12 replies to this topic

[mitoman]

Dear....

 

My work is to calculate thermodynamic properties for mixture of methane and n-octane using Lee-Kesler-Plocker equations and compare them to GERG calculations in Excel. So far I've found DWSIM, which does not work and I also have full version of REFPROP, but it doesn't contain LKP. 

 

No, I've found free editon of Prode, which is perfect because it can run in Excel (like GERG). 

 

http://www.prode.com/en/download.htm

 

I don't get it work properly. I'm using Win 10 and Office 2013 on 64-bit computer (tried also in VIrtualBox with Win 8 and Office 2013). After the installation of both files ppp.exe and proind.exe (password: "friend") and manageing add-in, I don't see "Properties" button and can't find in "Customize the Ribbon" and "All tabs" etc. Everything seems to work fine, but I cannot connect to Prode program. Is it possible to use it in free version?

 

Every time I run Excel, the add-in of Prode disappears and I have to locate it again. 

 

 

 

If this doesn't work would you suggest me any other free program for LKP calculations?

 

@PaoloPemi

[PaoloPemi]

the student version has some limits on max number of components (10) etc. but other ways works as complete version...

 

I have Windows 10 64 and Excel 2013 64,

Prode Properties works Ok but I think previous Excel 2010 64 was more stable....

if you need to go with Excel 2013 the 32 bit version offers (in my opinion) some additional stability (at cost of educed performance) , see also the comments (2013 32 vs 64) in Microsoft Technical forums

 

to install log in as admin,

note that for Excel 64 versions you need to install the 64 bit version of Prode Properties

while the 32 bit version is required for Excel 32 versions

 

the distribution of Prode Properties includes several  Excel pages,

you can open these pages in Excel and calculate phase envelope, distillation column etc.

 

to install the additional menu you need to load  properties.xla file

in Excel 2013 from add-in dialog  (see the instruction in operating manual),

 

if, after installing properties.xla, you do not see the additional menu

there could be some conflicts with other applications,

the solution could be to install Prode before other add ins or edit properties.xla file

 

Finally, the methods in Prode Properties thermodynamic library are accessible

as macros in excel,

you only need to instruct Excel about the exported macros in ppp.dll,

load an Excel application example provided with the software

then write,

= StrGD(1) 

to obtain the gas density of stream 1

 

in the same way you can activate Prode Properties editor,

= edSS()

 

etc....

 

corrected the name of method edSS()

Edited by PaoloPemi, 08 June 2016 - 06:22 AM.

[mitoman]

Thank you! I figured out reading some other forums, that you might not see the add-ins tab if you have disabled Analysis Toolpak. Strange....

 

Now I have other problem. I recorded a video of me using this program. Mouse pointer is not seen in the video and it's a bit annoying. You can see me pointing out Excel is 64-bit and so is my computer. In Excel I shoved "Properties" add-in, enabled macros and trying to connect with Prode. While installing Prode (both files) I've chosen 64-bit version both times. I'm 100% sure, cause I did that many times.

 

It's not working. When I change anything (and save), Excel stop working.

 

https://youtu.be/YB5TKoYm3r4

 

Change the quality of a video under setting in right bottom corner.

 

I'll try with Office 2010. Maybe it will work....

Edited by mitoman, 08 June 2016 - 02:38 PM.

[PaoloPemi]

it is stable on my computer, you may wish to update your Excel 2013 64 with latest service pack (should be SP2, not sure),

anyway my feeling is that Excel 2010 64 is much more stable (Excel 2010 64 rarely stops while Excel 2013 64 restarts regularly if you install some antivirus protection etc.) if you can switch to Excel 2010 64 that will work without those problems, otherwise you may install Excel 2013 32 ,

you may find useful also this link

 

https://technet.micr...y/ee681792.aspx

 

keep us informed about results with Excel 2010 64 or Excel 2013 32

Edited by PaoloPemi, 09 June 2016 - 02:15 AM.

[mitoman]

I've found Office 2010 64-bit which works normally. Finally...*fingers crossed*

 

I'm not really familiar with all the settings, but I've managed to calculate graph very similar to GERG (bottom graph). So, it's possible. Don't know why, but my settings weren't saved, even though I clicked every save button.

 

 

 

As I said previous my student work is only to calculate LKP equations using critical values and compare them to GERG (I've done this). Could you be so kind and comment my settings? Is it possible to change values of Y to MPa. I haven't found it under "units".

 

I get this:

 

 

My settings are:

 

If you'd be so kind and tell me what am I doing wrong. I can keep changing those settings until forever. I also don't understand this saving method. I always save everything and the last save is under "stream/operating", but some settings stay the same others don't.

Edited by mitoman, 09 June 2016 - 04:32 PM.

[PaoloPemi]

from Prode Properties Editor, when I modify something in a stream I only click on Save button on first page of editor , then Ok to exit dialog, no other actions are required,

 

the same for units of measurement, you can set the units then Ok to exit dialog

you can change all units of measurement from editor or by code (see operating manual for details)

 

before to close Excel to save all settings there are the archive files,

see the operating manual about how to use the archive files and specifically def.ppp which is the default archive that Prode Properties loads when starting (but you can create different files for different projects and load them when required).

