What s New in ChemSep TM 6.8

Transcription

1 What s New in ChemSep TM 6.8 January 2011 (Updated Marh 2011) Harry Kooijman and Ross Taylor In this doument we identify and desribe the most important new features in ChemSep. 1. New: GUI an diretly load ChemSep olumns from COCO flowsheet files 2. New: One-Clik Tables 3. New: Paked olumn HETP and HTU omparison plots 4. New: Pseudo streams 5. New: Combined feed frations in Stream Tables 6. New: Undo/Redo 7. New: Physial properties 8. New: Lambda plot 9. New: Tray stability alulation and plot 10. Revised: Display while alulating Marh 17,

2 New: Load ChemSep Columns Residing in COCO Flowsheet Files In Version 6.6 we allow diret loading of any ChemSep olumn from COCO flowsheet files! To see this lik on the File menu and selet COCO Flowsheet files from the list of file types: If multiple olumns reside in the FSD files a list of olumn names is shown for the user to selet. Clik on the button orresponding to the desired unit operation to bring up this message: Clik on Yes (assuming that having got this far you really did intend to load the olumn from the fsd file). The olumns are loaded in stand-alone mode i.e. they are saved as sep-files residing outside the flowsheet. This allows the user not only to inspet results but also to solve these olumns - but then using ChemSep internal property routines. 2

3 New: One-Clik Tables ChemSep 6.7 inludes new buttons that allow you to view tables with just one lik. The sreenshot below shows the button bar with some of the newly available buttons (those that are available after the program is installed). From left to right the buttons show: New file, Open file, Save urrent file, Run urrent file, Reload urrently open file, Previous Input/Output Panel, Next input/output panel, Liquid mole fration profiles, Vapor and liquid flow profiles, Temperature profile, Effiieny profile (whih appears here in grey beause this was an equilibrium stage simulation), MCabe-Thiele diagram. Additional symbols an be installed on the button bar representing nearly all of the tables that are defined in ChemSep. To add (or remove) buttons lik on Tools and then on Configure Tools to bring up the window shown below. The lower left half of this panel shows the buttons that have been seleted to appear on the button bar. The list on the lower right half shows available buttons. Buttons an be seleted, removed, and their order hanged using the grey buttons in the lower enter setion. 3

4 New: HETP and HTU omparison plots for paked olumns New to Version 6.7: two new plots that ompare HETP and HTU values. Examples and an explanation follow. P a k e d C olu m n H E T P C o m p a ris o n Stage H E T P (m ) H E T P (C h em S e p ) H E T P (T ra d itio n a l) P a k e d C o lu m n H TU C om p a ris o n Stage H T U (m ) H O G (C h em S e p ) H O G (T rad itio n al) H G (K ey p a ir) H L (K e y p a ir) 4

5 The traditional method used to predit the HETP of a paked olumn is to use the equation below ln HETP= H OG 1 where H OG is the overall Height of a Transfer Unit and is obtained from H OG = H G H L with H G and H L the Heights of Transfer Units of the gas/vapor and liquid phases respetively and =mv / L is the stripping fator. These quantities are defined by: H G =u G /k G a ' H L =u L /k L a' where u G and u V are the superfiial veloities of the respetive phases, k G and k L are the mass transfer oeffiients of the vapor and liquid phases (units m/s), and a ' is the interfaial area density (units m 2 /m 3 ). These equations allow the predition of the HETP from empirial orrelations of mass transfer oeffiients. Examples of suh orrelations inlude the well known orrelations, of Onda et al, Bravo, Roha & Fair, and of Billet & Shultes (and there are many more). The all important seond equation that allows the alulation of the overall Height of a Transfer Unit is a rearrangement of the well known additivity of resistanes formula: G 1 = 1 t K OG k G where m is the slope of the equilibrium line and where t is the molar density. The formula above for the addition of the inverse mass transfer oeffiients and the heights of transfer units apply only to binary systems; for multiomponent mixtures the equations are more ompliated and involve matrix generalizations. Worse, for systems of more than two omponents there is no unique way to define an overall mass transfer oeffiient (see pages 150 and 151 of Taylor & Krishna, 1993). For these reasons the multiomponent forms of these equations are rarely used; instead, an approximate value for the HETP is found by using the binary equations above with the properties of the key omponents. It is these approximate values for the HETPs and HTUs that are shown as HETP (Traditional), HOG (Traditional), HG (Key pair), and HL (Key pair). Also shown in these plots are the HETP (ChemSep) and the HOG (ChemSep). The latter quantity is alulated from the following equations: t L H OG =Z / where Z is the height of paking and is the Baur effiieny (for whih a separate plot is available) and is defined as: m k L y i, L 2 i =1 = i =1 y i 2 The Baur effiieny is the ratio of the length of the atual omposition profile to the length of the profile for a orresponding equilibrium stage. For a mixture in whih all the numbers (and heights) of transfer units are the same and for whih the liquid phase resistane is negligible it an be shown that the Baur effiieny is given by: =N OG =N G 5

