Calculation of Isentropic Efficiency / PV work term in HYSYS for blowdown modelling.

[Andrew291989]

Hello,

I am trying to improve on the accuracy of results produced when using the HYSYS dynamic depressurising utility. I understand the purpose of the isentropic efficiency option when setting up the model and that 100% efficiency gives the lowest temperature downstream.

It is my understanding that a perfect isentropic expansion is impossible and I have created a model that gives me the isentropic efficiency factor necessary to approximate an isoenergetic expansion (constant internal energy). I believed that modelling an isoenergetic expansion would give me a less conservative "worst case" downstream temperature, however I am getting isentropic efficiencies lower than I would expect. I was expecting the model to give me approx 80-99% isentropic efficiency for a gas expansion, whereas I am getting in the region of 50-70% for most gases.

I am reasonably confident that my model of the isoenergetic expansion works as advertised so I am beginning to question my initial assumption. Is an isoenergetic expansion a reasonable way to model a blowdown situation? If not, does anyone know a way to get a decent estimate of the isentropic efficiency (other than randomly picking a number between 87 and 100%)

I would appreciate any help you can offer.

Andrew.

 

[apetri]

you may consider a constant energy flash,
this operation is available in Prode Properties (and may be other software tools), see

http://www.prode.com/docs/pppman.pdf

for additional information

 

[MortenA]

Think about your thermodynamics!

A valve is NOT an isentropic process! Its as close to an isenthalpic process as you might get! Theres no point in producing a "worst case" that is not relevant (unless it dosnt give you any problems of course and the methode is significantly easier to use).

An expander is closer to an isentropic process, but not perfet.

A piston (could) be close to an isentropic process - so could the expansion takiing place UPSTREAM the valve be. Upstream the valve an isentropic efficency of close to a hundre (but i seem to recall that the HYSYS manual recommends someting around 90-95% and i cant remember where to input the valvue).

But what is isoenergetic?

Best regards

Morten

 

[Andrew291989]

Thank you both for your replies. I'm sorry I haven't responded sooner, however, I have been away. I will explore the Prode option further as it seems like it may give the results I require. Morten, In the HYSYS depressurising utility the isentropic approximation applies to the expansion of gas inside the vessel itself as the pressure is reduced during blowdown. Due to the fast nature of the depressurisation, this is much closer to an isentropic process rather than an isenthalpic one. By isoenergetic process I am referring to one in which the internal energy remains constant, giving a final temperature somewhere between the two extremes (isentropic and isenthalpic.) I understand how to adjust the isentropic efficiency in HYSYS, the problem is reliably calculating a number to set it at.

 

[MortenA]

I think isoenergetic is word you have invented yourself ? The situation that you describe seems to be isentropic to me.

IMHO i think you will have to accept a rule of thumb for the isentropic efficiency (for the upstream process) -or find some reference data for a similar systm? As you state youself its closer to isentropic (100%) than isenthalpic (0%) and i dont belive that has anything to sdo with the speed - its the nature of the system. Actually i think the ideal situation is a reversible process - usually best mimicked in rather slow systems.

The efficiency itself should be eff== (h2a - h1)/(h2s - h1)
h1 = enthalpy at the inlet/start
h2a = enthalpy of actual process at the exit/end
h2s = enthalpy of isentropic process at the exit/end

 

[apetri]

there are several different methodologies which you can consider, for example for supercritical fluids I prefer the one-dimensional isentropic nozzle flow assumption, see the annex B in API 520 (2008) and B.1.3 Numerical Integration Example
in my case Prode Properties (prode.com) provides quickly the solution to the multiple isentropic flashes and all the required properties,
for me the flexibility provided by Excel and Prode Properties is a real advantage in these cases.

 

[Andrew291989]

Thanks again for your assistance guys. I have no problem at all modelling an isentropic flash fo any fluid. I have built a few HYSYS models that allow me to do this reliably, althugh the Prode method is also interesting.