 

it seems you wish to calculate the phase diagram for a binary mixture (C1-C8), is that correct ?

the Excel page provided with Prode Properties allows to calculate T,P (equilibrium point) plus H, S, V  (at T,P)

 

in your example (C1 05 C8 0.5) Prode calculates the dew line but for some reasons the bubble line is not calculated,

have you tested different molar fractions, say C1 0.8 C8 0.2 etc.)  or different vapor/liquid fractions (the phase fract. field in Excel page) instead of 1.0 ?

 

Prode solves dew and bubble lines with specified tolerance (by default 1.0E-10) and this may originate numerical problems with mixtures as C1-C8 or heavier HC,

commercial version automatically redefines starting points and that helps when solving difficult cases  but also with Student version you may adopt some trick to calculate phase diagram,

for example you can calculate a line with, say,  specified fraction of 0.999 or 0.001to mitigate default settings (line with specified fraction 1.0E-10)

if that doesn't work another possibility is to solve a siothermal flash at specified P and different T to identify bubble (incipient vapor) or dew (incipient liquid) points

 

in your example, given that units are Pa, K the calculated values seems comparable with those produced by GERG, in Prode you may set units to Bar or other units if required,

 

Prode Properties (commercial) includes GERG 2008 (standard)  while the student version has only ISO 18453 (GERG 2004) which is based on a modified Peng Robinson model (while GERG 2008 is based on a more complex model),

 

note that VLE predictions with GERG 2008 (standard) in some cases may give errors larger than an optimized  EOS (see below),

see some discussion here

http://www.sciencedi...021961416000161

 

unfortunately GERG models have predefined parameters and there is no possibility to do data regression from experimental VLE data points as for other models

 

before to generate VLE with LKP you should calculate BIPs, LKP may require some large values of BIPs to fit C1-C8  VLE data,

for VLE / LLE / SLE data regression you may utilize the specific utility in Prode Properties,

 

for VLE of natural gas mixtures I routinely adopt the extended Peng Robinson in Prode Properties with NRTL for excess Gibbs and Huron Vidal mixing rules,

that model switches automatically from  NRTL to Peng Robinson with Aij = 0 giving a simple way to deal with mixtures of near ideal and non ideal chemicals, you may wish to contact Prode for details

 

in your case (binary mixture) you may evaluate also other models based on EOS + excess Gibbs,

note that these models may not be applicable to the whole T,P range as std. EOS

Edited by PaoloPemi, 10 June 2016 - 02:19 AM.

[mitoman]

This saving thing does not work for me. I'll give it another shot.

 

Don't know what is with those bubble points and LKP, but it doesn't work. Previously tried in DWSIM and I'm trying now with Simulis Thermodynamics and it's the same. 

 

Back to Properties. Experimenting with different molar fractions and different vapor/liquid fractions didn't give me desired results. It was better, but the line was not connected. 0,5 mol is used because it's the same value as in Gerg. Molar fractures being used for calculating are completely my choice. 

 

I'll try those options you mentioned...

[PaoloPemi]

in Prode Properties editor savings must work, otherwise you weren't able to define the C1-C8 composition as you did,

 

probably you mean saving to a file, with Prode you can save to archive files as mentioned in my previous post.

 

I have not my copy of Prode Properties available and I am unable to replicate your GERG diagram for C1 0.5 C8 0.5 with the GERG 2008 model in Prode, I can do that next week,

 

however iI would guess that your diagram shows only a dew line and not a bubble line, 

starting from initial lowest pressure and up to cricondentherm (the max Temp in VLE area, about 570 K in your diagram)  the values seem comparable with those calculated by Prode Properties with LKP (but note that you are comparing GERG 2008 with predefined BIPs against LKP with BIps = 0, as said a much better test would be to calculate BIPs for LKP with the data regression utility in Prode before to generate the diagram)

 

the absence of bubble line (or a lincomplete bubble line) may depend from several causes as discussed in my previous post,

some mixtures show LLE equilibria etc. and you need the commercial version of Prode Properties to calculate those phase diagrams (i.e. LLE * VLE),

 

as said, you can solve a series of flash operations to determine the extents of  VL area (or VLL)

this is another possibility, solving a flash operation you get also the fractions of the different phases (VL, VLL etc.) predicted by selected model

Edited by PaoloPemi, 10 June 2016 - 08:56 AM.

[staffel]

my experience with GERG model is that phase diagrams may be problematic,

I prefer the Extended Peng Robinson PRX in Prode Properties

for your C1 0.5  C8 0.5 mixture , PRX with BIPs = 0 predicts CricoT 544 K

lower than GERG,

however you can adjust Kij to replicate the data points

calculated by GERG

Attached Files

•  C1_C8_PRX.jpg   98.41KB   4 downloads

Edited by staffel, 11 June 2016 - 03:02 PM.

[mitoman]

I've already calculated phase envelope with some other Gerg program, bt would be nice, if I get them with Properties. Is this way I would have both methods calculated with the same program. But what I need the most is LKP phase envelope. Cannot use any different method, because my work is to compare Gerg to LKP.

