16 replies to this topic

[Root]

Hi,

We are facing compressor failure reaptedly, on studying the P&ID i found compressor suction scrubber is with out demister pad and liquid carryover is there. this facility is before 1970. I think if we install demister pad problem can be solved but then suction scrubber need to be PWHT and my company doesn't have PWHT equipments.

please see attached sketch and let me know further on this investigation and mitigation steps to stop carry overyover.

thanks

Toor

Attached Files

•  Comp KOD.pdf   44.54KB   140 downloads

[ankur2061]

Toor,

 

Besides not having a demister pad, there is no inlet device (half-pipe, vane, Schoepentoeter) in your KOD, the presence of which which will ensure momentum values are not very high in order to facilitate de-entrainment of liquid from gas and prevent carryover to the gas outlet.

 

An option would be to replace your KOD by a high efficiency filter coalescer which will ensure negligible liquid carryover. Refer the link below:

 

http://www.pecofacet...eparator-a4.pdf

 

 

Regards,

Ankur.

[fallah]

Toor,

 

It might the demister pad hadn't been put there to keep the suction pressure at a proper level, because the pressure at incoming line to KOD isn't so high (1 psig). Indeed, at the present situation it might the problem of liquid carry over would be solved by lowering the HH liquid level inside the KOD...

[Root]

Ankur & Fallah

Thanks for your quick reply, original design is 15 psig for suction scrubber but  gas is coming from crude oil surg and vent tank as i posted in tank venting forum.

the problem is  big because of this facility is installed  before 1970s and may be on that time process designer  does not consider potential of liquid carryover. any how i have not seen suction scrubber without demister pad.

Thanks

Toor

[breizh]

Toor ,

Consider theses references to support your work.

 

Hope this helps

Breizh

[Zauberberg]

If there is enough plot space, you could consider installing a vertical standpipe downstream of the KO Drum, equipped with level control and high level trip, to catch the liquid carried over from the KO Drum. I would try to do some filed testing to determine the range of liquid flows/volumes of liquid phase from the KO Drum - just to avoid installing something which is too small. I have seen this arrangement in two plants, working as intended.

[PingPong]

on studying the P&ID i found compressor suction scrubber is with out demister

A P&ID is not always in line with reality.

 

Check the drawing of the vessel to know for sure what was - or was not - originally installed in the vessel.

Talk also to people who entered the vessel during maintenance to hear whether there are internals present in the vessel.

 

Without knowledge of the gas flowrate and density, and the liquid level in the vessel it is not possible to estimate whether excessive liquid entrainment is indeed a possibility.

 

Do not get fixated on one possible cause of compressor problems. The unit already operated for decades, so why would now, all of a sudden, liquid entrainment be a problem?

[Root]

Hi,

Thanks for all valuable members for actively replying, LP compressor is designed for 2.4 MMSCFD and gas density is .68 but i have doubt on present value for gas density because this gas passing through surge and wash tank. moreover in vendor drawing no demister pad installed but for physical examination, I have discussed with maintenance guy and per him it is hallow vessel

Thanks

Toor

[fallah]

Toor,

 

As PingPong mentioned, at first it should be inspected if the KOD really hasn't equipped with a demister pad; but if it really hasn't and liquid entrainment is a concern, there might be two possible situations:

 

1- The KOD had already been equipped with demister pad and recently being removed due to, let say, plugging or...

2- The KOD hadn't been equipped with demister pad but the feed to KOD has recently been changed in composition respect to what had been used in design stage and liquid entrainment is a new concern due to such composition change...

[Root]

Fallah,

as i wrote in my posts, suction scrubber came without demister pad and no piping has yet ever changed, Brize attached one blutin for demister pad this looks sound and other zauberberg suggestion also looks sound by installing stand pipe with demister and complete with instrumentation.

Thanks

Toor

[PingPong]

LP compressor is designed for 2.4 MMSCFD and gas density is .68

With that fairly low flowrate you would need a demister with an area of only 0.25 m² (2.7 ft²).

The velocity in the 16 inch pipes is only 7 m/s (24 ft/s).

 

Without a demister you would need a cross sectional flow area of 0.5 m² (5.4 ft²) to minimise liquid entrainment.

Cross sectional area of the empty vessel is 1.4 m² (15 ft²) so even if the vessel would be half filled with liquid there should not be an entrainment problem. Gas velocity in the vessel would then be only 1.2 m/s (4 ft/s).

 

The vessel and piping seem generously sized for a gas flowrate of 2.4 MMSCFD.

 

As it worked for decades without a problem you should focus on what changed just before the recent compressor problem started.

[Root]

PingPong,

It is a cronic problem not new,any how thanks for your contribution in this thread.

thanks

Toor

[shan]

Install a strainer just upstream of the compressor inlet nozzle.

[Bobby Strain]

Maybe you are producing asphaltenes with the gas. These would quickly foul compressor valves. You didn't mention that it is a reciprocating compressor. So you might search for a solution with this possibility.

 

Bobby

[Root]

I'm sorry for this missing information, it is centrifugal compressor and compressor 14000 rpm speed.

Thanks

Toor

[chemsac2]

Toor,

 

Many of the experts on forum have already dwelt on aspects of getting rid of entrainment problem, I will introduce you to one excellent article on quantifying separator performance I recently came across. 

 

This is how I would proceed for this to estimate entrainment from compressor suction KOD to compressor:

1. Estimate liquid formed in inlet line to compressor KOD (probably due to heat loss/pressure drop) and entrained from upstream system. First part is easy and can be done with any simulation software using its pipe utility. Second part is difficult but few articles are available e.g. HTRI predicts entrainment rate in kettle reboilers.
2. If this is compressor KOD application, I would expect liquid in the form of mist.
3. Estimate droplet size distribution. Correlations for different two-phase flow regime are available which give mean droplet diameter. Droplet size distribution is considered Gaussian distribution and upper and lower limit of droplet diameter can be estimated from Gaussian distribution with corresponding volume fraction of liquid for each diameter.
4. Since rho*v2 is small, droplet size shift is not expected to be great and feed droplet distribution is expected to continue except that bulk liquid would separate if it forms at all.
5. From gravity separation equation, estimate which droplets would be separated and which would not.
6. From mass of liquid and droplet distribution, mass of liquid associated with entrained droplets can be estimated

This is essentially CFD, but a simplified procedure is given in articles by Mark Bothamley in his articles in "Oil and gas facilities". He has written 3 articles (August, October and December, 2013). Link below has these:

 

http://www.spe.org/ogf/print/archives/

 

Hope it helps.

 

Regards,

 

Sachin

[PingPong]

On request I explain below how I calculated a required demister area of 0.25 m²:

 

It simply follows the well known Souders-brown equation.

 

Allowable gas velocity through a demister is: v_G = K*√((ρ_L - ρ_G)/ρ_G)

 

in which K = 0.1 m/s or 0.35 ft/s.

 

It was specified that gas density is 0.68 which I interpreted to be the gas specific gravity.

So the gas MW is then: 0.68 * 29 = 19.7

 

Combined with a pressure of 1.08 bar abs (1 psig) and a temperature of (say) 40 ^oC this gives a ρ_G of 0.82 kg/m³

 

Assuming a liquid density of 850 kg/m³ this then gives: v_G  = 3.2 m/s.

 

Gas flow = 2.4 MMSCFD = 120 kmol/h = 2360 kg/h = 2880 m³/h = 0.80 m³/s

 

So the required free demister area (excluding support rings) is then: 0.80 / 3.2 = 0.25 m²