Pressure Drop across the Mist Eliminator Vessel with Internal Cyclone Bundles.

[Sruthish]

Hi All, We are supplying Mist Eliminator vessels to our Client with Cyclone bundle internals. The Flow parameters are given below.

Flow Rates
Max: 1471 m3/hr
Normal: 1421 m3/hr
Min: 1371 m3/hr

Operating Pressures
Max: 332 PSI(g)
Normal: 323 PSI(g)
Min: 313 PSI(g)

Operating Temperature: 248°F (120°C)
Liquid and particulate Removal efficiency : minimum requirement of 8 micron at 99.95% efficiency
Maximum pressure drop : 2 psi.

The Pressure vessel is having a Internal diameter of 660 mm and Shell length of 2,730 mm. Could anyone suggest equations to calculate the pressure drop across this pressure vessel. The inlet and exit nozzle sizes are 10" ( DN 250) x Sch 80S.

We shall calculate the pressure drop at the entrance of the inlet nozzle, across the cyclone bundle and at the exit of the outlet nozzle. Apart from these is there any area where the pressure is getting dropped? Please advise if the composition of the Air will impact on the pressure drop.
IF we are getting higher pressure drop as per the actual operation across the vessel ( around 3 Psi), what could be the reasons leading to higher pressure drop than the designed values ?
Let me know if require any more details regarding this issue.

 

[pierreick]

Hi,
Are you sure about your calculation?
To me:
h= K*V^2/2g > Delta P= Ro*g*h > Delta P= Ro*K*V^2/2.
Delta P1/Delta P2= Ro1*Q1^2/(Ro2*Q2^2)
Using the perfect gas law, Ro1/Ro2 = P1*T2/(T1*P2)
Delta P1/Delta P2 = P1*Q1^2*T2/(P2*Q2^2*T1)

Q1 = 1425 m3/h , P1 = 335.7 PSIA , T1= 393.15 K
Q2 = 1750 m3/h , P2 = 203.7 PSIA , T2 = 415.15 K
Delta P1 / delta P2 = 1.1474
At Customer site the delta P should be lower! less than 2 PSI
Pierre

 

[georgeverghese]

A few points to work on :
a) The process datasheet for this separator should have info on the wet air flowrate at approx 330psia and contribution from water vapor. Using the relative humidity value that you derive from this info, you can calculate an estimated value for water vapor content of the gas at 205psia and 140degC (assuming relative humidity remains constant - I would imagine RH would be >80% or so).

b) differential pressure should be measured using a differential pressure gauge ( dPG) or transmitter (dPT); not by PT_upstream - PT_downstream. Also, differential pressure instrument should have instrumentation impulse lines as top tap from process line, and impulse line should slope down to process line (ie diff press instrument should be at high point) in this wet gas service. What is actually installed there ?

c)You say FT is on compressor discharge. Is the operating discharge pressure from compressor the same as that used for sizing this FE ? What type of FE ? Is pressure and or temp compensation / correction done for this readout ?

d) Have assumed in these calcs I did, that the pressure drop across the swirltube deck, (which is the major contributor to separator total dp), is also, like the feed and exit nozzles, a linear function of Rho_V2 only. (ie dP_swirldeck = Ks. rho_V2). There is no contribution here in this assumption from residual liquid entrainment in the vapor/mist stream from the entrance nozzle beneath). This should be confirmed by swirltube deck supplier / designer. Suspect my assumption is not correct.

e)What was the justification for this 2psi dp allocated to this separator when operating pressure is approx 300psig ? Even though this is feeding a compressor, I would imagine a dp of up to 5psi for this multicyclone type separator would have been acceptable, as impact on compressor power would be negligible at max 5psi dp. Also, why is the operating pressure at this separator and compressor suction dropped from 300psig to 200psig? - is this a reciprocating compressor?

@1503/@pierre, To avoid making mistakes with all these corrections and double dipping, I used another approach the second time : Characterise this "composite resistance coeff K" for this separator as a length of a DN300 pipe (you can use any other length, any other dia as long as velocity remains low) to get this design dp of 2psi at 1470m3/hr at min op 330psia - adjust pipe length till dp matches. Once characterised, then change P and T and flow to suit any other operating condition to get new dp. Then do a new characterisation for the actual dp of 2.8psi at actual operating condition/ flow, and then work back what dp will be for design case operation. If you have a line sizing excel or mathcad sheet, these case runs shouldnt be too difficult.

