jeferson_seibel
Student
- Apr 17, 2022
- 1
[Problem description also attached to this post]
Hello!
I'm facing a problem in UniSim: I need to simulate a vegetable oil hydrotreatment process, but the triglycerides are not available in the basic databank attached to my university's software license. So, I went over the works on literature and found the Zhong method for triglyceride property estimation, which is frequently used in simulations done in Aspen Plus as a custom model. For consulting more about this property estimation method, please, see the links below:
Original article: Article correction:
The methods proposed by the author allow one to calculate properties such as molar volume, vapour pressure, viscosity and heat capacity by considering fragment contribuitions. Then, it would be possible, for every equation of the models, to calculate the constants by adding up the fragment contributions. Once the constants were calculated for a given component (ex.: triplamitin), one could put these in a property model equation and calculate its values over a given temperature range.
I've done this work within MS Excel for tripalmitin, which gave me accurate results - comparing the values reported by the Zhong's work. Nevertheless, I didn't get the same results when I put the model and constants in UniSim. I defined tripalmitin in UniSim as a hypothetical component, defining some component data such as boiling point, critical properties and so on (I got this informations from a NIST databank available through Aspen Plus). I also defined the UNIFAC structure of the component in order to compute de unknown properties. Last, but more important, I defined the T-dependent properties, namely vapour pressure, liquid density, liquid viscosity and liquid heat capacity. Let's take the liquid density to illustrate the process I've done:
In the published work, the molar volume is defined by the following expression, which varies with temperature (T):
, where the volume of a fragment is given by:
As tripalmitin is composed by 3 fragments of palmitic acid and 1 glycerol, the total molar volume (m³/mol) would be calculated as the volume for 1 unit of glycerol and 3 units of palmitic acid. So, we can do some algebraic work to derive an expression:
Having done this work, we can, in thesis, select the model Anto5 in UniSim to put the constants and calculate the property. So, we enter the values calculated from the fragment contribuitions, setting also the minimum and maximum temperature - the authors of the article point 253.15 and 516.15 k as the range in the case.
Calculations with the same model deduced above were made in an Excel spreadsheet over a temperature range of 25 ºC to 400 ºC. This results were in accordance with the ones reported by the authors of the original article. But, when generating a data table in UniSim to get the values of the liquid density property over the sme temperature rande, we get these values:
One may notice that the values calculated by UniSim differ slightly from the ones obtained in Excel calculations (red line with blue circles). Even the profile is a little different, with a slightly curved form in Excel data points (in accordance with the Zhong's reported values) and an almost straight line in UniSim. I can't understand why this difference takes place, as the equations and constants are essentially the same in both programs - the values should be exactly the same. This keeps occuring to other T-dependent properties:
- Viscosity:
-Heat capacity:
I'd like to know if somebody else has ever faced some trouble of this nature in UniSim, and if there is a manner of resolving it. Maybe I'm doing something wrong in the process of defining the component inside UniSim, I don't know. Maybe you can share a tutorial or point some way for solving this.
Thank you in advance!
Hello!
I'm facing a problem in UniSim: I need to simulate a vegetable oil hydrotreatment process, but the triglycerides are not available in the basic databank attached to my university's software license. So, I went over the works on literature and found the Zhong method for triglyceride property estimation, which is frequently used in simulations done in Aspen Plus as a custom model. For consulting more about this property estimation method, please, see the links below:
Original article: Article correction:
The methods proposed by the author allow one to calculate properties such as molar volume, vapour pressure, viscosity and heat capacity by considering fragment contribuitions. Then, it would be possible, for every equation of the models, to calculate the constants by adding up the fragment contributions. Once the constants were calculated for a given component (ex.: triplamitin), one could put these in a property model equation and calculate its values over a given temperature range.
I've done this work within MS Excel for tripalmitin, which gave me accurate results - comparing the values reported by the Zhong's work. Nevertheless, I didn't get the same results when I put the model and constants in UniSim. I defined tripalmitin in UniSim as a hypothetical component, defining some component data such as boiling point, critical properties and so on (I got this informations from a NIST databank available through Aspen Plus). I also defined the UNIFAC structure of the component in order to compute de unknown properties. Last, but more important, I defined the T-dependent properties, namely vapour pressure, liquid density, liquid viscosity and liquid heat capacity. Let's take the liquid density to illustrate the process I've done:
In the published work, the molar volume is defined by the following expression, which varies with temperature (T):
As tripalmitin is composed by 3 fragments of palmitic acid and 1 glycerol, the total molar volume (m³/mol) would be calculated as the volume for 1 unit of glycerol and 3 units of palmitic acid. So, we can do some algebraic work to derive an expression:
Having done this work, we can, in thesis, select the model Anto5 in UniSim to put the constants and calculate the property. So, we enter the values calculated from the fragment contribuitions, setting also the minimum and maximum temperature - the authors of the article point 253.15 and 516.15 k as the range in the case.
Calculations with the same model deduced above were made in an Excel spreadsheet over a temperature range of 25 ºC to 400 ºC. This results were in accordance with the ones reported by the authors of the original article. But, when generating a data table in UniSim to get the values of the liquid density property over the sme temperature rande, we get these values:
One may notice that the values calculated by UniSim differ slightly from the ones obtained in Excel calculations (red line with blue circles). Even the profile is a little different, with a slightly curved form in Excel data points (in accordance with the Zhong's reported values) and an almost straight line in UniSim. I can't understand why this difference takes place, as the equations and constants are essentially the same in both programs - the values should be exactly the same. This keeps occuring to other T-dependent properties:
- Viscosity:
-Heat capacity:
I'd like to know if somebody else has ever faced some trouble of this nature in UniSim, and if there is a manner of resolving it. Maybe I'm doing something wrong in the process of defining the component inside UniSim, I don't know. Maybe you can share a tutorial or point some way for solving this.
Thank you in advance!
