Rise in gas pressure at abrupt valve closure 4

[Iomcube]

In this p&id there is a constant feeding of H2 into the circuit. Now only one consumer of H2 gas exists i.e acid furnace, which requires a bit pressurised H2 gas. If there is any tripping at the furnace at that instant vessel pressure will increase abruptly beyond mentioned setpoint. How can we estimate that pressure rise?

I know volume of vessel & holdup volumes of piping as well. I can convert mentioned feed volumetric flow to furnace to mass flow & thereby can calculate rise in density of gas in the vessel for instance. Perhaps packages like CoolProp (or any equation of state package) can back calculate rise in pressure at that density (assuming T remains same)

https://files.engineering.com/getfi...9e9-43d5-8de4-db378e3bf233&file=H2Circuit.PDF

 

[Iomcube]

@ LittleInch
So in this case you could simply forget about your water seal and vent and the PCV at 250 mbar because it will never get there.

So all you need is your FCV before the burner, or maybe your cooler, but nothing else in terms of vents or pressure control.

No, Like I said upstream pressure is low. Setpoint of PCV is 250mbarg, that's furnace vendor requirement for gas pressure. In the document vwssel pressure setpoint indicate this. Also we need some redundancy to avoid high-pressure at furnace tripping thus the water seal. I previously added quartz burner will break at pressures close to 600mbarg


Also be aware that "safe vent" for hydrogen
As you can see N2 purging is included at water seal, Saturated steam also is introduced at H2 vent outlet to avoid any explosion during thunderstorms ...

 

[LittleInch]

Tom,

You seem to have decided all this stuff is required, but your information says it isn't. If the u/s pressure "cannot increase [more than?] 180mbarg" - Your words - then I cannot see how you can ever get to 250 mbarg when the PCV starts working, far less the 600 mbarg limit of your burner or the 500 mbarg of your vent seal.

Repeating pressure control and relief systems from the data you've provided just looks like a waste of money. IMHO.

Remember - More details = better answers
Also: If you get a response it's polite to respond to it.

 

[shvet]

@Iomcube
No

 

[Iomcube]

@ LittleInch
The compressor is a pd compressor (as indicated) which translates 160mbarg to 250mbarg. At the downstream of compressor till the furnace pressure will increase & can go beyond 250 easily if FCV closes at furnace tripping. The compressor (like all pd compressor) have a check calve at suction of H2 so u/s pressure is not issue but downstream is

 

[LittleInch]

OK, now we're getting somewhere.

Your system to me though isn't looking good.

Your burner trips and closes in say 3 seconds.
Your pressure starts to rise, but not ridiculously fast as you have a buffer vessel.
Lets assume your PCV is reasonably fast acting and also recognises a rising pressure so starts to close and is fully closed before the pressure exceeds say 275 or 300 mbarg.
your Note 2 vent PCV opens, but only to 30% which you've stated is only able to handle half of the flow rate.
So the PD pump keeps pumping and there is no apparent trip or other relief valve or vent u/s the PRV?

So this pressure rise on furnace trip will almost certainly lead to the vent opening and pressure will rise until the PD pump stops.

The rate of pressure rise and what it settles out to is dependant on the speed of closure of the PRV as far as I can see. Modelling that requires transient software like you linked to and all the parameters entered into it such as valve closure speed etc.

Remember - More details = better answers
Also: If you get a response it's polite to respond to it.

 

[Iomcube]

@ LittleInch
So this pressure rise on furnace trip will almost certainly lead to the vent opening and pressure will rise until the PD pump stops.

So at some u/s pressure an interlock with compressor motor can be thought of. However this 5m water seal should avoid pressures beyond 500mbarg

Note 2 vent PCV opens, but only to 30% which you've stated is only able to handle half of the flow rate.

This 30% value; I will review
Per our experience H2 egress for system is very quick & pressure loss occurs abruptly that's why this sizing of valve is done.

I will calculate & attach an Smath Sheet here ...which will include some calculations. Hope you are free to review. Will do it in 2 days

 

[LittleInch]

OK, Last point though is that on your simplified circuit the 5m water seal is Down stream of the PRV. Therefore should only see 250 mbarg if the PRV is working properly.

ALL PD pumps I have ever see have a spring relief valve system installed to provide a high integrity pressure relief. Your system doesn't seem to have this.

Simple Interlocks and set points for the vent valve are not normally considered sufficiently reliable.

Remember - More details = better answers
Also: If you get a response it's polite to respond to it.

 

[Iomcube]

Thanks for all your inputs

ALL PD pumps I have ever see have a spring relief valve system installed to provide a high integrity pressure relief. Your system doesn't seem to have this.

