ENCH 445: Separation Processes

     Professor Frey's Separation Processes WebBook

 


 

Chapter 5 (Part 3): Introduction to ChemSep and COCO.

 

The simplest way to obtain ChemSep-LITE is to download it directly from the ChemSep website described in the links given below. After downloading and opening the latest version of ChemSep-LITE, the user interface looks like the following:

The goal when setting up a problem is to convert all the red x's in the node tree (circled in red also above) into green check marks. The first node is the title node and completing it is optional. But it is good practice to give your file a name and add some comments so you don't forget what it does. The procedure for completing the rest of the file will be illustrated here for a single-stage flash calculation that solves the same problem as given in one of the links to part 1 of this chapter where a spreadsheet was employed for a single-stage flash calculation. More specifically, the following problem is solved in this tutorial:

  • A subcooled liquid feed composed of n-pentane, n-hexane, n-heptane, benzene, and toluene at 298 K and 5 atm enters a flash chamber whose pressure is 2 atm. The flow rate of the feed is 1 mole/s. A heat exchanger present in the flash chamber provides 20 kJ of heat per mole of feed. Determine the composition of the vapor and liquid product streams, the vapor product flow rate, and the temperature in the flash chamber. Use the virial equation of state with the Pitzer and Curl correlation for determining the virial coefficients (or the best available approximation to this equation of state) for determining the vapor phase activity coefficients and regular solution theory for determining the liquid-phase activity coefficients. Assume ideal mixtures when determining the enthalpies of the liquid and vapor phases.

The following screen shots are from version 6.96 of ChemSep since later versions have less clear fonts and text.

The first thing you should do is to go to the Units node and select a convenient set of units. The default set of units can be used but it is usually convenient to change the pressure units to atmospheres, the temperature units to degrees C, the flow to moles/sec, and the enthalpy to J/mol as shown below:

Next, go to the components node and select the components of interest for the problem being considered. You do this by transferring these components from the left window to the right window in the selection portion of the node. You can type a partial name in the find window to help in locating a component. Note that after you select the components, the red x for this node changes to a green check mark to indicate your progress as shown below. Note also that if you are using a version of ChemSep later than version 6.96, your user interface will look slightly different that the interface shown below.

Next, go to the operation node and select the flash button. You need to be cautious in this node and just make one selection and then leave it alone. If you toggle between selections in the operation node, some features of the file (like the degree of freedom analysis) can become corrupted.

 

After the operation node is complete, go to the properties node. Here you should select the Gamma-Phi option in the K-value window. This indicates you will use an equation of state for the vapor phase and an activity coefficient model for the liquid phase. Select the Tsonopolous equation of state since it is most similar to the virial equation of state with the Pitzer and Curl correlation for the virial coefficients used in Problem Set 3 in this WebBook. The a-ij binary interaction parameters for the Tsonopolous equation of state can be left as their default entries, which are the symbol *. This will cause ChemSep to use the UNIFAQ model to estimate the interaction parameters. Then select regular solution for the liquid phase activity coefficient model since again this is the choice in Problem Set 3 in this WebBook. You can then select the Antoine equation for the vapor pressure and ideal for the enthalpy method, which means ideal mixtures are assumed with no heat of mixing. In the Enthalpy/Exergy section, you can leave the default options in place, which will mean that the reference enthalpy is selected as the vapor phase at 298 K and a correlation for heat capacity as a function of temperature stored in the component database will be employed. The physical properties and reactions sub-nodes can be left with their default values in place. When you are done, the user interface will look as shown below:

