Reactions in HYSYS

Download as pdf or txt
Download as pdf or txt
You are on page 1of 7

Reactions in HYSYS

There are currently five different types of reaction that may be simulated in HYSYS and a number of reactor types that they may be used with (and one special reactor that does not require any equations). Reactions may also be used in Columns and Separators (though there are some limitations on the phases that may be used by the reactions in those cases). The five reaction types are as follows:

Conversion
This reaction type does not require any thermodynamic knowledge. You must input the stoichiometry and the conversion of the basis reactant. The specified conversion cannot exceed 100%. The reaction will proceed until either the specified conversion has been reached or a limiting reagent has been exhausted. Conversion reactions may not be grouped with any other form of reaction in a reaction set. However, they may be grouped with other conversion reactions and ranked to operate either sequentially or simultaneously. Lowest ranking occurs first (may start with either 0 or 1). Just as with single reactions, simultaneous reactions cannot total over 100% conversion of the same basis. Conversion reactions cannot be used with Plug Flow Reactors or CSTRs. In general, they should only be used in Conversion Reactors.

Equilibrium
Equilibrium reactions require that you know some sort of relation between the reaction's equilibrium constant, Keq, and temperature. You may specify Keq in a number of ways: As a constant. Enter either Keq or Ln(Keq ) As a function of Temperature. You specify A-D in the equation below Ln(Keq ) = A + B/T + C*Ln(T) + D*T
IMPORTANT: No matter what the units in your preferences, the T's used for this equation are in Kelvin and thus your coefficients A-D must be adjusted accordingly if the T's for the information you have are not also in Kelvin. If it becomes complicated to do so, you might choose to generate a table of K vs. T and enter that as tabular data.

As tabular data of Keq vs. T (as suggested in the note above and from which HYSYS fits the above equation) Have HYSYS determine Keq from the Ideal Gas Gibbs Free Energy Coefficients. This is similar to, but not exactly like what you get by attaching any equilibrium reaction to a Gibbs Reactor (which just takes the stoichiometry). The difference depends on the property package because the Gibbs reactor will take into account any non-ideal behavior predicted by a thermo package such as Peng-Robinson. An essentially ideal thermo package like Antoine would give almost exactly the same results for the two different methods. You may also search for the reaction among the pre-defined reactions in the HYSYS library (reached from the Library Page of the Equilibrium Reaction window) Supposedly, like Conversion reactions, equilibrium reactions may be calculated either sequentially or simultaneously. I actually did not see any means by which reactions might be designated sequential and suspect an error in the reference manuals in stating that it could be so. Equilibrium reactions also cannot be used with Plug Flow Reactors or CSTRs. In general, they should only be used in Equilibrium Reactors or General Reactors. They can, however, also be used in the special Gibbs Reactor. When a reaction set is attached to a Gibbs reactor, the stoichiometry involved in the

reactions is used in its calculations.

Kinetic
All three of the remaining reaction types can be considered kinetic, in that they deal with an expression for the rate of the reaction. Differentiating between the three becomes simply a matter of formulation. In this first and simplest form, the rate equation is the one to the left (this picture is taken from the Parameters Page of the Kinetics Reaction window). The first term on the right hand side refers to the forward reaction, the second term to the optional backward reaction. The k's are the reaction constants for which you must enter on the Parameters Page the activation energies, E and E', and the pre-exponential factors, A and A' (which are basically all of the constants lumped out front). The basis functions are not just functions of the Base Component (which you set on the Basis Page -- see Chapter 11 of RV1 for an explanation of the Base Component or anything else having to do with reactions), but are the products of the concentrations (or partial pressures, etc.) of any of the reactants or products to whatever power (negative numbers and decimals are fine). For example, it just so happens that for the reaction CO + Cl2 --> COCl2 the rate law might be rCO = k[CO][Cl2]3/2 You actually enter the form of the basis functions on the first page, Stoichiometry, of the reaction window. In the columns to the right of the one in which you enter the stoichiometric coefficients, you must enter the forward and reverse order. The HYSYS default is to assume an elementary reaction the stoichiometry parallels the order. Therefore for this reaction you would leave the forward order of CO at 1 and enter 1.5 for the forward order of Cl2. Though it is assumed there is no reverse reaction, you might, if you chose, leave a 1 for the reverse order of the COCl2. As long as you did not enter a value for the reverse E and A, no reverse reaction would take place. The Chemicals Tutorial in the Tutorials Book will take you through an example of the use of a Kinetics reaction in a CSTR.

