Taming a Thermal Expansion Relief Valve

[Bambie]

Relief valves installed on ASME VIII Div 1 vessels solely for thermal expansion relief may have significant inlet flow restriction so long as the required relief capacity is met and proper operation is not affected per Clause UG-135 (b)(1).

The tube side of a water heater operates at 900 psig with steam on the shell at 150 psig and 365 degF. The tubes are designed for 1000 psig corresponding to the heater pump shutoff head and the set pressure for the ¾”x1”nps thermal expansion relief valve.

Murphy designed an acoustic interaction between the channel volume, RV inlet line and relief valve dynamics that causes chattering, valve seat damage and chronic leakage after a brief pressure transient comes along.

I would like to introduce an orifice plate at the relief valve inlet flanged connection to mess with the acoustics and introduce vapour dampening. I would also like to install a take-off and normally closed valve between the relief valve inlet and orifice to permit manual re-seating once cycling is detected.

We have tried relief valves with reduced blow down and different seat materials with no luck.

Could this work Does anyone have experience with this configuration?

 

[Latexman]

Acoustics, or is it the normal reseating pressure, which is 20-30% of opening pressure in liquid service? I would recommend looking into a pilot operated PSV where reseating pressure is typically 95% of opening pressure and write off the current installation as a learning experience. I think that will avert a lot of potential future issues with your suggestion and bring some normalcy to the operation.

Good luck,
Latexman

Technically, the glass is always full - 1/2 air and 1/2 water.

 

[Bambie]

Latexman, would you please elaborate on the 'potential future issues' associated with the orifice plate and manual blowdown?

Another suggestion that we considered was installing an accumulator upstream of the relief valve to remove pressure spikes with compress-ability.

 

[hacksaw]

you'll will have to better explain just what the problem really involves.

Thermal relief can only occur when the tube-side block valves are closed with steam heating on the shell side. Once the thermal expansion is relieved, the BFW pressure cannot go above 150#. That is well below the design pressure of the tube-side piping/equipment.

 

[Bambie]

Hacksaw,

Your first sentence is correct, however, the tube side of the water heater operates at 900 psig and the tubes are designed for 1000 psig which corresponds to the heater pump shutoff head, so we have a thermal relief valve lifting during transients and not re-seating due to an acoustic interaction with the vessel and the connecting inlet line. I am interested in anyone's experience with deliberately increasing inlet pressure drop and using the dampening effect of flashing vapor to de-tune a chatter-prone relief valve.

 

[Latexman]

OK, operating pressure = 900 psig. Set pressure = MAWP = 1000 psig.

In liquid service, the PSV inlet pressure has to reduce to 70-80% of opening pressure to reseat. It’s not acoustics! It’s the pressure in the inlet, the momentum of liquid flow, and the design of the disk.

So, that makes the reseat pressure = 700 psig. But it won’t reseat, not if operating pressure = 900 psig. Every time a pressure transient comes along and pops the PSV it will not reseat until someone goes out and forces it to reseat. With the modification being sought, that will not change. Someone will have to go out and open the bleed on the inlet, so enough pressure is dumped, plus the orifice pressure drop to be under 700 psig. I guess the hope is - they will be able to force it closed on the fly. But, will the operators get tired of this one-off PSV (I hope this is not common) and eventually leave the bleed open all the time? This could be what you are setting it up for.

The “potential future issues” I was referring to will most likely be caused by the human factors that come into play from the operators having to compensate for a bad design. Plus, the extra wear and tear this PSV will see (while it chatters away waiting for someone to force it closed time and time again) will be more reason for “potential future issues” of a mechanical/repair nature.

As suggested, please look into a pilot operated PSV. Reseat pressure = 950 psig. When the transient dissipates (yes, this part is acoustics), it will close automagically. Oh, it will be a "close design", but at least it stands a chance! The operators will love you for it. I don’t think they will love baby sitting the proposed plan for the long term.

I have no experience putting a restriction orifice in the inlet to a PSV. Personally, that just seems wrong on several levels. It wasn’t long ago, this was against Code! Maybe, that’s my problem with it. You know, an old dog.

