Pressure Reducing Valve or Orifice

[plantprowler]

I am planning a pilot scale reaction that is carried out in a large excess of H2O at 150 C in a single CSTR. Obviously, that CSTR operates under substantial autogenous pressure. A limpet coil provides the heat.

The part that confuses me is the mechanical design on the outlet side. Obviously, I must cool down below 100 C else the H2O will flash into steam on dropping P to 1 bara. So could I cool a pressured stream in a S&T HEX (say vs CW) & then reduce pressure to 0 barg?

What sort of device do I take the final pressure reduction over? A valve in a pipe? An orifice in a pipe? Something else?

Absolutely accurate pressure is *NOT* a concern. For testing purposes, can I use some sort of manual valve to do this?

 

[Latexman]

Technically, you can use about anything that has a significant pressure drop (54.3 psi) at the desired flow rate. A valve, orifice, backpressure regulator, venture, flow nozzle, capillary tube, etc. In my experience, most use a control valve because it is adjustable or a regulator because it is self-adjusting or an orifice because it is cheap.

Good luck,
Latexman

To a ChE, the glass is always full - 1/2 air and 1/2 water.

 

[plantprowler]

Thanks Latexman!

Can you elaborate more about a regulator? Is that like the Pressure Adjusting Valves you see often on Gas cylinder manifolds?

 

[Latexman]

For your application, I was thinking of a "back pressure regulator". It is a special regulator. It has some similarities to "Pressure Adjusting Valves you see often on Gas cylinder manifolds", BUT there are some differences too. For example, a regulator controls pressure downstream. A "back pressure regulator" controls pressure upstream. Similar, but opposite.

Good luck,
Latexman

To a ChE, the glass is always full - 1/2 air and 1/2 water.

 

[plantprowler]

Thanks Latexman!

That would indeed be perfect. I wasn't aware these are available for liquid duties. That actually sounds perfect for this application.

I will research more.

 

[zdas04]

There are a few BP valves for liquid service. I use the Baird a lot because it is effective and reasonably inexpensive (around $700 for a 2-inch as I recall) and only comes in threaded versions. Masoneilan makes a flanged liquid BPV that works really well, but costs $6k. I've never had an application that the Masoneilian was 10 times better than the Baird, but some of my clients are so anti-thread that they spend it. You need to check the specs and make sure that the valve you pick is rated for your temperatures. Fisher has a BP valve as well, I tried to test one for an application but it wouldn't maintain pressure in my application, I replaced it twice in the test and all 3 did the same thing. I don't know if I just had very bad luck or if their valve isn't very good.

[bold]David Simpson, PE[/bold]
MuleShoe Engineering

In questions of science, the authority of a thousand is not worth the humble reasoning of a single individual. Galileo Galilei, Italian Physicist

 

[lilliput1]

How about using a blowdown separator where the flash steam is vented to atmosphere and the condensed water is cooled by a cooler with thermostatic valve controlling city water feed to cool the mix to below 140F to allow it to be discharged to the drain. This is what is typically done to boiler blowdown.

If it make sense to recover heat consider economizer flash tank to preheat feedwater to boiler.

 

[zdas04]

Any BP valve you find is going to be of limited size (the Baird only comes in 2-inch for example, the Masoneilan has a 3-inch). The 4 and 6-inch gas BP valves then to not be available for liquids. In bigger pipes this function can be done with simple programming and a valve like the Fisher V-Ball that is available up to 24-inch (DN 600). You'll need a PLC that can query an upstream pressure transducer and drive a valve positioner.

[bold]David Simpson, PE[/bold]
MuleShoe Engineering

In questions of science, the authority of a thousand is not worth the humble reasoning of a single individual. Galileo Galilei, Italian Physicist

 

[SNORGY]

It's easier to mitigate the risk of noise with control valve trim selection than it is to deal with it coming from an orifice plate. I stuck a restriction orifice plate into a brown field compressor suction system on the inlet of the suction pressure control valve with a view towards limiting the blow-by sizing of the downstream PSV (600# / 300# spec break issue). My valiant attempt at effecting a quick and easy safeguard was so loud that they had to pull it and do the PRA / LOPA that everyone was trying to avoid. Otherwise, the PSV orifice grew from J to Q. Then a new flare system would have been required...on and on it goes.