Kinetic (Rev Eqm)


This form of the rate equation is fairly similar to the standard kinetic form. The difference is that instead of getting information about the reverse rate constant, we use the relation: Keq = kforward /kreverse or, as is actually substituted into the standard form, kreverse = kforward /Keq Of course, in doing so we have implicitly assumed an elementary reaction as that is inherent in the definition of Keq . Therefore there is no place to enter reaction orders. Keq is determined by HYSYS in the same it was done for the equilibrium reaction above, except that this time you are forced to enter the data in the Ln(K') format (again, if that's not what you have, don't despair. Try creating an equilibrium reaction, defining it how you are able and then copy over the constants that HYSYS generates). For an example of the use of this type of reaction, see my Plug Flow Example.

Langmuir-Hinshelwood
This is the most complicated of all the reaction forms and is therefore the one that is not even mentioned in any of the manuals (either on-line or off). Therefore it falls to me to explain it. I go through all of this in detail in the Plug Flow Example, so if you decide you do need to use this form, I recommend you work through that. Langmuir-Hinshelwood is mainly used to model heterogeneous catalysis. The rate of reaction is slowed when you have a finite number of active sites on the catalyst, some of which may become blocked to reaction by the products being formed. Hence, to the standard rate equation is added a denominator (this is almost exactly like the form for enzyme catalysis for those of you who have had Biochemistry).

General Information on Reactions


If you are looking for a step by step instruction on every stage of the creation of a reaction and the use of reaction sets, you won't find it here. You will find explanations of that stuff in any of my reactor examples (Plug Flow Example, Gibbs and Equilibrium Reactors, Conversion Reactor), or you may look in Chapter 11 of Reference Volume 1 (does an excellent job with this kind of stuff), or work through the Chemicals Tutorial in the Tutorials Book. What I am going to list here are the little pearls of wisdom I picked up while working through the reactions myself and that may or may not be mentioned in the manuals. 1. You do not need to go back to the basis environment every time you want to edit reactions. You can create or change reactions in the simulation environment using the Reactions Package under the Flowsheet menu. You cannot, however, import or export reactions except from the basis environment. 2. Modifications can be made to reactions on a specific reactor's property view pages that are local only and do not apply globally. Local changes always take precedence over the global settings. (Not an option in PFR). See Section 13.13 of RV2. 3. The components you have in a reaction need not have been previously included in the fluid packages' component list. When you finally associate a reaction set with the fluid package, whichever of its reaction's components are not already there will be added then to the fluid package. 4. There are places (such as when entering the kinetics parameters) where you shall find that no matter how many significant digits you enter in, only 2 sig. figs will be displayed. This is a short coming of HYSYS and will be fixed in future versions. The extra unseen digits are used in the calculations, but the only way to see them is to export them to a Spreadsheet (using the secondary mouse button, simply drag the number from the reaction window to a cell of the Spreadsheet Page of the Spreadsheet). 5. Under the Stoichiometric Page of a reaction window, the item called "Reaction Heat" is not the Heat of Reaction as we know it. In fact, they have opposite signs (an exothermic reaction has a positive Reaction Heat). 6. Remember whenever defining an Equilibrium constant using Ln(Keq ) = A + B/T + C*Ln(T) + D*T, that T is in Kelvin. 7. HYSYS does not show you the units of the Activation Energy it displays. If you would like to display it in

the same units you entered, you must go into your preferences set and set the units of Molar Enthalpy to the units you would like Activation Energy displayed in, as Activation Energy does not have its own category. Reaction rate does have its own category. The Pre-exponential factor has the same units as Reaction rate (the units set on the basis page, NOT necessarily the units set in the preferences which seem to only be for reporting purposes) divided by the Basis Units.