Correct, it has 800mbarg setting

 

[georgeverghese]

In addition to volume boosters at PCV supply to furnace (SP = 250mbarg) and compressor PCV to vent (SP = 270mbarg), suggest the following to improve transient response to sudden loss in demand at the furnace
a)Fully automate the PIC - PCV control loop to vent and raise the SP here to say 300-350mbarg
b)Check that this vent PCV has a quick opening trim (which will give you high flow at low opening).

Why do you call the H2 supply to furnace control valve as "PRV" ? - it should be a fully automated standard PIC - PCV control loop. No controller shown either on sketch. Using a pilot operated regulator cannot give you good response for sudden loss in downstream demand.

 

[Iomcube]

@ LittleInch
@ georgeverghese

I have updated the drawing & added further details which represent field actuality. I had also included paper / per calculation which is based on previous discussion

@ georgeverghese
Thanks for your recommendations.

Fully automate the PIC - PCV control loop to vent and raise the SP here to say 300-350mbarg
I am unable to understand this

Why do you call the H2 supply to furnace control valve as "PRV"
because the loop maintains pressure at buffer vessel which is pressurised gas source to furnace. In future we will hook up another furnace such that x2 furnaces running in parallel after the vessel indicated, with each requiring certain flow (FCV) depending upon HCl production required. However pressure must be regulated at 250mbarg

https://files.engineering.com/getfi...e-b6a2-a832208d2c93&file=H2Circuit-merged.pdf

 

[LittleInch]

Where is flow / mass control on this system?

I get your 936nm3/hr to be about 90kg/hr, so if your inlet flow is 101 you are going to pretty fast have an overpressure in the system up stream the PRV?

If the PRV is actually in operation, i.e. pressure U/s the FCV is ~250 mbarg then the system will work better, but its not clear how you control the pressure / flow into the system to avoid venting it all the time. your 300 - 35o mabrg setpoint is too close to the 250 set point and will go off frequently or continuously if your inlet flow is > the flow out to the burners.

Alternatively if the burners want more than 101 kg/hr then your system will rapidly go to the pressure form the compressor and your PRV will be fully open. How it then reacts to a sudden pressure rise is uncertain, but will be slower than if it is already closed to some extent to limit d/s pressure to 250 mbarg.

Remember - More details = better answers
Also: If you get a response it's polite to respond to it.

 

[Iomcube]

@ LittleInch
I get your 936nm3/hr to be about 90kg/hr, so if your inlet flow is 101 you are going to pretty fast have an overpressure in the system up stream the PRV?
I think one thing is missing here. When you burn Cl2 with H2 ...H2 is always 7-10% excess to avoid Chlorinated HCl. That should account for the discrepancy approximately.

Alternatively if the burners want more than 101 kg/hr then your system will rapidly go to the pressure form the compressor and your PRV will be fully open. How it then reacts to a sudden pressure rise is uncertain, but will be slower than if it is already closed to some extent to limit d/s pressure to 250 mbarg.
This scenario you should not assume...

your 300 - 350 mabrg setpoint is too close to the 250 set point and will go off frequently or continuously if your inlet flow is > the flow out to the burners.
How close or far PCV setpoint is to be from vessel / PRV setpoint I think requires dynamic analysis as mentioned by other users. As I mentioned before (per our experience) rate of H2 egress is too fast that it is to be very finely controlled

 

[LittleInch]

The point is that you have flow control at the end, but no apparent flow control on the inlet side of your diagram from the electrolyser / compressor.

Over time even a small discrepancy between flow in and flow out will lead to either high pressure or low pressure in your system. This is the downside of using a PD compressor with a flow control valve....

Remember - More details = better answers
Also: If you get a response it's polite to respond to it.

 

[georgeverghese]

"Fully automate the PIC - PCV control loop to vent and raise the SP here to say 300-350mbarg
I am unable to understand this"

Right now, you have only partial auto control of this vent control valve, as you've stated - beyond 30% open, it is on manual control from DCS operator. This vent control loop ought to be fully automated so that the PIC here has full control of this vent valve throughout. Obviously the "PRV" feeding the furnace should also be operated on full auto by a standard PIC at DCS.

What is max normal compressor discharge pressure (ie upstream of "PRV") at design flow of 101kg/hr ? The setpoint of the vent PIC - PCV on compressor discharge has to be
a) Some what higher than this max normal compressor discharge pressure
b) Some what lower than 500mbarg, which is when the liquid seal in the seal drum will break.

To improve sensitivity of the PRV control loop, it should be obvious that (a) the control valve should not be oversized and (b) the pressure transmitter feeding the PIC should have a range appropriate for this service (say 0-500mbarg). The bigger the range, the less sensitive the control loop will be.