Next, go to the feed node. The name and stage entries can be left at their default values and the two-phase feed entry box should be left alone. You can select T & P in the state selection box, which means you will select the temperature and pressure of the feed. Then select 5 atm and 25 C for T and P, which will yield subcooled liquid feed in this case. Note that after you type the value of these parameters in the appropriate box, you need to either hit "return" or click on another box to actually input the value. If you want to have a feed that is saturated liquid feed at, say, 5 atm, you can first select pressure and vapor fraction in the state selection box. Then select 5 atm for the pressure and 0.01 (or a similar small number) for the vapor mole fraction. This forces ChemSep to make the feed essentially all saturated liquid at 5 atm pressure (in reality 1 mole-% of the feed will be entering the flash tank as saturated vapor). This small amount of vapor feed will not significantly affect the calculation. This method is a lot easier to do than using the T & P option and trying to determine the temperature of saturated liquid feed at 5 atm. This simple trick for making the feed saturated liquid at a desired pressure also works in ASPEN. Finally, select the molar flow rates in the feed of all the feed components (here it is 0.2 mole/s for each component). When you are done, the user interface will look as shown below where all the lettering is black (notice also the progress shown by the check marks on the node tree):

Next, go to the analysis subnode of the specifications node. As shown below, this subnode indicates there are two degrees of freedom left if all the degrees of freedom associated with the feed are fixed. This agrees with the Description Rule result described in Part 1 of this chapter.

 

The following screen shots are from version 7.01 of ChemSep since earlier versions contain some bugs that are fixed with this version.

 

In the flash specs subnode of the specification node, you need to specify two items (which uses the remaining two degrees of freedom) to fully specify the problem. In this example the pressure in the flash tank and the heat transfer in the heat exchanger are specified as shown below:

Before solving the problem, go to the solve options node as shown below. Notice you can limit the step size used in Newton's method to promote numerical stability if needed. Also, you can select numerical methods other than Newton's method. Another useful feature here is the initialization method (the starting guess). In cases where convergence is difficult, you can first solve a simplifed version of the problem under consideration (such as assuming ideal vapor and liquid). Then change initialization from "automatic" to "old result" and solve the actual version of the problem (accounting for nonideal behavior in this example). ChemSep will then use the results from the simplified version of the problem as the starting guess for the actual version of the problem.

Finally, using the main menu, select Solve/Check Input. After confirming you have correctly specified the problem, select Solve/Quick Solve. Your solution will then appear in the result node as shown below. Notice in this problem V/F = 0.215 and T = 95 C (or 368 K) are obtained. This is in good agreement with the Excel spreadsheet version of this problem described in part 1 of this chapter where V/F = 0.26 and T = 369 K were obtained.

 

Links of interest:
  • ChemSep-LITE.   This free version of ChemSep, called ChemSep-LITE, simulates multicomponent distillation, absorption, stripping, liquid-liquid extraction, and single-stage flash processes and was developed by Ross Taylor at Clarkson University and Harry Kooijman at Shell Global Solutions International. Be sure to also look at the rest of the ChemSep website, which contains lots of interesting information. ChemSep is a CAPE-OPEN (Computer Aided Process Engineering Open Source) compliant software program and is part of the COCO (CAPE OPEN to CAPE OPEN) suite of software described in the next link.
  •  

  • COCO.   COCO (CAPE OPEN to CAPE OPEN) is a free suite of computer aided process engineering open source software programs that includes ChemSep as described in the previous link. For maximum software stability, you should uninstall any previous versions of ChemSep on your computer before you install COCO so you have only one copy of ChemSep on your computer. COCO is an example of the newest generation of software for computer-aided process engineering that consists of "plug and play" components that are primarily open source and free.
  •  

  • COCO: The power of component based flowsheeting.   This material, which comes from a presentation by Jasper van Baten and Richard Baur, describes the history of COCO from its beginning as a software component testing platform to its current status as a full (and free) chemical process simulation platform. Aslo included is an introduction to COFE (Cape Open Flowsheet Environment).
  •  

  • Using ChemSep, COCO and other modeling tools for versatility in custom process modeling.   This material, which was presented at the 2010 AIChE meeting in Salt Lake City by Jasper van Baten, Ross Taylor, and Harry Kooijman, includes a description of how to create an Excel spreadsheet model of a process unit that can be incorporated into a COFE flowsheet.