DDBST User Meeting Oldenburg, Germany

Transcription

1 DDBST User Meeting Oldenburg, Germany DDBST Software & Separation Technology GmbH DDBST GmbH, Marie-Curie-Str. 10, D Oldenburg, FRG

2 2 9/21/ only for personal use DDBST Session 2

3 NEW DEVELOPMENTS IN VERSION 2009 New Developments 2008 to 2009 Presented by: Jürgen Rarey DDBST GmbH brought 2 U by: The DDBST-Team OVERVIEW 20 Years of DDBST GmbH A New Website DDB 2009 DDBSP 2009 Business Education, Training 3 9/21/ only for personal use DDBST Session 3

4 20 YEARS OF DDBST GMBH December Years of DDBST GmbH A New Outfit A Corporate Identity DDBST GmbH DDBST GmbH THE NEW WEBSERVER A New Website 4 9/21/ only for personal use DDBST Session 4

5 THE NEW WEBSERVER A New Website THE NEW WEBSERVER Huge amount of references to scientific papers. Extensive downloadable material (info-material, tutorials, bugfixes and updates, presentation slides, reports, MSc. and PhD. Thesis, demo videos,.). Regular news entries. Personalization via user registration and login. Extended functionality planned (pre-paid): property estimation and calculation download of exp. data from the DDB, -profiles, Multi-language interface (planned) 5 9/21/ only for personal use DDBST Session 5

6 CAN WE WISH US A BIRTHDAY PRESENT? Customer Testimonials would greatly increase the value of our Internet-site! If you could leave us a short testimonial (hand written is sufficient) before you leave? CAN WE WISH US A BIRTHDAY PRESENT? List of Customers.. May we disclose the identity of our customers? 6 9/21/ only for personal use DDBST Session 6

7 Data Sets DDBST User Meeting 2009 Oldenburg 2009 DDB THE GROWTH (DATASETS) DDB Gain 2007 to Gain 2008 to 2009 VLE 27,140 28,165 1, % , % HPV 25,915 26,975 1, % , % ELE 6,350 6, % % LLE 17,330 18,545 1, % , % ACT 48,325 50,850 2, % , % ACM 1,110 1, % % GLE 17,710 18, % % EGLE 1,220 1, % % CRI 1,395 1, % % SLE 19,580 23,675 4, % , % ESLE 19,070 23,805 4, % , % AZD 49,560 50, % % KOW 7,665 7, % % ADS 3,545 3, % % POLY 15,830 16, % % he 18,100 18, % % VE 29,470 32,995 3, % , % CPE 2,245 2, % % PCP 159, ,840 10, % , % total , % % DDB THE GROWTH , ,0 PCP CPE VE he POLY AAE KI AZD ESLE SLE CRI EGLE GLE ACM ACT LLE ELE HPV VLE ACT and AZD: 10 data points taken as 1 data set , ,0,0 Year 140 data sets per working day 7 9/21/ only for personal use DDBST Session 7

8 log (Number of VLE Data Sets) DDBST User Meeting 2009 Oldenburg 2009 GROWTH OF VLE DATA BANK 4,6 4,5 4,4 4,3 4,2 4,1 4,0 3,9 number of VLE data sets in the DDB is doubling every 16.3 years exponential growth 3, year Moore s Law of Vapor-Liquid Equilibria DIFFERENT WAYS TO KEEP THE DDB GROWING Constant further back-evaluation of older sources Integration of unpublished data from chemical industry MSc. and PhD. thesis from well known research groups (several hundred) Evaluation of new non-english scientific publications Digging for Data 8 9/21/ only for personal use DDBST Session 8

9 part of data with greater deviation Ideal Gas Heat Capacity (J/(mol*K)) DDBST User Meeting 2009 Oldenburg 2009 EVALUATING NEW INFORMATION SOURCES To which extent can quantum chemistry provide reliable thermophysical data? 150 Ethanol Gaussian03 B3LYP G6-311G(d,p) experimental data (DDB) Chloroform Gaussian03 B3LYP G6-311G(d,p) experimental data (DDB) Ethane Gaussian03 B3LYP G6-311G(d,p) experimental data (DDB) Temperature (K) EVALUATING NEW INFORMATION SOURCES 1 quantumchemical calculation 0.1 regressed polynomials (DDB) relative deviation in ideal gas heat capacity (%) 9 9/21/ only for personal use DDBST Session 9