 

[pierreick]

Hi,
To add to the discussion, information about compressor capacity and effects of altitude, % RH
@ George,
I did the math following your proposal,
DN 350, Q1=1421 m3/h and Q2 =1750 m3/h, P1 = 335.7 PSIA and P2 = 203.7 PSIA, T1= 393.15 K and T2= 415.15 K Delta P1 = 2 PSI and Delta P2 = 2.8 PSI
rugosity= 0.15 mm , f based on Swamee equation f1 # 0.016656 and f2= 0.01686, turbulent regime, incompressible model
L1eq = 1677 m and L2eq = 2660 m
Pierre

 

[georgeverghese]

@pierre,
Okay, dp1 matches, but dp2 seems different. But in any case, dp2 is much more than spec limit of 2psi.

 

[pierreick]

Hi,
To get 2psi max, the flow rate @ customer's site should be less or equal to 1474 m3/h.
Pierre

 

[georgeverghese]

"There is a cooler upstream of this mist eliminator vessel, which they shows as S&T HEx in some P&IDs and some other projects shows as "Radiator type"."

Further to item (d) in my post on 7th June, if this is a HX type cooler, then liquid water feed to this separator vessel will not be the same at 200psia / 140degC operating pressure as compared to 330psia / 120degC. This will most likely make some difference to the liquid water related pressure drop component at the swirldeck, and complicate things further re performance guarantee on pressure drop. Process engineers should be able to calculate what the liquid water feed rate change will be at this separator for this lower pressure operation.

So far, you havent posted the complete process datasheet for this vessel / PFDs' for this entire section of plant.

Given these complications, also ask these plant operators to run this entire unit at 330psia to check for performance guarantee dp value.

 

[Sruthish]

Hi George and Pierre.
Thanks for the supports and also appreciating continued information/suggestions. We will ask customer as you suggested as well. We have limitation of providing the confidential information such as process datasheets/ P&ID/ PFDs. I could explain the matters further if you could provide me a contact number or email ID. Or please drop me an email at sruthish.p.s@gmail.com

 

[Sruthish]

@George, I would like to reply to your message on 7th June as separately as below.

a) Client hasn't provided any such firm info. We have asked earlier also this info to them and still awaiting the response.
b) Impulse lines are taped from centers of the process nozzles and DPIT ( we are using electrical DP transmitter) is situated above the process nozzles.
c)Yes that was the value which Client provided to us before. But now the pressure is not the same. And still we are investigating how much pressure is input to the Mist eliminator vessel on the upstream end. I understand by FE you mean Flow element. kindly confirm. We are checking more info to Client on this anyway.
d) This is something we have to check with our Bundle supplier. But could you please explain what are the chances that this entrained liquid contributing to the pressure drop. Condensate and solid particles supposed to be collected at the bottom chamber below the swirltube deck.
e) We stick to 2 psid as it was required by Client's specification. We are requesting them to consider this as 3 psid as we also thinking it could have more allowable pressure drop without affecting the performance of the systems.
It seems operating pressure is getting adjusted based on the feedback from the overall system. Yet to confirm and type of compressor also yet to check and confirm with Client.

Thanks for the supports.

 

[Sruthish]

Hi Pierre,

Thanks for the support. We are also having similar feelings from our calculations and analyzing the trend given by the client.
"To get 2psi max, the flow rate @ customer's site should be less or equal to 1474 m3/h". Also the operating pressure should be more closer or same as what we have designer for.
We are investigating how the client is measuring the flow rate inletting to our mist eliminator vessel and the accuracy of that data. Also, why it has been changed from the designed operating parameters.

Once again thanks for the supports.

 

[georgeverghese]

Re your response to (d)
The feed nozzle internals (half open pipe or some kind of turning vane assembly) is not a perfect vapor - liquid separation device. Some of the liquid in the feed will be entrained in the vapor stream going up to the swirldeck. The more the liquid in the feed, the more liquid will be entrained up into swirldeck. Liquid entrainment fraction will also increase as feed gas velocity increases. Pressure drop at swirldeck will increase with increasing liquid fraction in the feed from upstream feed nozzle internal - this major contributor to swirldeck dp has NOT been taken into account in the estimations made.

 

[Sruthish]

Hi @George verghese. Thanks for the information. Do you have any papers/methods/calcs available to calculate this pressure drop due to liquid entrainment?

 

[georgeverghese]

Documentation I have has a method for calculating the gas phase dp at the swirl tubes, which is directly related to rho_v2 of the gas phase - taken into account in these estimates in this thread. No method given for entrained liquid phase dp - it only states a range.
Pls note that liquid entrainment into the vapor stream arriving at the swirl deck is dependent on type of feed nozzle internal also. A primitive half open pipe will entrain far more than a specialty turning vane type assembly, such as Shell's proprietary schoepentoeter device.