trends from DCS show the instability and fluctuation of discharge flow of stage 5, discharge P stage 5, suction P stage 5, DP caustic tower. Moreover, field operators reported frequent closing and opening of the check valve at the discharge of caustic tower during the start-up.
2) Compressor casing drains are routed to ditch( ground) and all working well.
3) If you mean steam jacket by heat tracing on piping downstream of ASV, yes it is working well. In other case if you mean heat tracing in discharge stage 5, there is no heat tracing in this stage.
4) No, there is no input signal from steam turbine to stage 4 and 5 antisurge loops. The input signals to both antisurge controllers are suction T&P and discharge P & DP (or flow).
5) In case of low pressure on stage 1S (less than 0.2 bar) the stage 4 antisurge loop recycle gas to suction to prevent ESD activation but the same concept does NOT apply to the second antisurge loop i.e. stage 5 antisurge. However, the turbine speed did not decrease to min speed during the start-ups.

"The surge line in this compressor is defined by X-axis : sqrt (DP * Pdischarge)/Psuction and Y-sxis: Pdischarge/Psuction
The anti-surge control uses the signals of the differential pressure transducer in the discharge line of the compressor , the suction pressure
transducer and the discharge pressure transducer. The anti-surge control valve is opened pneumatically, which means that it is necessary to convert the
electrical controller output into a pneumatic signal. This function takes place in the current to a pneumatic (I/P) converter and acts directly to the pneumatic
positioner mounted at the control valve. The solenoid valve mounted at the anti surge valve has to ensure a quick opening of the anti surge valve.Quick opening occurs if a potential free contact from the compressor PLC to the solenoid valves fails for any reasons. This contact is normally closed as the compressor has reach minimum speed."

 

[rotw]

Typically, train is started with stage 1 and 2 in re-circulation. Turbine ramps up. Probably there is idle speed sequence for turbine as well.
Turbine reaches minimum operating speed. Discharge Check valves should be open. Anti-surge valves gradually close. Turbine speed is ramps up gradually. At normal operating speed, anti-surge valves are normally closed.

It seems at this point, discharge pressure and temperature of stage #5 continued to increase, flow reduced. Anti-surge valve no.2 went fully open. Anti-surge valve no.1 opened at 60%. Which one occurred first is not clear.

This is just guess. So, could you post a brief / more correct summary of the starting sequence of CGC if it is minimum effort; I cannot say that it will really help solve the issue. It would just be interesting to investigate the start up sequence in details and maybe get a hint of which point exactly things went wrong.


Life is not about waiting for the storm to pass. It's about learning dance in the rain.