liquid butane flow control 1

[rika kose]

We have a liquid butane tank 6 barg pressurized by N2.
We want to have constant flow injecting into a reactor, e.g. 0.05 kg/s. At this moment, there is only one flow control valve in the injecting line. Butane flashes in the valve. downstream the valve, icing. flow is very unstable. a mass flow meter is near the storage tank, around 30 m before the control valve.

What is the right setup? how to improve the system?

 

[LittleInch]

At what pressure are you injecting / d/s the control valve.

What temperature is your butane tank?

If the butane is flashing then you will have cold vapour which can cause issues such as icing.

Only way around that is to heat the butane before you flash it or heat it straight after the valve, but before is better so long as the nitrogen pressure prevents it flashing.. 6 bar butane you can get to prob 60C without it flashing upstream the valve. You would need to do some flash calculations or simulations to see what temperature the vapour would be. Or have a higher pressure d/s the control valve.

 

[georgeverghese]

Delete N2 pressurisation. Set up a heating coil in the holding vessel to heat up this liquid butane and control - maintain vapor space pressure at 6barg. Draw off vapor leading to injection control valve, not liquid.

 

[shvet]

1/ i-butane or n-butane?
2/ Is pressure at the tank constant?
3/ Is pressure at the injection point variable?
4/ Is flowrate variable?
5/ Are total liquid conditions at CV inlet nozzle guaranteed?

 

[rika kose]

At what pressure are you injecting / d/s the control valve.

What temperature is your butane tank?

If the butane is flashing then you will have cold vapour which can cause issues such as icing.

Only way around that is to heat the butane before you flash it or heat it straight after the valve, but before is better so long as the nitrogen pressure prevents it flashing.. 6 bar butane you can get to prob 60C without it flashing upstream the valve. You would need to do some flash calculations or simulations to see what temperature the vapour would be. Or have a higher pressure d/s the control valve.

Thanks for the reply.

pressure across the CV : upstream : 5.6 barg (due to the piping insulation problem, CV upstream pressure changes somehow with the ambient condition), downstream 0.2 barg
Temperature in the storage: winter time: ~5ºC, summer, ~20ºC. Tanks has no insulation.

Butane flashing in the control valve is very unfavorite. flow is not stable.

What would be the disadvantages if we place a RO or something downstream the control valve (to prevent flashing).

 

[rika kose]

Delete N2 pressurisation. Set up a heating coil in the holding vessel to heat up this liquid butane and control - maintain vapor space pressure at 6barg. Draw off vapor leading to injection control valve, not liquid.

Thanks,

but modifying the storage tank is really a big chance, no one wants to do it.

 

[rika kose]

1/ i-butane or n-butane?
2/ Is pressure at the tank constant?
3/ Is pressure at the injection point variable?
4/ Is flowrate variable?
5/ Are total liquid conditions at CV inlet nozzle guaranteed?

1. it is n butane.
2.Pressure in the storage tank is quite stable.
3.Injection point is the reactor. around 1 m below the liquid level. so it is stable.
4. flow should be constant.
5. before the CV, it is liquid. but because of insufficient isolation, the P-T varies with ambient temperature.

 

[shvet]

flashing CV is indesired but stiil a normal option for the process industry as those are sensitive to a control loop adjustments and actuall Cv rangebility
make sure that existing CV inlet conditions are total liquid accounting the flow acceleration due to CV's DN is smaller than upstream piping

add a pressure gauge and hand operated CV downstream main CV so way the main CV's outlet pressure would be >2barg
plus
talk to a reputable experienced manufacturer to replace the main CV with another one dedicated for cavitation/flashing into the seat
plus
test with an experinced I&C engineer adjusting the responce time and other parameters of the flow control loop or try to control the flowrate manually

this way is cheaper and faster than modifying the process design

 

[pierreick]

Hi,
Consider Pulsafeeder pump with associated flow meter to transfer your butane to reactor. This was the way we were supplying Butadiene to a polymerization reactor where the injection had to be stable and overshoot not acceptable.