Can you tell us if you have some kind of capacity control at this compressor ? Internal recycle valve or some kind of suction unloader or similar? And does this capacity controller get its feed signal from compressor discharge pressure?

 

[LittleInch]

George,

The liquid seal in the seal drum (the 5m water column I assume you mean) is downstream of the PRV which should limit pressure in this section to 250 mbar.

So the set point for the vent line is not connected to that seal?

Remember - More details = better answers
Also: If you get a response it's polite to respond to it.

 

[georgeverghese]

@LI,
For safety reasons, and since loss of the liquid seal leads to release of flammable H2 to the atm, the set point of this vent pressure controller should be below 500mbarg for the case when this "PRV" doesnt work as it should; ie if it fails open or tends to open during a sudden pressure transient such as this.

Suspect the capacity controller for this PD compressor (waiting on @iomcube reponse)also has a SP less than 500mbarg for the same reason.

 

[Iomcube]

@ LittleInch
Over time even a small discrepancy between flow in and flow out will lead to either high pressure or low pressure in your system. This is the downside of using a PD compressor with a flow control valve..
True, I think this discrepancy can be assumed not to be present. Even if it is there the furnace can be operated 20% more than the stated capacity. Although the outlet HCl (aq) w.t% which is normally 32% will be reduced to 30.5% [at higher temperature less solubility of HCl gas]

So the set point for the vent line is not connected to that seal?
No, the seal only protects the overpressure scenarios after the PCV. As you can see the relief valve at compressor before PCV is at 800mbarg ...so this seal is a relief valve at 500mbarg after PCV

@ georgeverghese
Right now, you have only partial auto control of this vent control valve, as you've stated - beyond 30% open, it is on manual control from DCS operator.
True

This vent control loop ought to be fully automated so that the PIC here has full control of this vent valve throughout. Obviously the "PRV" feeding the furnace should also be operated on full auto by a standard PIC at DCS.
Per the illustration there are x2 PTs one for PSV & one for PCV. PSV need to operate at emergency situations thus closed loop I don't see as necessary, however it can be close looped with its own PT before PCV. I will review

Can you tell us if you have some kind of capacity control at this compressor ? Internal recycle valve or some kind of suction unloader or similar? And does this capacity controller get its feed signal from compressor discharge pressure?
In the chloralkali industry the electrolyser produce H2 / Cl2 as per Faradays Law & operates at 16kA. So depending upon kA, no. of electrodes, cell efficiency the exact amount of gases can be easily calculated accounting for water saturation. Images:

https://www.eng-tips.com/viewthread.cfm?qid=451851

Now gases must be consumed downstream to maintain pressures across highly sensitive electrolyser membrane. Thus this H2 compressor has a manual recycle valve [for startup purposes] but no auto recycle valve; rather auto vent valve
For Cl2 [which cannot be release to atmosphere] its vent valve is opened to a suction header where it goes to CaCl2 or NaOH batch towers producing ...hypo

What is max normal compressor discharge pressure (ie upstream of "PRV") at design flow of 101kg/hr ?
Currently we have one electrolyser one year after we will have x2 thus H2 will be 202kg/hr. So compressor name plate says 220kg/hr 0.6barg

To improve sensitivity of the PRV control loop, it should be obvious that (a) the control valve should not be oversized and (b) the pressure transmitter feeding the PIC should have a range appropriate for this service (say 0-500mbarg). The bigger the range, the less sensitive the control loop will be.
Transmitter range is -10 - 90 kPag i.e -100 to 900mbarg. Regarding sizing, I think I should try a bit undersized valve on a bit bigger connecting line. So that latter I can use same line for a larger valve [x2 electrolysers]

 

[georgeverghese]

So I would suggest the fully automated vent PIC-PCV be set at no more than 400-450mbarg.
Would be better if the full calibration scale for PT used for the PIC-PRV be reduced to 500mbarg (as compared to the current 900mbarg) to improve sensitivity.
And the PRV be rate checked for the future operation to enable 220kg/hr at 300mbarg upstream and 250mbarg downstream at about 70-80% open.

 

[Iomcube]

@ georgeverghese
Thanks for recommendations

 

[LittleInch]

Iomcube,

"True, I think this discrepancy can be assumed not to be present. "

errr, don't think so unless your flow set point controller is somehow linked to the pressure in the cylinder or upstream.

As shown, it controls to a set flow rate.

Also now we suddenly find the compressor is sized for 220 kg/hr. How are you going to get it to only do 101 kg/hr?

I'm really starting to hope that this is all just experimenting here and that this will go through a proper review, HAZOP etc??

Remember - More details = better answers
Also: If you get a response it's polite to respond to it.