Morten, I can assure you that I have not invented the word isoenergetic. I will admit it is not a concept I was familiar with before I started this project, however, it definitely exists and I have even modelled it. Unfortunately I don't seem to be able to paste diagrams on here as I have found several papers explaining the concept better than I can. As it happens I have now determined that whilst my isoenergetic expansion model is accurate, It will not provide the approximation for isentropic efficiency that I require.

I have now moved on to using the temperature drop of the vessel contents over a time step(given in HYSYS)to calculate the enthalpy loss. Comparing this relative to the enthalpy loss of an isentropic flash gives me the isentropic efficiency. This seems to work early in the blowdown, however, later on my efficiency tends to reduce to around 20%. I'm not sure if this latest approach will work and to be honest I'm close to abandoning the project and just estimating a number but if you have any ideas I'd be interested to hear them.

 

[Longinthetooth]

I have followed this discussion with interest, although I have not got to grips with exactly what the point is. To my mind isenthalpic/isentropic/isoenergetic (Havent heard of that word before) only applies to fixed ammounts of gas. If you are depressuring through a valve, you are losing mass and energy from the vessel as you go, so describing the vessel contents as iso-whatever is meaningless. Am I mssing something?

 

[apetri]

the isentropic nozzle flow assumption is the basis of PSV std. model,
the solution is complex cause real fluids have not constant properties and / or state,
with a simulator (in my case Prode Properties) you can solve at different pressure steps and then integrate numerically (see above) this gives accurate results.

 

[MortenA]

Totally derailing the tread isoenergetic is not word that is in wikipedia (OK that may not be a conclusive argument) but my trusty old Smith& Van Ness does not recognise it. And right now i cant see the difference to isenthalpic? When googling i get a lot of hits concerning diets and metobolism...


But i guess that my general problem is that i still cant see what the purpose of the exersise is - so i will just shut my mouth

 

[Latexman]

Andrew,

I do not understand why the Internal Energy (or specific Internal Energy, for that matter, if that is what you meant) of a vessel's contents would remain constant during a depressurization. Can you explain please?

Upload your attachments to ENGINEERING.com using the link below "Reply To This Thread" and then copy/paste the url.

Good luck,
Latexman

 

[Andrew291989]

Hi everyone,

I seem to have unwittingly caused a great deal of confusion with this thread. I'll try and explain myself so it is a bit more clear. Firstly, I am not 100% sure that isoenergetic is the "official" word for the concept that I was referring to although it does seem to be used a lot in the papers I have read. Secondly. I am not trying to model an isoenergetic (or whatever the word may be) process anymore. I can't fully explain why but I am relatively certain this isn't correct. I have put in an url for the paper I was reading which describes isoenergetic expansion of water. As you will be able to see, the results for the isoenergetic expansion were somewhere inbetween isenthalpic and isentropic. I never meant to suggest that the expansion in the vessel was actually isoenergetic, I just thought it may be a half decent way to approximate the pv work term or "isentropic efficiency" that HYSYS asks for.

The aim of my project as a whole can be summarised as follows: When using an isentropic efficiency of 100%, the coldest temperatures are achieved requiring the most costly materials for piping.etc If a lower efficiency is used, this can occasionally result in the use of cheaper materials. I was looking for a method to give a conservative number for the isentropic efficiency.ie Still erring on the side of caution but less than 100% (or the 98% HYSYS recommends). I realise I can jusst guess at a number but I have to be able to back up this figure or I cannot use it.

Finally, when I say I have modelled an isoenergetic expansion I simply mean that I have been able to replicate the water results in this paper using HYSYS, I am not claiming to have done anything fancy or groundbreaking in that department.

 

[PaoloPemi]

when simulating a blow down the isentropic efficiency factor required by your simulator is just a correction for keeping in account the differences from real processes and the standard procedures coded in the software (capable to solve at specified enthalpy, entropy etc.),
different tools (see Winsim etc.) may require different corrections.
With the standard set of flash operations available in simulators (for example my copy of Prode Properties is able to solve the multiphase equilibria with different H,S,V and phase fraction specifications) you can replicate the results of blow down utility in your software, once you have defined the model for the blow down you can decide to introduce a global correction factor as the isentropic efficiency or improve the accuracy of the model with additional code, by the way this is my preferred solution working with Excel.