Oh well...fast track brown field project, we gave it a shot. It worked and failed simultaneously.

Different situation for liquids, maybe, but it's worth double checking the noise level prediction.

 

[zdas04]

SNORGY,
I've run the noise calcs on liquid BP valves and the reality is never close to the models. I'm not sure that noise modeling in liquids is really ready for prime time.

[bold]David Simpson, PE[/bold]
MuleShoe Engineering

In questions of science, the authority of a thousand is not worth the humble reasoning of a single individual. Galileo Galilei, Italian Physicist

 

[plantprowler]

@lilliput:

Can't flash and vent the steam: The water is a reaction medium. If I vent I would lose some of my product with it.

@zdas

Thanks! The sensor + programming a valve sounds like one good solution.

I've run the noise calcs on liquid BP valves and the reality is never close to the models. I'm not sure that noise modeling in liquids is really ready for prime time.

Is the noise in reality less or more than the predictions?

 

[zdas04]

It seems to be a crap shoot. I ran a Baird in Queensland that was twice as noisy as I expected for the flow rate. I ran another Baird in New Mexico and it was quieter than background when the model suggested noise might be an issue. Masonelian has a calculator for their valve and it was about 30% low for the one valve that I ran it on. No idea how close it would be on a second test since the valves are so expensive that I don't have many takers for that version.

I was pretty surprised by these results since I find the manufacturer's noise calcs in gas to be very reliable. I just don't think on reflection that there has been enough focus on noise of liquid backpressure valves for the models to have matured because applications for the valves seem to be pretty rare. I've looked at this quite a lot since a liquid backpressure valve is crucial to the operations of an invention I'm marketing.

[bold]David Simpson, PE[/bold]
MuleShoe Engineering

In questions of science, the authority of a thousand is not worth the humble reasoning of a single individual. Galileo Galilei, Italian Physicist

 

[Jordan Hribar]

I'm curious Plantprowler, are you more interested in having the hot water under pressure stay in liquid form, or flash to steam ?
When you say venting the flash would make you loose product, is that the reasoning behind searching for a liquid BP Valve?

Venting to a closed loop receiver could be an option.
What is the flowrate and initial pressure of your CSTR outlet at 150C ?

Any BP valve you find is going to be of limited size (the Baird only comes in 2-inch for example, the Masoneilan has a 3-inch). The 4 and 6-inch gas BP valves then to not be available for liquids.

Armstrong has a GD-20R that goes up to 6". Minimum pressure differential required across the valve to work is 7 psig.

Fisher has a BP valve as well, I tried to test one for an application but it wouldn't maintain pressure in my application

Likewise ! Where I find the Fischer's blow everything else away, is when application calls for tight shut-off.

Different situation for liquids, maybe, but it's worth double checking the noise level prediction.

If you speak to the manufacturer's, I would venture to ask them at what liquid velocity they recommend is the maximum to avoid premature erosion when you do your pipe sizing. Have you seen noise occur under 10 ft/sec velocity ?



« Rien ne se perd, rien ne se crée : tout se transforme ».
— Antoine Laurent de Lavoisier (1743-1794)

 

[plantprowler]

I'm curious Plantprowler, are you more interested in having the hot water under pressure stay in liquid form, or flash to steam ?
When you say venting the flash would make you loose product, is that the reasoning behind searching for a liquid BP Valve?

Venting to a closed loop receiver could be an option.
What is the flowrate and initial pressure of your CSTR outlet at 150C ?

@Jordan:

Maybe I need to describe the goal better: There is a reaction between an organic chemical & water carried out in a large excess of water. The higher the T the faster the reaction. And hence 150 C is the target T. The pressure is the corresponding saturation pressure at 150 C.

If I flash & vent some amount of the organic chemical (reactant as well as product) will be lost in the vent steam I am thinking. Although both reactant & product are high boiling liquids with atmospheric Boiling Points of >175 C.

Hence the desire to cool & then drop pressure.