 

The problem is everything is new to me, those VLLE, BIP etc. and I really don't know what I'm doing. I'll try of course....I hope it will work eventually.

 

Did anyone might tried calculate LKP with Properties and get good results?

[DanW]

Hi mitoman,

 

I'll try to explain what you're doing, what it seems what you were asked to do:

 

- Create a Pressure-Temperature phase diagram of C1-C8. This diagram typically shows a confined region where you'll have vapor and liquid in equilibrium for all temperature and pressure conditions inside it. To the right of it, only vapor can exist. To the left, you'll find only liquid, depending on the mixture. For the C1-C8 mixture the diagram is a little weird, because the bubble pressure (pressure at which the liquid begins to evaporate as you lower it) goes to infinity. This means that an all-liquid mixture is pratically impossible to obtain at lower pressures and temperatures, which is what one would normally expect to get.

 

- To build a phase diagram, you'll have to calculate dew and bubble lines. Both lines will cross each other at the critical point, where the liquid has the same properties as the vapor phase, that is, you cannot tell who's who because they'll have the same density and "look" the same. Dew points are calculated by considering your mixture all-vapor and finding a pressure at a specified temperature, or vice-versa, where liquid begins to form. Bubble points are calculated by considering your mixture all-liquid and finding the conditions where vapor begins to form. These conditions are fulfilled when the compounds have the same fugacity in the liquid and vapor phases. In order to check that, you'll need fugacity coefficients.

 

- You were asked to compare LKP with GERG probably because they're very different in nature, but they will ultimately give you the fugacity coefficients you'll need to calculate the equilibrium points. This is how you calculate fugacity cofficients in the LKP model:

 

 

The LKP model calculates averaged properties for mixtures, and here is where the interaction parameter goes. It exists to adjust model-calculated properties so they match real-world data. If the model was perfect, no interaction parameter would be needed.

 

 

 

And this is how GERG does:

 

 

I guess this is why you won't find a free or open-source GERG alternative out there. It is just too complex for anyone to code and give it for free... not worth the effort.

 

I can't really comment about GERG because I`ve never worked with it, but for LKP you'll get very different results if you use a default value of 1 for the interaction parameter between C1 and C8. The value of 1.423 is in the original paper and must be the one that gives the best results when compared to experimental data.

 

If you manage to get DWSIM working with Excel, I can show you how to generate a phase diagram with IP = 1 for the LKP model just for comparison purposes.

Edited by DanW, 21 June 2016 - 06:33 PM.

[staffel]

Mitoman,

I use regularly Properties (mainly with Office 2016  64),

the phase envelope utility in Prode allows to calculate VL, VLL plus SV and SL lines,

I believe that student version, the free version you are utilizing) allows only VL + VLL lines, 

I know there are some limits in free version (see the above post of Paolo)  

but you should be able to obtain similar results as other software adopting the same models,

may be you are entering some different parameters or values.

As said in my previous post i prefer the PRX model for these mixtures (see the resulting phase diagram in my previous post),

anyway there are many different models available in Properties, student version and you can compare.

If you don't require VLL lines (only VL) goto Editor and in models tab set Multiphase Equilibria to No Multiphase, only two phases, in the same way one can require VLL, SL etc. lines

For BIPs Prode includes several databases (VLE, LLE,  SLE,, Hydrate)  and you can set the binary interaction pairs required for your application, 

there is also a procedure to regress BIPs from VLE-LLE-SLE points.

Edited by staffel, 23 June 2016 - 07:22 AM.

[PaoloPemi]

Mitoman,

why do you wish to obtain that strange shape for C1-nC8 VLE phase diagram ?

The bubble line (VL or VLL) should start from a low pressure,

make sure to extract the correct root and do phase stability tests on each calculated point,

In Prode Properties Lee Kesler Plocker model  includes contribute of BIPs as

Kij = 1.0+BIP

then you must subtract 1.0 from values given in original paper,

you may use  0.55 for C1-nC8 , howewer better values are available to reduce errors,

to test follow this procedure,

1) download and install the software from prode.com (if not installed)

2) run Microsoft Excel (I have Excel 2013 64), load properties.xla and phasenv.xls (see the manual for details)

3) from Properties menu in Excel start Properties Editor

4) in components dialog include C1 and nC8

 

5) in models dialog select Lee Kesler Plocker

 

6) in BIPs dialog insert 0.55 as interaction coefficient between components 1 and 2

 

7) in operating dialog click on button Save to store the data for this stream

 

click on Ok button to exit Properties Editor and, finally, in Excel

click on Compute Phase Diagram button to generate the phase diagram

 

as you see with BIP = 0.55 you obtain values not too different from the Peng Robinson model

reported by Staffel, there is about 10% differences in P,T values at critical point

but this is acceptable considering the differences between Lee Kesler and Peng-Robinson,

anyway you can do better : changing BIPs to find the optimal value as suggested in my preious post,

Paolo

Edited by PaoloPemi, 11 July 2016 - 11:50 AM.