Home

Trust Pulsafeeder, a global leader in fluid handling technology. Pumps, controls, systems for industry. Safe and reliable. Precise control.

pulsafeeder.com

Pierre

 

[LittleInch]

Sounds to me like you have the wrong control valve. What is it and what does its data sheet say?

0.2 barg n butane has a temperature of about -20 from what I can find out. Add in the latent heat of evaporaiton into hat and you've got pretty cold vapour.

But that should affect the control of flow. What percent open is your control valve at during this flow? Should be between 30 to 70%.
What sort of "unstable" flaw are you talking about? 5% or 50% of rated flow?

 

[rika kose]

flashing CV is indesired but stiil a normal option for the process industry as those are sensitive to a control loop adjustments and actuall Cv rangebility
make sure that existing CV inlet conditions are total liquid accounting the flow acceleration due to CV's DN is smaller than upstream piping

add a pressure gauge and hand operated CV downstream main CV so way the main CV's outlet pressure would be >2barg
plus
talk to a reputable experienced manufacturer to replace the main CV with another one dedicated for cavitation/flashing into the seat
plus
test with an experinced I&C engineer adjusting the responce time and other parameters of the flow control loop or try to control the flowrate manually

this way is cheaper and faster than modifying the process design

Thanks!

 

[rika kose]

Sounds to me like you have the wrong control valve. What is it and what does its data sheet say?

0.2 barg n butane has a temperature of about -20 from what I can find out. Add in the latent heat of evaporaiton into hat and you've got pretty cold vapour.

But that should affect the control of flow. What percent open is your control valve at during this flow? Should be between 30 to 70%.
What sort of "unstable" flaw are you talking about? 5% or 50% of rated flow?

Yes, very cold vapor, full of ice outside the pipeworks after the CV.
flow set point is 0,03kg/s. flowmeter shows: 0.005kg/s to 0.5 kg/s. jumps every second.
The CV is not properly sized, working range is below 10%. no position sensor on the CV. the exact opening is unknown.

 

[TiCl4]

There might be a few issues with the following solution, but this might be your best "cheap" try at fixing this:

1. Put a restriction orifice in the line directly on the reactor inlet nozzle. Size this such that the majority of your pressure drop occurs through the orifice, not in the CV. N-butane has a vapor pressure of ~2 bar at 20 C, so you'd want to keep pressure downstream of the CV around 3 barg or so. Check CV trim to ensure liquid flow can be controlled with lower dP. Check reactor inlet nozzle and reactor constituents to ensure the very cold liquid jet of flashing butane that is going into the reactor doesn't pose icing issues or other issues from the injection style.
2. If injection of cold butane directly into the reactor poses a problem, move the restriction orifice to just downstream of the control valve and look to see if you can heat trace/insulate the line downstream of the R.O. to warm the butane back up after it flashes. I'm not sure why flashing into the reactor would be an issue given you already send very cold vapor into it (you mention your piping is icing), but I thought I'd mention that. Going with option #1 would also relieve you of needing to install heat tracing.

My $0.02.

 

[LittleInch]

But before that look at your system, the control valve and what sort of control you have in your PID loop controlling the valve.

It does sound very much to me like your control valve is too big and your controller can't exercise the control required at very low opening percent.

Please provide some details of the control valve. That's a very small flow so you need an incredibly small contol valve.

 

[rika kose]

I am thinking to replace the CV with a manual CV (needle valve). as we need only the constant flow.
the upstream pressure is relativly stable, although it changes slightly with ambient temperature.
Adjust the manual CV to the desired flowrate in the startup.