Venting to a closed loop receiver is an option but then I'd still need to cool & condense it so what's the advantage over not cooling first & then dropping pressure?

 

[Jordan Hribar]

@plantprowler:

OK I understand now. The final result is that you want a mixture of 1) water (in excess) and 2) organic chemical at atmospheric & room temperature to eventually use. Re-reading your application and understanding further, I don't see much use of a closed loop system for the reasons you have mentioned.

Allowing the mixture to flash and vent to atmospheric in my eyes would essentially be a distillation process. If you were to recuperate that vented steam and check it for product lost once cooled, I don't think it would have. Humidification via direct injection is widely used and works under the same principle.

My thought process :
- Back Pressure Regulator
- Let mixture flash in receiver and steam can vent through a free floating lever-type air vent.
- Depending the cost of product lost (if any worth caring about), you can always go through a simple drain separator after.
- Following these, even a steam filter can be used to further purify. Steam filters can be periodically back washed.
This hot (and hopefully product free) condensate can than be re-used for next batch if required.

@ 302F, 55psig your H20 will be at 9% flash steam. 9% of your H20 flow rate (lbs/hr) will in fact flash. From there you derive the receiver size required.



« Rien ne se perd, rien ne se crée : tout se transforme ».
— Antoine Laurent de Lavoisier (1743-1794)

 

[lilliput1]

If batch cooling the entire content of the CSTR is ok you can pump to a heat exchanger cooled with cooling tower water then back to the CSTR and pump back again the HEX till the temperature drops to the level desired. Locate tank outlet to pump at higher elevation than the return to the CSTR. Talk to a cooling tower manufacturer for help in sizing the cooling tower. It is illegal to use once through city water for cooling. Water makeup the cooling tower is required since the cooling effect is by evaporation of the cooling tower water. Compare plate type heat exchanger vs shell and tube.

 

[georgeverghese]

The mechanical design pressure of the downstream equipment should match up with the that used for calculating the relief load for the case when there is total loss of cooling medium to the upstream HX. Credit may be taken for the 100% wide open Cv of the throttle valve in this reilef load estimate.

 

[Jordan Hribar]

@georgeverghese:

If bypass line is installed it may trump the wide open capacity of the throttle valve. If so, this capacity should be used to size SRV.
As per code, safety relief should be set at 5% higher (but 5 psig minimum) above the equipment with the lowest pressure rating downstream of the throttle valve.

@plantprowler:

If budget is above $ 15k, I would look into a flooded-type HX. Throttle valve on outlet (condensate side) of HX which modulates condensate level within the HX according to desired pressure set points within your process. It actually would eliminate the need for a tank all together and would ensure no loss of product and control of your process temperature.

« Rien ne se perd, rien ne se crée : tout se transforme ».
— Antoine Laurent de Lavoisier (1743-1794)

 

[georgeverghese]

Hi Jordan, Agreed, the throttle valve bypass (if installed) should be checked also. It should preferably be locked close and opened only with work permit approval.
Without much more information, it does look like this throttle control valve should be operated on constant flow control or constant level control (level at CSTR) in order to get the required residence time for this reaction within the reactor, with minimum pressure override from CSTR pressure. CSTR temp would be heating medium flow regulation in the heating coils.

 

[plantprowler]

Thanks @george @Jordan

Those are very useful comments. I will try and make a P&ID / sketch based on these ideas and post it here.

One point that confuses me: I've used CSTRs before in Liquid-reaction atmospheric pressure duty where a certain liquid level is maintained.

But when used in pressurized mode is the vessel entirely filled with liquid? i.e. Up to the shaft seal at the top? Or is it typical to operate with a mixed Liquid-Vapor space inside the CSTR? (i.e. in some sense like a boiler)

i.e. Will there be a Liq-Gas interface or none inside the CSTR? All the mixing handbooks etc. speak of stirred vessels assuming a liquid level exists. I wasn't sure what happens in the pressurized and fully filled case. e.g. Normally baffles run up to liquid level. Here will they run till the top head?

I am wondering mainly from a operational perspective. My reaction won't care so long as I give it enough residence time.