Hydraulic expansion - blocked in liquid (water) scenario - API Clause 5.14.4. 1

[johcatrac]

We have a blocked in liquid scenario between two isolation valves and the medium is water. In our case the temperature is rising from 20°C to 110°C.

My issue with the formula {5.14.4.1-d-(3)} is the variability of the following three variables with respect to pressure and temperature.

αv {is the cubic expansion coefficient of the liquid, expressed in 1/°C (1/°F)}
αl {is the linear expansion coefficient of metal wall, expressed in 1/°C (1/°F)}
χ {is the isothermal compressibility coefficient of the liquid, expressed in 1/kPa (1/psi)}

How should I account for the above 3 variables' variability in the formula? Any input or help would be greatly appreciated!

 

[LittleInch]

see this FAQ

http://www.eng-tips.com/faqs.cfm?fid=1339

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

 

[Latexman]

The physical properties of water are well understood compared to other chemicals. Physical property variability is covered by ASME safety factors in the piping design. A thermal relief is sized at 10% overpressure. The piping is hydro tested to 30% (or 50%) overpressure. The theoretical failure point is what? 200-400% overpressure? Don't worry, just get the best data you can find and carry on. More than likely what you put in will be over sized anyway.

Good luck,
Latexman

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

 

[johcatrac]

Thanks LittleInch & Latexman.

We want to stay away from an RV for now, as installing an RV will be a lot of work in a nuke facility such as ours. We want to analyze as much as we can before we throw in the towel and go with an RV.

 

[Latexman]

Here's a thread from My Thread Archive on the subject: thread124-171158 .

Good luck,
Latexman

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

 

[jari001]

This looks like a job for integral calculus! But going through the info LittleInch and Latexman provided, I think you have some decent estimation techniques to use.

 

[don1980]

If the liquid temperature can rise from 20C to 110C, then I think these calculations are a forlorn hope. Focus on preventing the system from being totally isolated, if installing a thermal PRV isn't a practical option. Alternatively, consider options for ensuring that the system isn't actually 100% liquid-full. We generally try to eliminate gas pockets from liquid piping systems, but in this case such a pocket can effectively prevent overpressure.

 

[Bill3752]

The API 520 calculation is very straightforward. Regarding the sensitivity of the data, you will need specific volume data to several significant figures; as indicated above this should be easy to find.

You will find in API 520 (2008) p. 34 equation (3), the calculation for pressure. The several times I have checked for different fluids, the pressure has far exceeded the design pressure of the piping - pressure simply rises quickly as temperature increases. You didn't say if you were protecting piping or a vessel, but the same logic would apply to the latter.

If you are discharging to atmosphere, then you will encounter flashing flow in the discharge piping (relief temp = 110°C). You will want to consider in your piping calculations.

 

[LittleInch]

The last time I did this it is somewhere in the region of 3 to 5 bar per degree C., so maybe up to 400 bar or so assuming a low initial pressure (ouch)

However the volume involved is very small - some sort of closed volume expansion vessel will take care of 90% of that pressure rise. If the volume in the pipe isn't very big, then neither will be the expansion vessel.

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

 

[gelsi]

Often, the order of magnitude is greater than 5 bar for 1 °C of temperature rise, so 90°C causes overpressure, without doubt.

Just to think about alternatives:
Can you install a bypass check valve arround one gate?
Can you drill one of the two or more gates?

 

[johcatrac]

I am sorry, but how would the bypass valve work? We do not have gate but globe isolation valves that bound the isolatable volume.

Also, when one gets a resultant pressure of say 13,000 psig for the blocked in liquid what code stress limits are used for the evaluation of the piping and the associated fittings/valves? Is it the normal stress limits or are you allowed to flirt with the yield/tensile stresses of the applicable materials?