Reactors
Though I plan to tell you about all of the reactors, pay special attention to the info on the Plug Flow and Gibbs Reactors, as they are the ones you will likely use the most. With the exception of the Plug Flow Reactor, the property views of the various reactors are essentially the same. You must specify a liquid and/or vapour product stream for all reactors (again, except for plug flow which has one and only one product stream). All reactors are assumed adiabatic until an energy stream is attached (at which point some further specification such as outlet temp or energy input is required).With the exception of the Plug Flow Reactor and the Gibbs Reactor, all of the reactors also have a number of pages in their property views related to the reactions in them. There you may make changes to the reactions that only take effect in that particular reactor. Local changes always take priority over global changes (so make sure you are aware of whether the change you are making is local or global). Appearance in PFD / Description Object Palette Button The Plug Flow Reactor can be used with Kinetics, Kinetics (Rev. Eqm.), or Langmuir-Hinshelwood reactions (any number and combination of the three types can be used in the reaction set). An excellent description of each of the PFR's inputs can be found in Section 13.10 of RV 2. You can also get a good idea of the way to go about setting up a PFR in your simulation by working through my Plug Flow Example. Quirks of the PFR: HYSYS "integrates" over the length of the reactor by dividing it into a number of sub-volumes (like a series of CSTRS). The default is 20 sub-volumes. The most noticeble effect of this to the user is in the reactor profiles. The various characteristics are given as values vs reactor length. The lengths listed are the midpoimts of the subvolumes. For example, a 10 m length reactor with 20 subdivisions would give profiles starting at .25 m and incrementing by .5 m (the size of a subdivision), finishing with 9.75 m. Not realizing this at first I was irritated that the profiles were not showing me the entering and exiting values (0 and 10 m). Not to worry, the .25 and 9.75 m values are, in fact, the same as 0 and 10 m. Everything within a subvolume is the same (like a CSTR). Another quirk of PFR is that on very rare occasions (see the note in the gibbs reactor example). The integration mechanism gets confused. To fix the problem, try looking at the Reactions page of the reactor property view. Under "Initialize segment reactions from:" make sure the re-init radio button is chosen. Normally, you would not pick this option as it takes the longest time to run. For details on how HYSYS handles the three options see Section 13.10 of RV2 (p. 464). The CSTR can be used with Kinetics, Kinetics (Rev. Eqm.), or LangmuirHinshelwood reactions (any number and combination of the three types can be used in the reaction set). An excellent description of each of the CSTR's inputs can be found in Section 13.13.3 of RV 2. You can also get a good idea of the way to go about setting up a CSTR in your simulation by working through the Chemicals Tutorial in the Tutorials Book. In addition you might want to take a look at my Case Studies Example,

where I build on the tutorial by adding a case study. I use the Spreadsheet feature to access the actual conversion % of the CSTR in the tutorial. Quirks of the CSTR: CSTR is primarily for liquid reactions, of course, but HYSYS will conduct the gaseous reactions as well. The less the "liquid" volume, the more of the total volume available for the vapour phase reactions (i.e. HYSYS uses the total volume minus the volume you set for the liquid to calculate the volume of the gas, whether or not any liquid is actually present in the stream).

The Gibbs Reactor (like the one in Aspen) is unique among the reactors in that you are not required to enter a reaction set for it to work. The Gibbs reactor works by finding the equilibrium state with the lowest Gibbs Free Energy. It appears to be akin to finding all the possible equilibrium reactions and allowing them all to equilibrate. It's nice because you do not need to know anything about the individual equilibrium constants. On the Composition page you can set the production of components or set any of them to be inert. You may also set, on the Reactions page, the Gibbs reactor to behave like an equilibrium reactor (you must then attach an equilibrium reaction set, also see quirks below), or like a separator (no reaction). See the example which will teach you about the Gibbs reactor, the Equilibrium Reactor, and my "switch" technique. The "switch" is also mentioned on my tools and tricks page. Quirks of the Gibbs Reactor: There is something very important to note when attaching equilibrium reactions to the Gibbs reactor. The Gibbs reactor takes only the stoichiometry of the attached reactions and applies its own free energy minimization technique to it. Only components listed as reacting in the reaction set undergo any reaction. Note that HYSYS will not allow you to attach a reaction set which would include all of the possible independent reactions as that would simply duplicate the effect of setting the reactor to full Gibbs reactions. The part of this that is important to you in the design classes is that the results of the Gibbs calculations come extremely close to the values obtained in the equilibrium reactor using correct data, while not making use of any data on Keq . Thus if you need to simulate a reactor in which you want certain reactions equilibrated, but not others (for instance, because a certain catalyst is employed allowing those particular reactions to equilibrate quickly, but not aiding any other reaction) and yet have no or untrustworthy data on the equilibrium constants, you are better off using the Gibbs reactor set on "Specify Equilibrium Reaction" than using the Equilibrium Reactor. Two postscripts on this: 1) If you use the above technique, it doesn't matter what you set the equilibrium constant, a fixed constant of 1 is fine, because you only need the stoichiometry. 2) Before using the above technique, check the equilibrium reaction library. The one you need may already be there, in the temperature range you desire (even then, however, the Gibbs way may still be better). One last note, there appears to be a minor bug in HYSYS, in that, when operating the Gibbs Reactor in Equilibrium Reactor mode, a button appears that would show you the % conversion, reaction extent, etc. Unfortunately, even when the Gibbs Reactor had completed its calculations, the matrix remained blank.