IMO, it’s a bad idea. An engineering control hamstrung with human factors. I doubt the reliability of this final protection layer will be up to standards after some time. If there are PHA scenarios this PSV is addressing, I’d recommend asking the PHA experts if this set up will affect the credits a PSV normally possesses.


Good luck,
Latexman

Technically, the glass is always full - 1/2 air and 1/2 water.

 

[hacksaw]

"thermal reliefs" are not designed for the service you describe. You need to consult the Boiler/Pressure Vessels Codes in your area or get a PE involved to review your piping/equipment design.

Of greater concern is that your proposed "solution" is not permitted in relief valve installations, period. Were anyone to bring it up in a project meeting, for example, they would be asked to leave the room.

Life is tough, but blowing a heat exchanger or damaging adjacent piping and valves carries heavier penalties....

 

[Latexman]

Bambie,

There are others in the Safety Relief Valve engineering forum, forum1203 ,who know Code and reliefs much better that I. Please consider opening a new post there and putting a link to this post, thread391-358149 , to see if you can get their input.

Good luck,
Latexman

Technically, the glass is always full - 1/2 air and 1/2 water.

 

[Bambie]

ASME VIII Div 1 and API 520 are very clear on requirements for thermal expansion relief valves, but thanks for the suggestion Latexman.

 

[abeltio]

From the description: could it be that the application is not "thermal expansion relief valve"?
For steam at 1000 psig the Saturated Steam Temperature is 546.43 °F
So, the tube side will never get to steam phase (will always stay liquid)

The "thermal expansion relief valves" are designed for a change in temperature in a blocked pipe which rises much slower than the rather sudden change in pressure due to a valve shutoff, which is a case of "water hammer".

From this reference:

http://www.irc.wisc.edu/file.php?id=343‎

All ASME stamped vessels and components require over pressure protection [UG-125]
•Vessels & other stamped components that are to operate completely filled with liquid shall be equipped with pressure relief devices designed for liquid service, unless otherwise protected against over pressure. [UG-125(f)]
•Relief valves must be certified [UG-129]

So, using a "thermal expansion relief valve" designed for piping (non ASME rated equipment) on an ASME stamped vessel is not allowed.

What’s a liquid service relief?
•Liquid service relief valve: a pressure-actuated valve closed by a spring or other means, designed to automatically relieve liquid pressure in excess of its setting.
•Intended for use with ASME [rated] pressure vessels or other stamped equipment.
•Certified by the National Board and rated in U.S. gallons per minute of 60°F [16°C] water flow at 110% of its setting

Perhaps we need to go back to the basics here and see what whas actually installed in the piping and in the heat exchanger.
If the piping is fitted with a hydrostatic pressure relief device that would not allow the pressure to rise above MAWP of the vessel, and it is impossible "lock" the pressure inside the vessel during operation (e.g. manual valves), then valve in the vessel could be eliminated ("unless otherwise protected against overpressure").


saludos.
a.

 

[Latexman]

No problem, I'm not that familiar with Div. I. All my experience is Div. VIII.

Good luck,
Latexman

Technically, the glass is always full - 1/2 air and 1/2 water.

 

[don1980]

I think you're likely dealing with an acoustic resonance. You can find info on the internet on how to calculate the acoustic line length. You don't want the line to be longer than that value.

This is becoming a well recognized problenm for PSVs in liquid services, and it one that likey explains the distructive chattering that you see. In the past, this chattering was attributed to pressure drop, but it's starting to look like accoustic interaction is the more likely cause. The solution is to shorten the inlet line, below the acoustic line length.

 

[Latexman]

Interesting, this is new to me. Bambie, please let us know how things turn out so we can learn along with you.

Good luck,
Latexman

Technically, the glass is always full - 1/2 air and 1/2 water.

 

[Bambie]

Abeltio,
This water heater has motorized isolation valves on the tube bundle inlet and outlet piping. The relief valve is installed on an 18" long 3/4"nps pipe welded to the heater channel and is specifically for the purpose of providing over-pressure protection in the event these valves are closed when steam enters the shell.