10 DDB NEWS GPA (Gas Processors Association) Package officially accepted and available. Data Bank and estimation, regression, visualization by DDBST GmbH GPA reports presentation, 3-phase flash software provided by EPCON Ltd. NEW COMPETITOR (2008) Availability of NIST- TRC Data in Aspen Plus Currently (for the next 5 years) not considered a threat: significantly less data. after first inspection: significantly lower quality. (DDB is intensively used by many customers and within research for the development and test of predictive methods) NIST will use data from journals (ThermoML) and data input by approx. 40 part time students. Huge administrative overhead (2 million lines of code, ). NIST plans to add ALL available data, even technically irrelevant data from organic literature. Financial resources similar to that of DDBST but publicly sponsored within a project of fixed duration. 10 9/21/ only for personal use DDBST Session 10

11 ASPEN PROPERTY PLUS THERMO ENGINE Calculation using Aspen Property Toolkit to verify physical property issues prior to process simulation VLE, HPV 2009 h E, c PE, v E Pro/II-Interface mostly finished. NEW ESTIMATION METHODS IN ARTIST Water Solubility (25 C) by Moller, Rarey, Ramjugernath Water Solubility by mod. UNIFAC (corrected for alkanes, cycloalkanes and alkenes) Octanol-water partition coefficients by UNIFAC, mod. UNIFAC, and COSMO-RS(Ol). Vapor Pressure by Moller, Rarey, Ramjugernath -Profile Estimation for COSMO-RS(Ol) and COSMO-SAC by Mu, Rarey, Gmehling 11 9/21/ only for personal use DDBST Session 11

12 ARTIST Added Sublimation Pressure Calculation by Myrdal/Yalkowsky Rarey/Nannoolal Rarey/Moller NEW REGRESSION EQUATIONS 12 9/21/ only for personal use DDBST Session 12

13 MISCELLANEOUS IMPROVEMENTS AND CHANGES License management has been thoroughly reengineered. (Reprise License Manager (RLM)) Test DDB installation tool. XML-based parameter export for ProSim Simulis software. Search for components with specified predicted properties (preparation for CAMD). DEVELOPMENT OF DDBST GMBH Since the last User-Meeting Some losses due to the difficult economic situation worldwide. But: Licensing of the nearly complete DDB and DDBSP by research groups in Kaiserslautern and Paderborn. Licensing of a large part of DDB and DDBSP by ENEL Several large new sales and updates for existing customers. Many new customer contacts through the Internet-site Large number of potential customers directed to our site via Wikipedia-articles. 13 9/21/ only for personal use DDBST Session 13

14 TRAINING COURSES Thermodynamics for Process Synthesis and Simulation + D E C H E M A - C o u r s e THANK YOU Thank you for attending this talk and please excuse all the things I forgot to mention! You are welcome to a demonstration of the DDB and the DDB software package, the introductory course and special topic workshop tomorrow or a discussion about your special interests! 14 9/21/ only for personal use DDBST Session 14

15 DDBST-SESSION 2009 DDBST-Session 2009 In Depth New Developments and Research Presented by: Jürgen Rarey DDBST GmbH brought 2 U by: The DDBST-Team 1 OVERVIEW Status of last years topics Further Development of g E -Model Parameter Estimator (GC2GE) Solute Activity and Solubility in Water Combinatorial Effects in Alkane Solvents Solute Activity and Solubility in Alkanes Estimation of Surface Tension and Thermal Conductivity Development of an exact Mixing Rule for VTPR Demonstration /21/ only for personal use DDBST Session 15

