Using 'real' Cp/Cv Ratios 3

[jcaiken]

Since 2000 HYSYS (and possibly other simulators) has calculated a 'real' Cp/Cv Ratio as its default and this can create potential problems if used inappropriately.

As pressure increases the 'ideal' Cp/Cv Ratio typically reduces from 1.4 towards 1.2 but the 'real' ratio can increase from 1.4 to over 3 for the same scenario.

The use of a 'real' ratio for RV orifice sizing can result in under-sized orifices and valves being fitted.

As well as process sizing calculations for RVs, compressors, etc. this ratio is also used in estimating the reactive loads caused by RV operation and could affect the piping stress analysis.

 

[UmeshMathur]

You are correct on all points. That is why the older published methods can be simplistic, even dangerous (e.g., the current GPSA handbook shows tables using ideal gas Cp values to get to the Cp/Cv ratio for use in compressor calculations). The horsepower may easily be off by quite a bit using an ideal Cp/Cv, especially for high pressure service and/or high compression ratios.

A recent, detailed discussion is available in thread391-151370 for compressors. You may find this quite interesting. For RV and related system sizing, it is again very important not to skimp on the accuracy of the thermodynamic calculations.

 

[jcaiken]

Thanks for the direction to the other thread. The discussion there is very detailed but focused on compression.
I'm concerned more about the impact on sizing relief valves. Using the 'real' ratio in the standard API 520/521 equations for sizing the orifice required for a defined flowrate can lead to the installation of an undersized orifice with potential impact on the plant safety. Unfortunately, an inexperienced engineer extracting data from HYSYS for use in an external program or using HYSYS's Dynamic Depressuring utility to size RVs may not be aware that they have used inappropriate data.

An paper in 'Chemical Engineering' magazine (November 2003 issue) recommended that the 'real' ratio should not be used for RV sizing.

 

[MortenA]

jcaiken

But the "real" k is the _real_ k and the valve should thus not be undersized - but using the ideal will result in an oversized valve and this could cause other problems as have been discussed.

While a undersized valve most likely will be worse than an oversized chattering etc. can also cause problems.

Im therefore not sure that i agree with you. You could take your argument one step further: Why use a Z factor? This also increases the desity and thus lead to a smaller valve size! Why not just use a z factor of one - nah lets ad a mirgin of 10%.....

Best regards Morten

 

[MortenA]

jcaiken

Had to doubecheck this myself and found my old "Smith & Van Ness" themodynamics textbook. Since the formula for calculating the critical pressure ration is based on an ideal gas assumption then im not sure on the implications of calculating k using real gas assumtions.

So i withdraw my first comment (this is however a little difficult in this forum).

Best regards

Morten

 

[Latexman]

jcaiken,

While your concern (sizing relief valves) and the subject of the past thread (sizing compressors) are not the same, they share an identical root problem. That issue which UmeshMathur explained so well about compressors is in my words - should we use the old, simple, conservative, closed form, analytically integrated equations which were derived with many ASSumptions (ideal gas behaviour, constant Cp/Cv, second order terms ignored, etc.), or should we go back to the start of the derivation of these simplified equations (energy balance, momentum balance, continuity equation) and derive a numerical method that is devoid of any assumptions?

Whatever the answer that is "good enough" for a particular problem, what should NOT be done lightly is mix the two, i.e. use "real" Cp/Cv's in the simple methods, without accounting for the differences in the ASSumptions. I believe this is exactly your point, and I agree 100% with you.

It is perfectly fine to use rigorous EOS's that accurately model a fluid in an equally rigorous numerical solution of the fluid dynamics of the relief system. This is UmeshMathur's point, and I agree 100% with him too.

Good luck,
Latexman

 

[EGT01]

For reference, the details of the article mentioned by jcaiken are...
"Using the ideal gas specific heat ratio for relief-valve sizing"; by Aubry Shackelford, Berwanger, Inc.; Chemical Engineering, November 2003. Just my opinion but I think it is a good one to have.

The article was motivated by the differences in the recommended practices of API 520 Part 1 between the 6th edition (1993) and 7th edition (2000)...in particular, using the specific heat ratio at 60F and atmospheric pressure vs using the relieving conditions. But, the article also covers when it is acceptable to use the ideal gas specific heat ratio and when it is more appropriate to the use the isentropic exponent for non-ideal situations.

Both editions of API 520 include an appendix which discusses the basis for the flow equations used in sizing pressure relief valves for gas and vapor and both advise of the limitation to using the ideal gas specific heat ratio. I don't have my old 6th edition handy, but it seems I remember it attempting to give more guidance than what the 7th edition gives for using the isentropic exponent. However, I believe there were problems with the guidance that the 6th edition was giving.

