19 replies to this topic

[Yogesh V]

I came across a safety valve sizing problem with following parameteres:

 

Flow= 65710 Kg/hr of Process Gas with mol wt = 28.02

Set pr = 36 Kg/cm2g with 10% overpressure

Normal temp = 41

Rel temp = 240 C

k= 1.4

z= 1, no superimposed back pressure

 

As per API 520, the area comes out to be 26.7 cm2. Hence orifice N is selected having areaa of 28 cm2.

 

When we put this data in Aspen Flare Net for Flare header sizing, the rated flow for this size comes out to be 59498 Kg/hr. and hence the software requires either normal flow to be changed to a value lower than this rated flow or the orifice designation to be changed to P type with higher area.

 

i think that the PSV size selected is adequate enough for the relief rate of 65710 Kg/hr. However i do not understand that how the rated flow is coming to be less.

 

please advice.

[shan]

The difference of the expected process flow (65710 Kg/hr) and PSV rated flow (59498 Kg/hr)is less than 10%, which is acceptable in the industry practice because all the assumptions, uncertainties and calculation methods. If you really want to fool Aspen FlareNet, you may increase your release pressure or lower release temperature to have the rated flow higher than the expected process flow.

[latexman]

The difference of the expected process flow (65710 Kg/hr) and PSV rated flow (59498 Kg/hr)is less than 10%, which is acceptable in the industry practice because all the assumptions, uncertainties and calculation methods. If you really want to fool Aspen FlareNet, you may increase your release pressure or lower release temperature to have the rated flow higher than the expected process flow.

 

I do not think this is acceptable.  If the process gas flow of the scenario is right, this tells me the PSV is too small.  What do others think?

Edited by latexman, 23 April 2013 - 09:49 AM.

[shan]

The PSV is adequately sized per API 520.

[PaoloPemi]

make sure that z (or density, depending from formulation you prefer) and other thermodynamic properties in formulation are the same on both cases (your calc's and the values adopted by software)

then decide which is the correct value to adopt,

you may find useful a previous thread discussing these topics, see

 

http://www.cheresour...ng-temperature/

[latexman]

Could someone point me to the section in API 520 that says this please?

[shan]

To me, the question is how to resolve the discrepancy between API 520 and FlareNet and not how to size, select, or verify a PSV.

[Bobby Strain]

You should check FlareNet discharge coefficient for the valve.

 

Bobby

[latexman]

No, it's a very simple question.  The flow rate required by the scenario is 65710 Kg/hr.  The rated flow of the PSV is 59498 Kg/hr.  Notice rated flow < sizing flow.  Where in API 520 does it say this is okay?

[Bobby Strain]

Latexman,

   If you calculate the required size for the required capacity using API formula, you will find that 26.7 cm2 is correct. This is using a discharge coefficient of .975. So, FlareNet is using some different calculation, which may not be correct.

 

Bobby

[latexman]

I have no experience with FlareNet, but that assumes that care has been taken with the inlet and outlet losses/line sizes inputted to FlareNet and there is no corresponding loss in capacity, which, if care was not taken, could be another explanation.

[fallah]

I came across a safety valve sizing problem with following parameteres:

 

Flow= 65710 Kg/hr of Process Gas with mol wt = 28.02

Set pr = 36 Kg/cm2g with 10% overpressure

Normal temp = 41

Rel temp = 240 C

k= 1.4

z= 1, no superimposed back pressure

 

As per API 520, the area comes out to be 26.7 cm2. Hence orifice N is selected having areaa of 28 cm2.

 

When we put this data in Aspen Flare Net for Flare header sizing, the rated flow for this size comes out to be 59498 Kg/hr. and hence the software requires either normal flow to be changed to a value lower than this rated flow or the orifice designation to be changed to P type with higher area.

 

i think that the PSV size selected is adequate enough for the relief rate of 65710 Kg/hr. However i do not understand that how the rated flow is coming to be less.

 

please advice.

 

Yogesh,

 

Appears the software has taken certified capacity of the PSV (as per ASME code limited to 90% of the measured/calculated flow and here would be 59498 kg/hr i.e. almost 65710*0.9) as its rated flow and because based on ASME rule the required capacity cannot be higher than certified one, has selected the next orifice designation (P type) to meet the code rule.

