I understand you are looking to the second term of API formula.

I think your problem hasnt a simple answer- and for this fact API doesnt enter in details. Lets simplify the discussion looking to a "dummy "fluid mechanics model. One pipe is connected to a large vessel and the other end of a pipe exits into the atmosphere. The vessel is large enough so even in steady-state flow the vessel pressure ("large enough") remains constant. So the pipe is connected to an "infinite source" with density0, p0 ,T0 and discharge into atmosphere with p_atm. T_atm. The pipe is adiabatic insulated. There is a steady-state flow trough the pipe. At the endpoint, the gas pressure cannot drop to match the atmospheric pressure without the gas accelerating to critical conditions. A shock wave forms at the end of the pipe, resulting in a pressure discontinuity. In my opinion that discontinuity results in energy dissipated in the atmospheric turbulence so the pressure that would be taken into consideration in API reactive force formula is the "upstream shock" static pressure. This approach would be questionable but anyway is conservative. It is true also that, the choked flow at the endpoint can be count as Laval. However it is difficult to apply these equations to choked conditions, because the local conditions upstream are not known at the point of choking. Anyway, to perform this calculation, one must be able to calculate the stagnation pressure and temperature at the end of the pipe, upstream of the shock wave. How? The real behavior of gas flow in adiabatic pipe is the gas accelerates along the length of the pipe. As the pressure drops, the gas density will also drop and the dropping density must be balanced by an increase in velocity to maintain mass balance. So to make a calculation means to evaluate pressure, temperature and gas density as "local conditions" upstream of the shock wave. In this case, important is to simulate the flow trough the pipe. Even so, the fluid mechanics model is a little bit strange since you know one boundary condition at the end of the pipe (Mach=1). BTW, if you take a look in the section "Discussions" of the famous article "Steam Flow Through Safety valve Vent Pipes" by Brandmayer and Knebel, Mr. G.S.Liao said "In any case, the conditions at the vent pipe inlet can not be determined without knowing the conditions at the vent pipe outlet. Therefore, the calculations are always backward". True, thats why I think that article is just a fluid mechanics model that makes assumptions hard to be confirmed. For adiabatic conditions, the calculation for this model can be performed by hand. Some guidance you can found in the article "Gas-Flow

calculations: Dont Choke". Ive downloaded the article from AFT site. Mr. Trey Walters is the President of AFT (Applied Flow Technology). Anyway, the real case of PSV is a much more complicated simulation, so I would prefer to have the results from AFT package- Arrow is the module for Gas-Flow calculations. I dont know if what I've written clarifies your doubts. Probably neither of your formulas is what you need. In my understanding it's important to simulate what is happening through the piping system rather to know what is happening in final shock wave. Attachments CE%20Gasflow%20Reprint.pdf (145 downloads)

Edited by mariog (03/03/10 03:14 AM)

Top #33488 - 03/03/10 06:34 AM

Reply Quot e Quick Reply Quick Quote Notif y Email Post

anindya stress
Member Registered: 04/12/04 Posts: 486 Loc: London, UK

Mariog, Thanks for your input. I would like to share two papers with you on this subject. Due to copyright issues, I cannot post it on the forum. It will be great if you can share your email with me. Regards _________________________ anindya

Top #33509 - 03/04/10 01:43 AM

Reply Quote Quick Reply Quick Quote

Notify

Email Post

mariog
Member Registered: 09/29/07 Posts: 115 Loc: Romania Top

Ill be honored to share ideas about the subject. Ive tried to send you a message via My stuff/ Messages. I get a nice "Private messages are disabled". It is a setting of my profile or Coade service is disabled? Reply Quote Quick Reply Quick Quote Notify Email Post

#33510 - 03/04/10 02:06 AM

Re: Safety Valve reaction force [Re: mariog]

anindya stress
Member Registered: 04/12/04 Posts: 486 Loc: London, UK

Dear Mariog, I also could not send you a private message as against your name it is disabled. My email id is: [email protected] You can send a test mail at this address and I will Send you the papers. Regards _________________________ anindya

Top #33533 - 03/05/10 06:39 AM

Reply Quote Quick Reply Quick Quote

Notify

Email Post

anindya stress
Member Registered: 04/12/04 Posts: 486 Loc: London, UK

Mariog, I will send you the papers today. Regarding my post, this is how I could get 1=2 well at least nearly. Since k= 1.4 ,the expression for PX A that I got earlier can be written as : = W x 0.71 X SQRT( k x R' X Texit) The expression in my post is : (W x 1.88/g) X SQRT( k x g XR' X Texit) This is equal to : ((W X 1.88 )/ SQRT(g)) X SQRT( kxg XR'XTexit) Since g=9.81m/sec^2 and inverse of SQRT g= 0.319, the above expression comes to : WX1.88X0.319 X SQRT( kx R'XTexit) = W x 0.6 X SQRT( kx R'XTexit) Hope my algebra above was correct. Regards _________________________ anindya

Top

Reply Quote Quick Reply Quick Quote

Notify

Email Post

#33633 - 03/15/10 02:53 AM

Re: Safety Valve reaction force [Re: anindya stress]

mariog
Member Registered: 09/29/07 Posts: 115 Loc: Romania

Dear all, Anindya has sent me two articles about the subject. They are valuable, no doubt. However, the author has developed a quite complex system eight equations, all based on the assumptions of perfect fluid in steady-state, adiabatic, frictionless and isentropic flow. If the goal is to evaluate the choked sonic speed and choked fluid pressure, a serious simplification can be applied. There is a p_choked formula that supplements API formula being developed in the same way as that leading to the choked sonic speed expression. I have no idea why API does not detail the reaction force formula with p_choked expression. I attach my interpretation. The result is identical with the articles conclusion and is- in fact- exactly what Anindya has posted previously. So my contribution was just to simplify everything to "one page" presentation. In addition it is not subject to copyright. This discussion is not about where the choked flow and reaction force may appear. The formulas are evaluating the sonic speed and choked pressure in isentropic choked flow section- they are not showing where is that point. About this aspect, I think the reaction force must be count just where there is "the free jet" effect- at the end of open discharge piping and in that point there is a "choked" flow in many cases. It may be a long discussion about this aspect- anyway is what I am thinking. The formulas are not counting friction effects. Are interesting also the remarks I found in the articles "As mentioned before, the relief valve body has a very complex geometry. This complex geometry causes the relieving fluid to have a complex flow pattern. The fluid must change direction several times and there are certainly some frictional effects within the relief valve body. However, by making these assumptions, analytical expressions can be developed and these expressions yield interesting information and valuable insight about a rather complex problem. More complicated methods can certainly be utilized to analyze the body of the relief valve. However, these added complexities will certainly require a more detailed description of the geometry within the relief valve body and will almost certainly require numerical methods for computations." I second this interpretation, the formulas are just what we can do analytically and this fluid mechanics model is just one possible. The fact API has accepted this way is encouraging. Thank you, Anindya for your help.

Best regards. Attachments Choked flow pressure.pdf (147 downloads)