Further information can be found in the Technical Inquiries for the API publication to which API provides a technical interpretation upon request. Here's a link to the Technical Inquiries for the current 520 Part 1, 7th edition...

http://committees.api.org/standards/tech/docs/520ati.xls

As I understand, API does not maintain support for Technical Inquiries for older editions but I happened to make a copy of the Technical Inquiries from the 6th edition that I thought was worth keeping. Here's an excerpt regarding Appendix E which discusses the problems with the guidance given in the 6th edition and suggests another method for determining the isentropic exponent...

http://img223.imageshack.us/img223/2512/apirp520p11993tin6ku.png

You'll need to right click the image link above and "save target as" to be able to view it easily.
.

 

[Ashereng]

MortenA,

If you want to "retract" a post, simply red flag it.

"Do not worry about your problems with mathematics, I assure you mine are far greater."
Albert Einstein
Have you read FAQ731-376 to make the best use of Eng-Tips Forums?

 

[UmeshMathur]

jcaiken:

I believe Latexman has summarized the issue very well, and I am gratified that we invariably are on the same page when it comes to use of theoretically sound and rigorous procedures for process calculations. In this day and age, when process simulation tools are readily available, it seems quite unnecessary and even dangerous to insist on using simplistic procedures that could easily go wrong. Of course, the sophistication of the available software imposes a responsibility on the user to be fully aware of the pros and cons of adopting one option over another. In this instance, choice of the proper thermodynamic equation of state would be a real issue requiring judgement and caution.

Additionally, there is the issue of complying with the established norms and standards of practice for professional engineers: failing to be diligent in performing one's work can lead to very unpleasant consequences in the event that problems develop after the fact.

If there were to be adversarial litigation, the work of those who took unnecessary shortcuts might be subjected to legal attacks too horrid to contemplate. This is a general point that I believe needs to be stressed in many of the threads in these forums. I have been involved as an expert witness in several litigations and, believe me, seeing the attorneys on my side tear apart engineers on the other side for sloppy work left a lasting impression on me.

 

[jcaiken]

UmeshMathur,

Whilst I agree with your sentiments, the standard methodologies used within much of the contracting industry are aimed at reducing the need to apply 'judgement and caution'. Whilst corporate policy dictates the use of internal standard methods, these may not address the use of appropriate data. Even the application of the rigorous EOS models in simulators may be compromised by the re-use of previous models without fully assessing whether their basis is still appropriate for a new project.

The current approach of "just apply the API/ASME/GPSA standards" is not 'wrong' it is just that more rigorous thermodynamic models are becoming available. It is difficult for companies to justify the additional cost of revising their standard methods based on these better models, unless the standards themselves drive the change. The interaction between clients, contractors and vendors may also result in a 'lowest-common-denominator' approach based on ignoring the better models in order to maintain consistency.

I'm sure that the better models will gradually be adopted, initially for applications at the limit of the standards (e.g. high pressure, dense-phase, fire-case relief valve sizing). However, the senior management levels of many companies will not be aware of the potential consequences and therefore may not be prepared to shoulder the costs of making the necessary changes to their standard methods.

In the mean time there appears to be a considerable risk associated with using the 'wrong' data in the 'right' model and vice-versa. This could create a whole new area for the attorneys to attack.

 

[monaco8774]

Interesting discussion, I have come across this problem of which Cp/Cv value to use specifically on blowdown systems, in the past we have validated the blowdown system on some platforms by actually blowing down and timing it. I have done this on two seperate platforms, fairly high volume inventories and up to 360 bar(g). We found both of the calculations were out against reality in some cases design had blowdown in 15 minutes and actual took over 40 minutes. The thing that puzzles me is since a lot of parameters used for flow calculations were worked out empirically (such as valve Cvs and Piping fitting losses) do we have to use the ideal Cp/Cv when calculating the flows on these since the Cv would have been back calculated using ideal Cp/Cv not real Cp/Cv. Any ideas?

 

[Ashereng]

Valve Cv is a measured value, and varies with each valve. There really isn't a "ideal" Cv value for a valve - it is what it is.

"Do not worry about your problems with mathematics, I assure you mine are far greater."
Albert Einstein
Have you read FAQ731-376 to make the best use of Eng-Tips Forums?

 

[Latexman]

It is difficult for companies to justify the additional cost of revising their standard methods based on these better models, unless the standards themselves drive the change.

In a large chemical or petroleum company, I disagree. My company switched from their standard method to internally developed VAX and PC based tools that combine an EOS that has wide application with a robust numerical solution of PSVs and pipe flow (rupture disks are a simple K value now) about 10-15 years ago. We probably average spending $750,000,000 a year on capital projects. By using PVT models better than the ideal gas law and using solutions to PSVs and pipe flow that are devoid of simplifying and conservative assumptions, there is no doubt in my mind that the development costs of these tools have been paid back many, many times. The average relief device is smaller, the inlet and outlet pipe is smaller, vent headers are smaller, effluent treatment systems are smaller, etc.