16 STATUS OF LAST YEARS TOPICS Verifying Physical Property Data Prior to Process Simulation 2008 VLE, HPV 2009 h E, c PE, v E Pro/II-Interface mostly finished. Work on COSMO-RS-Type Models > profiles in version 2009 Group contribution estimation of -profiles in ARTIST 2009 (COSMO-SAC and COSMO-RS(Ol)) Group Contribution Methods for Pure Component Properties Vapor pressure estimation - published Thermal conductivity under construction Surface tension under construction Activity Coefficients at Infinite Dilution in Solvent Families Water solubility and in water available in ARTIST 2009 Alkane solubility finished 3 CURRENT RESEARCH GROUP OF J. RAREY /21/ only for personal use DDBST Session 16

17 FURTHER DEVELOPMENT OF G E -MODEL PARAMETER ESTIMATOR (GC2GE) Brian Satola 5 PROCESS SIMULATION Typical Problem: /21/ only for personal use DDBST Session 17

18 G E -MODEL PARAMETER ESTIMATOR 7 PROCEDURE Process Simulator Calculation of g E -model parameters from predicted -values. DDBSP Generation of activity coefficient values over the whole concentration and large temperature range and regression to all values (initiated by A. ten Kate). Generally very good results but different functional forms of the models sometimes leads to differences /21/ only for personal use DDBST Session 18

19 Frequency DDBST User Meeting 2009 Oldenburg 2009 NRTL-REGRESSION TO MOD. UNIFAC PREDICTION 9 SYSTEMATIC DEVIATIONS FOR MOD. UNIFAC Frequency of RAD of γ calculated with different models, 7123 data points approx. 8.7% SAC TM BP SAC G03 B3LYP SAC G03 BP OL TM BP OL G03 B3LYP OL G03 BP UNIFAC Mod. UNIFAC(Do) /21/ only for personal use DDBST Session 19

20 PROBLEM DESCRIPTION Can the conversion of mod. UNIFAC predictions to g E -model parameters be optimized to yield more reliable results? /21/ only for personal use DDBST Session 20

21 SOLUTE ACTIVITY AND SOLUBILITY IN WATER Bruce Moller 12 MODEL DEVELOPMENT Data used for the model development Source Number of data points Gamma infinite - DDB 269 SLE - DDB 106 SLE - Beilstein 197 SLE - Chemspider 58 Total 630 Typical compounds Cl HO O N + O - O OH F F O HO OH N F O H OH O OH HN Niflumic acid N O O O Bifenox Cl N + O Dinoseb O - O F H Triamcinolone /21/ only for personal use DDBST Session 21

22 RESULTS Experimental and fitted ln 2 at K ( - This work, Mod. UNIFAC (Do) Δ COSMO-RS) multifunctional compounds /21/ only for personal use DDBST Session 22

23 COMBINATORIAL EFFECTS IN ALKANE SOLVENTS Bruce Moller 15 PROBLEM (2008) of high boilers in low boilers is difficult to determine except from LLE if is high from SLE if high boiler is high melting Abundance of of low boilers in high boilers is available (GC, static measurements, ). Most frequent solvents: Water, Hexadecane, Squalane, NMP, Sulfolane,1,2-Ethanediol,[HMIM]- [TF2N], Phthalic acid dinonylester, DMF, NFM, 19,24-Dioctadecyldotetracontane,Tetracosane, Tetraethylene glycol dimethyl ether Can values be safely converted between different solvents inside one solvent family? (e.g. in squalane to in hexane) /21/ only for personal use DDBST Session 23

24 8 8 DDBST User Meeting 2009 Oldenburg 2009 OPTIMIZATION OF THE COMBINATORIAL PART 1.2 UNIFAC 1.2 Mod. UNIFAC 1.0 heptane in n-alkanes number of carbon-atoms heptane in n-alkanes number of carbon-atoms 17 REDUCTION TO A COMMON SOLVENT - HEXANE 18 Combinatorial and residual contribution Solution of groups concept: Residual contributions for a single solute in any alkane solvent at infinite dilution are identical CH 2 CH CH 2 2 CH 2 CH CH 2 2 CH 2 CH 2 CH 2 CH CH 2 CH 2 Solute CH 2 2 CH 2 CH 2 CH 2 CH 2 CH 2 CH CH 2 CH 2 2 CH /21/ only for personal use DDBST Session 24