6 New: Pseudo Streams New in Version 6.7 are so-alled pseudo streams. These virtual produt streams may represent any internal stream and, when used in a CAPE OPEN flowsheet environment, an be onneted to other unit operations. However, they are not real produt streams in that they are not inluded in the olumn material balane. To define a pseudo produt stream add one (or more) sidestreams to the olumn: and then define the sidestream: Note that the name appears on the flow diagram must first be entered as the stream name on the sidestream panel. The phase must be seleted as either Pseudo Liquid or Pseudo Vapour. To obtain a pseudo produt with the same flow rate as the atual internal stream selet Flow ratio as the stream speifiation type and type in 1.0 on the line below as shown above. 6

7 Pseudo streams flows and onditions are summarized on the stream table but, as noted above, are not inluded in the olumn material balanes: 7

8 New: Combined feed frations in output tables Now inluded in the stream tables are the ombined feed frations in eah stream. The ombined feed fration of any ompound i n stream j is defined as: Molar flow of ompound i in stream j Combined Feed Fration of ompound i in stream j= Molar flow of ompound i in all feeds This quantity helps one see where a partiular ompound leaves a olumn. 8

9 New: Undo/Redo New in Version 6.8 is a omplete Undo/Redo faility. Any hange to input an be reversed up to the point at whih the file was last saved. It is, perhaps, easier to demonstrate this feature rather than try to explain how it works. Consider this input sreen from a file modeling a omplex distillation olumn. We will now hange the flow rates of feed 1. The numbers in the illustration below are purely for illustrative purposes and have no signifiane in the problem at hand. 9

10 We now deide that it was a mistake to have hanged these flows and we want to reover the old values. To do so we use Ctrl-Z one to get: Notie that the flow rate of n-hexane in Feed 1 has reverted to the old value of 0.5 lbmol/h. Use Ctrl-Z a seond time: and a third: 10

11 Using Ctrl-Z two more times and we will reover the original set of omponent feed flows. Now, suppose that we deide that the first of the feed flow hanges was, in fat, orret. If he use Ctrl-Y we will reverse the last Undo as shown below: Undo and Redo an be used to reverse (and re-reverse) any hange to the input inluding, for example, speifiations as shown below. Before: Change to: Reverse the hange with Ctrl-Z (twie beause the first use enters the default value): Redo the hange with Ctrl-Y (twie): 11

12 Undo and Redo are also available from the Edit menu: New: Physial Property Models In a famous paper published in 1972 Soave presented a modifiation of the Redlih-Kwong Equation of State: where P= RT V b a V V b a=a T a T = RT 2 A P A = B = = T r b= B RT P The Peng-Robinson Equation of State (EOS) is a modifiation of the onepts pioneered by Soave with a view towards improving the estimates of liquid density provided by the basi EOS. The Peng-Robinson EOS is: P= RT V b a V V b b V b with the parameters given by the same equations as for the SRK EOS with the following differenes: = T r A = B = For mixtures the parameters are obtained from the following simple mixing rules : a= i =1 x i x j a i, j j=1 12

13 b= x i b i i=1 where k i, j a i, j = a i a j 1 k i, j is the binary interation parameter for the i-j pair of ompounds. Volume Translation for Liquid Density It has long been reognized that the liquid density predited by the SRK and PR equations of state are inadequate for proess engineering work (this despite the fat that both models are apable of exellent preditions of vapor-liquid phase equilibrium). The idea of orreting the liquid phase densities was first proposed long ago. In the so-alled volume translation methods the volume is alulated as follows: V =V EOS EOS where V EOS is the volume omputed from the original equation of state (i.e. the unmodified SRK or PR models) is the volume translation parameter for that partiular EOS and ompound. Several methods of and EOS alulating the volume shift parameter have been proposed. ChemSep 6.8 inludes the pioneering method of Peneloux and a generi approah. Peneloux proposed the following orrelations for the volume shift parameter: PR = RT z P Ra SRK = R T z P Ra where z Ra is the Rakett parameter, values of whih have been published for many ompounds. In ase the Rakett parameter is not known, it an be estimated from orrelations or set equal to the ritial ompressibility to whih it is often very nearly equal. ChemSep also inludes a generi approah to volume translation in whih we ompute the shift parameter from: EOS =V V EOS where V is the pure omponent molar volume. The pure omponent volume needed for this alulation may be obtained from data or an be estimated from an alternative method for prediting pure ompound volumes. In ChemSep the pure omponent density is estimated for eah ompound at a redued temperature of 0.7 from the method seleted to estimate pure omponent liquid densities. The default hoie (when no seletion is made) is an aurate orrelation of pure omponent density data. For mixtures the volume translation parameter is alulated from: = i =1 x i i 13