We use contract engineering firms to size reliefs when we are in the peak of the chemical cycle, like now. Since we are legally responsible for our plants (contractors have limited liability), we train them on our tools and insist they use them.

Good luck,
Latexman

 

[jcaiken]

Unfortunately, I'm currently working for an international engineering contractor company and we cannot justify the expense of updating the basis of our standard in-house engineering methodologies without the driver of working for a client who requires it.

Production clients (such as yourself) are able to balance the additional costs of implementing the improved methods against the Opex and Capex benefits.

I'm sure that these improvements will become part of the appropriate standards over time and therefore increase the potential market to the point it is worth implementing them. In the mean time we engineers will just have to take care to ensure that the basis for the property data is appropriate to the correlations being used.

 

[Latexman]

jcaiken,

Well said! Sorry about the wrong reference on the quote.


Good luck,
Latexman

 

[Ashereng]

Latexman,

As a contractor (EPC, consultants, whatever you want to call us), we always use the tools that our client wants us to us. If, such as your employer, there is an in-house tool that you want us to use, we will of course use it. If you don't have a tool, then we will use our own if it is acceptable.

Not everyone wants or needs a BMW. We are just as happy selling you a Honda, Hyundai, Kia, or bicycle. ;-)

"Do not worry about your problems with mathematics, I assure you mine are far greater."
Albert Einstein
Have you read FAQ731-376 to make the best use of Eng-Tips Forums?

 

[UmeshMathur]

Having spent a number of years both in operating companies and engineering & construction companies, I again revert to the basic theme that the owner's interests should drive design decisions. These interests are best served when their contractors use the most appropriate and reliable engineering methods that are based on proven and reliable advances in the state-of-the-art.

Many industry standards documents are often simplistic, having been developed many years ago - sometimes even before decent PC-based computing tools became commonplace. Today, most large-scale projects require a very high degree of automation to enable numerous software products to communicate effectively with each other. This need is especially vexatious when confronting the lack of CAPE-OPEN compliance on the part of the software vendors. However, failure - or worse refusal - to enable options to handle the latest advances in the name of "productivity" does the client community a grave disservice, in my opinion, and opens the door to serious challenges in case something goes wrong.

At the very least, improving the technology assessement process and making ongoing enhancements should be a primary concern everywhere. I recall working in this area, as part of a large technology department, for a large E&C over 25 years ago. Alas, such groups are now mostly disbanded and many slipshod practices have re-emerged in basic areas like distillation, heat transfer, and process safety engineering.

In my assessment, the ability of the E&Cs to reduce the extent of over-design, and thus improve their competitive position significantly, has resulted from large advances in the automation of the design process and the incorporation of better tools and methods.

Also, in the majority of the reputable E&C companies, the management has generally been quite progressive when engineering staff have requested permission to upgrade existing methods. However, the onus remains on the specialists to alert their management as the situation evolves and to be aggressive in pointing out the technological and legal risk of not keeping abreast.

While we may be drifting from the rather narrow scope of the original question, I still think this issue must continue to receive attention.

 

[jcaiken]

I couldn't agree more.

Perhaps I have a jaundiced view of the capitalist market but too often the 'progressive' attitudes of management suffer when the company becomes publically-quoted on the stock market. A short-term approach becomes the norm, driven by the need to continue past profit growth rates in order to maintain share price.

This often changes the historic 'quality engineering' philosophy behind the original company growth to 'quality product with minimum engineering' and produces a 'use the standards' approach to minimise the need for expensive senior engineers with their years of experience and good judgement. Too often MBAs and Marketing are seen as more important than engineering experience and expertise.

It is interesting to see comments by Trevor Kletz and others that many of the recent fires and explosions are the result of failing to learn from past mistakes.

 

[sethoflagos]

Some very good points, but there are also some interesting certification issues to consider.

In general, the majority of us are only asked to ballpark equipment design. Compressor performance is guaranteed not by us, but by the compressor vendor; PSV rated capacity is certified by the valve manufacturer. It's useful if our preliminary design comes close, but should we expect our methodologies to be equal to or better than the specialist suppliers who guarantee and underwrite the performance of their equipment? Most of us just aren't given the time in this life to do more than a conservative isothermal-ideal gas approximation per ASME, API, BS, Indian Boiler regs or whatever.

 

[jcaiken]

Reliance on the others to guarantee the performance implies that they are considered to have 'better' experience/methodologies/specialist engineers etc. Unfortunately, such companies are no less subject to the same economic constraints and given the competitive nature of the market may actually in a worse position. If you can already sell enough equipment based on 20-30 year old design parameters there is little economic justification for improving the basis/accuracy of the design correlations and capability/experience of the engineers.