25 REDUCTION TO A COMMON SOLVENT - HEXANE 19 Expressing this mathematically: Activity coefficient made up of a combinatorial and residual contribution: R, C, 1, hexane 1, hexane 1, hexane R, C, 1, squalane 1, squalane 1, squalane WCCE8 Montréal REDUCTION TO A COMMON SOLVENT - HEXANE 20 Expressing this mathematically: Taking the ratio, the residuals cancel: R, C, 1, hexane 1, hexane 1, hexane R, C, 1, squalane 1, squalane 1, squalane WCCE8 Montréal /21/ only for personal use DDBST Session 25

26 2 DDBST User Meeting 2009 Oldenburg 2009 REDUCTION TO A COMMON SOLVENT - HEXANE 21 Expressing this mathematically: In general form: R, C, 1, hexane 1, hexane 1, hexane 1,hexane R, C, 1,reference 1, squalane 1, squalane 1, reference WCCE8 Montréal DIFFERENT COMBINATORIAL EXPRESSIONS Ethanol (2) in alkanes (1) hexane hexadecane Exp. Data UNIFAC Mod. UNIFAC FV 45 squalane (ref.) The UNIQUAC surface area for the alkane solvent q 1 FV Free Volume /21/ only for personal use DDBST Session 26

27 DIFFERENT COMBINATORIAL EXPRESSIONS Ethylcyclohexane (2) in alkanes (1) Error starts to become noticeable when q 1 /q 2 < pentane decane hexadecane Exp. Data UNIFAC Mod. UNIFAC FV squalane (ref.) q 1 =q q 1 23 NEW EXPRESSION Propose the following correction to the combinatorial: ln ln ln C-Cav, C, Cav, Free-Volume Expression 2 3, 2 * 3 i m i i V ' FV V V FV V V V * i m, i i ln FV FV V V' FV FV V1 V' 1 Cav, V mi, * V i - Liquid molar volume (cm 3 /mol) - Van der Waals volume (cm 3 /mol) Cavitation term empirically modified: q q r r Cav, Cav, exp corrected 1 2 Step function allows an activity coefficient of 1 for pure component /21/ only for personal use DDBST Session 27

28 MODIFYING THE UNIFAC COMBINATORIAL Blindly (no changing of the parameters, only swapping the combinatorial expressions) applying this combinatorial expression to original UNIFAC: Combinatorial Relative Mean Deviation (%) in ln Number of data points Original ,344 This work ,336 Note: Only data for water were excluded due to the combinatorial expression in this work giving unrealistically large values. 25 PERFORMANCE OF THE NEW COMBINATORIAL EXPRESSION O N hexane HN N'-(4-Chlorophenyl)-N,N-dimethyl-urea Cl Exp. Data UNIFAC Mod. UNIFAC FV New Expression hexadecane (ref.) q /21/ only for personal use DDBST Session 28

29 SOLUTE ACTIVITY AND SOLUBILITY IN ALKANES Bruce Moller 27 RESULTS FOR INFINITE DILUTION ACTIVITY COEFFICIENTS IN HEXANE /21/ only for personal use DDBST Session 29

30 ESTIMATION OF SURFACE TENSION AND THERMAL CONDUCTIVITY Eugene Olivier and Onellan Govender /21/ only for personal use DDBST Session 30

31 ESTIMATION OF SURFACE TENSION 38 ESTIMATION OF THERMAL CONDUCTIVITY Relative mean deviation (%) for the different alkane families /21/ only for personal use DDBST Session 31

32 PHYSICAL SIGNIFICANCE OF GC PARAMETERS (SUFACE TENSION ESTIMATION) /21/ only for personal use DDBST Session 32

33 DEVELOPMENT OF AN EXACT MIXING RULE FOR VTPR Alan Foster and Andre Schedemann 41 BASIC IDEA g E -mixing rule for equations of state Works flawlessly at subcritical conditions. How can the translation be applied at supercritical conditions? /21/ only for personal use DDBST Session 33

34 DEMONSTRATION 43 THANK YOU Thank you for attending this talk and please excuse all the things I forgot to mention! You are welcome to a demonstration of the DDB and the DDB software package, the introductory course tomorrow or a discussion about your special interests! /21/ only for personal use DDBST Session 34