opens

then,

If you would like to experiment with the similarities and differences between the Gibbs reactor and the Equilibrium Reactor yourself, see the example for a good way to go about it.

opens

The Equilibrium Reactor uses reaction sets with only, surprise, equilibrium reactions in it. You can read more about it in Section 13.13.4 of RV 2. You can also see the example in which I compare it to a Gibbs Reactor. In general, I recommend making use of the Gibbs Reactor over the Equilibrium Reactor.

then,

The Conversion Reactor deals with, yep, you guessed it, conversion reactions. You use it when you know how much of the reactants will be converted into products. As mentioned in the section on conversion reactions, it can handle multiple reactions which may be ranked to occur simultaneously or sequentially. Reactions with the same ranking are simultaneous and the total conversion of the same reactant can not exceed 100% (all subject to limiting reagents, of course). The product of one reaction can be the reactant of another reaction. Quirks of the Conversion Reactor: Though the specified conversion cannot exceed 100%, the actual conversion can. This is because the actual conversion is the percentage conversion over the original amount of base component present. However, if that base component is the product of a lower ranked (meaning reacts first) reaction, there may be more available than was originally there. This allows the actual conversion to exceed the specified conversion (it's still behaving correctly, so don't panic). The conversion could thus be much greater than 100%. You can see this in my conversion example. If none of the base component was initially present, the actual conversion field will remain blank. Conversely, if the base component of a reaction is a reactant in an earlier occuring reaction, or if there are limiting reagents, the actual conversion will be less than the specified conversion. If the reaction producing the base component of another reaction is ranked equally with it, the second reaction would not convert any of the component produced by the first reaction, but would only convert the specified percentage of the original amount present. Ranking cannot be changed locally in the reactor. It may only be altered in the reaction set on the global level.

opens

then,

The General Reactor is like a combination of the CSTR and the Equilibrium Reactor. If you put in all kinetic style reactions, it acts like a CSTR. If you put in all equilibrium reactions, it acts like an Equilibrium Reactor. The General Reactor, unlike any other type of reactor available to you, will also allow you to mix equation types. You can combine kinetic and equilibrium reactions into one reaction set (you still cannot combine conversion reactions with any other type, though you can attach a conversion reaction set to the general reactor as well). That set will then be a mixed type and can be attached to the general reactor. Unfortunately, verifying the accuracy and method is somewhat involved (though not necessarily difficult). I leave it to you to investigate this. Rather than create an example for you. This time I simply added the General Reactor to Not Accessible from the my file comparing all the other reactors. That way, you may use all the reactions and Object Palette. Can only reaction sets that were created for the other examples as a way of exploring the results be reached via the Unit that the General Reactor gives you. The name of the file is AllReactors.hsc and it is Ops View (obtainable by located under \\Hartsook\Hysys\SAMP403. pressing <F12>, "Add Quirks of the General Reactor: A strange problem, somehow tied into the number an Operation" under the of product streams attached, results if you try to attach an equilibrium set immediately Flowsheet Menu, or after a conversion set. It fails to find a solution even if it had done so before. If from the Unit Ops Page something like this happens to you click off then back on again the Act as a Separator of the Workbook. when cannot solve button and it should solve. The same quirks of the CSTR apply when using any kinetic reactions in your reaction set. Return to the HYSYS Page | Return to the CENG 403 homepage

You might also like