14 The Rakett Models for Liquid Density The well known Rakett method as modified by Spener and Danner is as follows: V sat = RT P F z 1 1 T F z =z r 2/7 Ra The Rakett equation has been extended for mixtures as follows: V sat =R i=1 T i x i F P z i In order to alulate the funtion F z we must selet a method of estimating the ritial temperature of the mixture and the mixture Rakett parameter. Sine there are several ways to estimate mixture ritial properties there are, neessarily, several ways to use the Rakett equation for mixtures (all of whih give the same results when applied to pure ompounds). Three different mixing rules now are available in ChemSep: those in the DIPPR handbook (whih is the version that was available in earlier versions of ChemSep), the mixing rules of Li et al. as given in PGL5 (page 5.23), and the mixing rules of Chueh and Prausnitz (also as reported on the same page in PGL5). The mixing rules of Chueh and Prausnitz are: Z Ram = x i Z Rai i=1 where T m = i=1 j=1 i j T i, j i = V i x j V j j=1 T i, j = 1 i, j T i T j V i V j 1 i, j =8 V i V j 1/ 3 The mixing rules of Li et al. are similar to those of Chueh and Prausnitz but with: and the DIPPR Handbook suggests using: Liquid Visosity T m = i=1 i T i T m = x i T i i=1 ChemSep 6.8 inludes a new method for mixture liquid visosity: The Letsou-Stiel method for mixtures. This method used the same equations as the Letsou-Stiel method for pure ompounds (whih has always been available in ChemSep) but uses the mixing rules of Li et al disussed above to get the pseudo-ritial properties. 14

15 New: Lambda plot New to Version 6.8: The Lambda plot. An example and an explanation follow. L a m b d a C h e m S e p 2 0 Stage L a m d b a L a m b d a S trip p in g fa t o rs C h e m S e p 2 0 Stage S trip p in g fa t o r (S i = K i. V / L ) P ro p a n e Is o b u ta n e n -B u t a n e 1 -b u te n e Is o b u te n e Tra n s -2 -b u t e n e N e o p e n t a n e Is o p e n t a n e n -P e n t a n e 15

16 The parameter known as Lamdba appears in many different distillation alulations and is defined as follows: = m V L where m is the (loal) slope of the equilibrium line. For multiomponent systems the slope of the equilibrium line depends on the hoie of the key omponents (and on whether or not the omponents are lumped together or not. For more disussion of the latter point we refer readers to the disussion of the MCabe-Thiele diagram in What's New in ChemSep 6.5. It will be immediately apparent that the Lambda parameter bears a strong relationship to the stripping fators illustrated for the same separation in the seond plot above and defined by: S i = K i V L The only differene between these two definitions is that the equilibrium slope is replaed by the K-value. There is, therefore, a stripping fator for eah omponent. For more on the differenes between stripping fators and Lambda see Distillation Design (H.Z. Kister, MGraw-Hill). 16

17 New: Tray Stability Plot ChemSep 6.8 has a new version of the Operation Limits plot. An example for a large tray olumn appears below. O p e ra tio n lim its 5 C h e m S e p 1 0 Stage O p e ra tio n F ra tio n o f flo o d S y s te m F a to r S y s te m L im it F F Je t F F B a k u p F F D C V e l. F F D C re s. tim e F F D C h o k in g F F W e ir L o a d F F W e e p F a to r S ta b ility TD M in.h liq TD S e a lin g TD In a paper entitled Tray Stability at Low Vapor Load by D. Summers, L. Spiegel, and E. Kolesniko (Pro. Distillation and Absorption 2010, p611, I. Chem. E., 2010) defined a tray stability fator by: = P dry h l The signifiane of this parameter is that for a tray with a stability fator greater than 0.6 has less than 30% weepiing. It is onsidered that the adverse effet of weeping on tray effiieny will be small if weeping is less than 30%. 17

18 We have, therefore defined the Stability TD as: TD Stability =0.7 W F were W F is the weep fator and is given in the plot shown above. We have used the more onservative value of 0.7 rather than the value of 0.6 of Summers et al. in defining this turndown ratio. With this definition weeping should not be a problem if this turndown ratio is less than 1. In the example shown above the Stability TD is always less than 1, showing that the tray effiieny is not adversely affeted by weeping. The Sealing TD is defined by: where d l is the downomer learane. TD Seal =W F d l h l The Minimum liquid height turndown ratio is defined by: ={ 2d W hole F h TD l hl,min 2d W bub,max F h l if 2 d hole d bub if 2 d hole d bub,max} with the maximum bubble size set to 30 mm. 18

19 Revised: Display While Calulating The information displayed while atually arrying out a simulation has been improved. Some of this is simply textual hanges to the words printed to the sreen, but there is one hange that will be more notieable. The number displayed to indiate the urrent error (the extent to whih the model equations are not yet solved) has been revised as shown in the sreen image below: The number that represents the error is defined as follows: Error=log Norm log Tolerane where Norm is the square root of the sum of the squares of the errors in eah separate model equation and Tolerane is the (user defined) value to whih Norm is ompared to determine onvergene. It an be seen that last number shown in the image above is negative demonstrating that the simulation has onverged to within the speified tolerane. 19