[PaoloPemi]

I did suppose (see previous post) there was a difference in values utilized by software and those utilized by Yogesh V when solving API formulation,

that was also the comment by Bobby Strain (different discharge coefficients),

the last comment by Fallah introduces another possibility,

are the details of the procedure not discused in software documentation ?

[chemsac2]

Yogesh,

 

It is due to k value. Value of 1.4 seems to be a guessed value. If you have used HYSYS, you might have considered Cp/Cv.

 

Flarenet uses Cp/(Cp-R) which is around 1.25. Using 1.25 in PSV sizing equation for your case, calls for P orifice.

 

Try Flarenet Cp/(Cp-R) and Z values in your PSV sizing spreadsheet and you will get same results as Flarenet.

 

I have seen this happen quiet a few times with Cp/(Cp-R) and Z being the cause every time.

 

Link given by PaoloPemi would give background on relevance of Cp/(Cp-R) and other k definitions. This is also covered in an article in November 2003 issue of Chemical Engineering titled "Using the ideal gas specific heat ratio for relief valve sizing".

 

Regards,

 

Sachin

[Bobby Strain]

Fallah makes reference to ASME relief requirements. But the description is not quite accurate. If one is to size the relief valve per ASME, then that code requires that the device to meet requirements must have the required capacity with 90% of the rated discharge coefficient. All valve manufacturers publish values for discharge coefficient along with the valve actual area in what is called "the red book". But you can probably research the subject for the exact language for both the API method and the ASME method.

 

Bobby

[Erwin APRIANDI]

Just to add,
 
You have to make sure that all the input for calculating the required area by using API 520 (I believe it is using an excel spreadsheet) and whatever input and calculation method used in Flarenet are same in order to have Apple to Apple comparison.
 
There are several types of calculation method used in Flarenet, they are (citation from Flarenet Help files):
 
The four sizing method options available are:

• 
  API_1976 – American Petroleum Institute method in the 1976 edition of RP 520 pt 1. No account is made of liquid flashing as it passes through the relief valve, thus this method is not recommended for either two phase or flashing fluids.
• 
  API_1993 – American Petroleum Institute method in the 1993 edition of RP 520 pt 1. Liquid flashing is handled by a simplified approach in which the fluid is flashed to the outlet pressure. The relative quantities of each phase at the outlet condition are then used at the inlet of the valve to determine the two phase capacity
• 
  API_2000 – American Petroleum Institute method in the 2000 edition of RP 520 pt 1. This method is often referred to as the Diers or Leung method. This is the recommended method for all two phase fluids.
• 
  HEM – Homogeneous Equilibrium method.

If the input and sizing methods used are same, but you still get different value then you can think of other possible causes. But also to make sure that you contact the Flarenet Developer (Aspentech) thru their support portal support.aspentech.com to revolve the problem in more detail. It is usually only takes one day for them to resolved it.

Edited by erwin.apriandi, 24 April 2013 - 08:24 PM.

[Yogesh V]

Thank you all fopr your responses.

 

If we forget the Flarenet for a moment, the PSV sizing is adeqaute enough as per the API. So there must not be any problem with respect to PSV sizing. (Apart from the possibility given by Fallah and a possibility of me having wrong values of k & z as said by Sachin).

 

Fallah, if you can provide me the clause of ASME about the rate flow certification please.

[fallah]

Yogesh,

 

Attached you can find the pages from ASME refers to what you did request...

Attached Files

•  Pages from ASME SECTION VIII D11.pdf   84.15KB   282 downloads

[Yogesh V]

Thanks fallah..... Thank you all..

[ChemEng01]

The API equations are generally quite conservative. It's common to get different answers when using software such as flarenet and HYSYS because they are dynamic.

If you used the HEM homogeneous equilibrium model) method for checking the rated flow of your RV it would probably come out much closer to that predicted by flarenet.

Note that if you use the HEM method fot RV sizing for a gas breakthrough case, HEM should also be used to determine the breakthrough rate through the failed valve. Vica versa if an equation such as that from API is used to determine breakthrough form a failed valve, an API equation should be used for RV sizing. You can not mix dynamic / equation methods when determining size of an RV as this can lead to over/under sized RVs.

Flarenet is fine to use for back pressure calculations, but if API equations have been used to determine the rated flow use the API calculated flow for determining the backpressure.

HEM is a great method for RV sizing as it can be used for gas, liquid, two phase and dense phase relief.

All the best