ASME VIII Division 1 Clause UG-135 (b)(1) states that significant pressure drop is permissible upstream of these valves so long as the required relief capacity is met and proper operation is not affected.

The transient pressure spike that upsets this relief valve is fractions of a second in duration, however, the relief valve cycles until the seat is damaged and a chronic and expensive leak results.

I would like to introduce an orifice plate at the relief valve inlet flanged connection to change a single phase acoustic response to a two phase acoustic response. This should eliminate the resonance between valve fundamental frequency and pressure wave by slowing it down, which should allow time for re-seating to occur.

As a contingency, the orifice holder would be modified with a branch line and isolation valve to provide operators with the ability to manually re-seat. "Leaving the bleed on" as Latexman suggests might not be a bad idea because a very small flow would keep the inlet hot and ready to flash.

 

[marty007]

Bambie,

I read through all of the posts so please forgive me if I missed it, but are you able to tell us specifically what type (brand? model?) of relief valve you currently have installed? This would provide more information to better define your problem.

Cheers,
Marty

 

[LittleInch]

I've read this post quite a few times and still come to the same conclusion that you appear to be trying to fix the symptom and not the cause.

The cause, as far as I can see from the info supplied, is that you are operating (900 psig) too close to your design limit (1000) for the equipment you've installed - seemingly simple spring relief thermal pressure safety valves. essentially I'm with latexman on this one.

Your transient event may only last a short period, but it is the key trigger to your problems which then occur as a result. Given that this seems to happen either with the pump still running or with a large reservoir of pressure, your valve may end up being affected by acoustics and similar pressure waves, but I think you'll spend a lot more time and effort trying to fix this rather than addressing the key causes of preventing the transient I the first place or replacing your simple valve with a more sophisticated valve able to re-seat at a much higher pressure or one not affected by continuous flow through the valve. It might be a larger valve will do this...

Don't rule things out too early in your assessment of which is going to be the cheapest and best solution.


My motto: Learn something new every day

Also: There's usually a good reason why everyone does it that way

 

[abeltio]

The thermal expansion relief valves are very small
The liquid service relief valves are bigger and certified by the National Board and rated in U.S. gallons per minute of 60°F [16°C] water flow at 110% of its setting

I would make sure that the valve installed is the "liquid service relief valve" and not the "thermal expansion relief valve", two different applications that use different valves.

saludos.
a.

 

[Bambie]

Marty007,

The relief valve is a Consolidated 3/4" x 1" model 19096M, capacity certified to deliver 81 USgpm through a 0.096 sqin orifice (0.35" dia) and 1000 psig set pressure.

LittleInch,

The cost of eliminating all process transients that upset this valve would be prohibitive, that is why an unconventional approach is necessary.

 

[LittleInch]

Ok, you can't tame your transients, but you can modify the valve.

Reading the 19000 series manual shows the following item on page 7:

Differential Between Operating and Set
Pressures—Valves installed in process
services will generally give best results if
the operating pressure does not exceed
90% of the set pressure. However, on
pump and compressor discharge lines, the
differential required between the operating
and set pressures may be greater because
of pressure pulsations coming from a
reciprocating piston. The valve should be
set as far above the operating pressure as
possible.

Also when was the last time this valve was dead weight tested?

This is a pretty basic TRV and it seems to me that you're very fixed on going down the acoustic vibration route instead of properly considering the possibility that your valve is just the wrong type for your particular application...

Also there's no info about what your back pressure is or whether you can get transients coming back at you for the discharge piping


My motto: Learn something new every day

Also: There's usually a good reason why everyone does it that way

 

[Latexman]

Why is there resistance to a pilot operated PSV? It is obvious to me it would help immensly. Yes, you still may need to dampen the acoustics, but it would be to a lesser degree with a pilot operated PSV. Have you considered baffling the inlet line, similar to a car muffler?

Good luck,
Latexman

Technically, the glass is always full - 1/2 air and 1/2 water.