 

[rika kose]

There might be a few issues with the following solution, but this might be your best "cheap" try at fixing this:

1. Put a restriction orifice in the line directly on the reactor inlet nozzle. Size this such that the majority of your pressure drop occurs through the orifice, not in the CV. N-butane has a vapor pressure of ~2 bar at 20 C, so you'd want to keep pressure downstream of the CV around 3 barg or so. Check CV trim to ensure liquid flow can be controlled with lower dP. Check reactor inlet nozzle and reactor constituents to ensure the very cold liquid jet of flashing butane that is going into the reactor doesn't pose icing issues or other issues from the injection style.
2. If injection of cold butane directly into the reactor poses a problem, move the restriction orifice to just downstream of the control valve and look to see if you can heat trace/insulate the line downstream of the R.O. to warm the butane back up after it flashes. I'm not sure why flashing into the reactor would be an issue given you already send very cold vapor into it (you mention your piping is icing), but I thought I'd mention that. Going with option #1 would also relieve you of needing to install heat tracing.

My $0.02.

Thanks!
Cold gas going into the reaction liquid gives no problem.

 

[TiCl4]

But before that look at your system, the control valve and what sort of control you have in your PID loop controlling the valve.

It does sound very much to me like your control valve is too big and your controller can't exercise the control required at very low opening percent.

Please provide some details of the control valve. That's a very small flow so you need an incredibly small contol valve.

Good point. What is the normal operating % of the control valve, and what is the CV value at 100% open?

My rough math on this: 0.05 kg/s -> 1.3 gpm (s.g. of liq butane at 0.6). dP = 85 psig, maybe take 5 psig away from line losses = 80 psig. Q = Cv * (dP/s.g.)^0.5, estimated Cv = 0.11

I would expect the valve to be a 1/4" or 1/2" nominal size. Larger valves with smaller trims can be installed, of course, but 1/4" valves can get up to 0.3 Cv max or so, and 1/2" tend to be somewhere near 2-3 Cv max.

 

[rika kose]

But before that look at your system, the control valve and what sort of control you have in your PID loop controlling the valve.

It does sound very much to me like your control valve is too big and your controller can't exercise the control required at very low opening percent.

Please provide some details of the control valve. That's a very small flow so you need an incredibly small contol valve.

Good point. What is the normal operating % of the control valve, and what is the CV value at 100% open?

My rough math on this: 0.05 kg/s -> 1.3 gpm (s.g. of liq butane at 0.6). dP = 85 psig, maybe take 5 psig away from line losses = 80 psig. Q = Cv * (dP/s.g.)^0.5, estimated Cv = 0.11

I would expect the valve to be a 1/4" or 1/2" nominal size. Larger valves with smaller trims can be installed, of course, but 1/4" valves can get up to 0.3 Cv max or so, and 1/2" tend to be somewhere near 2-3 Cv max.

control valve DN15.
I don't have a proper datasheet for this CV. The CV value I found is 0.4.

 

[TiCl4]

Typically the Cv listed on the valve body and datasheet will be the 100% open Cv. A Cv of 0.4 does not sound oversized to me - an equal % CV will get down to 0.1 Cv around 60-70%, while a linear CV will likely be near 30% open to get a Cv of 0.11. In either case, the valve seems sized appropriately.

Further questions:
What is your flowmeter type?
What is the CV doing when operating? By this, I mean two things. First, you need to physically look at the valve during operating and see how much the stem is moving. Second, can you tell us what the output value is during normal operation. Does the output to the CV jump around when the flowmeter reading goes from 0.005 to 0.5?

Putting a restriction orifice downstream of the control valve will help the flow instability due to flashing in the valve, but it won't help an improperly designed flowmeter, a control valve that is sticking or needs other maintenance, or bad control loop tuning parameters.

 

[LittleInch]

Then something strange is happening with your flow meter as that size valve simply cannot be flowing the flows you're quoting at changes of 100 times.

Again that flow meter needs to be very small and what type is it? Maybe its cycle time is too high and its reading are not smooth enough to count enough pulses or whatever it is doing. If its a turbine then maybe it is jumping or just not enough pulses per second to make it accurate.

Or your valve is actually pulsing open and close rapidly to try and react to this very strange flow input (too high, too low, too high....) and the whole thing is having a heart attack....

Get your data and control system understood before you start messing around with things like orifices and manual valves.