VPPD Lab -The Chemical Product Simulator

Krist V. Gernaey, Jakob K. Huusom and Rafiqul Gani (Eds.), 12th International Symposium on Process Systems Engineering and 25th European Symposium on Computer Aided Process Engineering. 31 May – 4 June 2015, Copenhagen, Denmark © 2015 Elsevier B.V. All rights reserved.

VPPD Lab -The Chemical Product Simulator Sawitree Kalakula, Rehan Hussainb, Nimir Elbashirb and Rafiqul Gania a

Department of Chemical and Biochemical Engineering, Søltofts Plads, Building 229, Technical University of Denmark, DK-2800 Lyngby, Denmark b Department of chemical Engineering, 162 Texas A&M Engineering building, Texas A&M University at Qatar, 23874 Doha, Qatar

Abstract In this paper, the development of a systematic model-based framework for product design, implemented in the new product design software called VPPD-Lab is presented. This framework employs its in-house knowledge-based system to design and evaluate chemical products. The built-in libraries of product performance models and productchemical property models are used to evaluate different classes of product. The product classes are single molecular structure chemicals (lipids, solvents, aroma, etc.), blended products (gasoline, jet-fuels, lubricants, etc.), and emulsified product (hand wash, detergent, etc.). It has interface to identify workflow/data-flow for the inter-related activities between knowledge-based system and model-based calculation procedures to systematically, efficiently and robustly solve various types of product design-analysis problems. The application of the software is highlighted for the case study of tailor made design of jet-fuels. VPPD-Lab works in the same way as a typical process simulator. It enhances the future development of chemical product design. Keywords: Chemical product design, Blended product, Jet-fuels.

1. Introduction In chemical product design, it is not only important to find the chemical product that exhibits certain desirable properties but improving the product performance and making products more versatile have become growing concerns in recent times. Chemical products can be classified in terms of molecule and mixtures. Several frameworks have been developed to design basic chemical/functional chemical products (Smith et al., 2009, and Hill, 2004). However, for more complex chemical products such as cosmetic creams, films, and detergents, no generic methodologies exist. While product design is still based on trial and error experimental approaches, it is now generally accepted that application of model-based methodologies can help to design/improve products so as to reach the market faster by reducing costly and time-consuming experiments (Gani, 2004). That is, experiments are only performed during the last stage as a verification tool. Since thousands of products involving mixtures/blends of chemicals need to be designed and/or evaluated, a huge amount of data on physico-chemical properties of chemicals and their mixtures and/or models that can reliably predict them are needed so that model-data based methods for tailor-made design of products can be developed and routinely applied. This is a challenging task requiring data acquisition, data testing, model development and model-based design method development, etc., that needs to be integrated in a computer-aided framework so that chemicals based products can be designed, analyzed and verified in a fast, efficient and systematic manner (Conte et al., 2012, Mattei et al., 2014). These motivate the development of a systematic framework for product design and evaluation.

1416

S. Kalakul et al.

In this study, the systematic framework employs its in-house knowledge-based system to (1) identify target properties of a desired product; (2) provide suitable property models to calculate the necessary properties classified as primary, secondary, functional, and mixture properties; (3) formulate and solve the design-analysis problems in a fast, robust and systematic manner; and (4) guide the user to the final experimental verification tasks. The framework has built-in libraries of product performance models and product-chemical property models. The built-in knowledge base has a suit of databases containing properties of different classes of chemicals (lipids, solvents, aroma, etc.) and design templates for molecular (solvents, refrigerant, etc) and blended products (gasoline, jet-fuels, lubricants, hair spray, etc.). The property models and the database together provide a very wide range of application for the various chemicalsbased products. The large amount of data, models and calculation-algorithms are managed through specially developed ontology, which is implemented as the framework within VPPD-Lab. It has interface to identify workflow/data-flow for the inter-related activities between knowledge-based system and model-based calculation procedures to systematically, efficiently and robustly solve various types of product design-analysis problems. Furthermore, the flexibility of its architecture allows it to be integrated with toolboxes from ICAS (Integrated Computer Aided System) (ICAS Documentation, 2003) such as ProPred (property prediction tool), ModDev and MOT (modeling tools), ProCAMD (hybrid methodology for computer aided molecular design), and process design tools (Kalakul et al, 2014). VPPD-Lab uses template approach for specific type of chemicals-based products. The current version has design templates for microcapsule, active ingredients uptake in plant, homogeneous formulated products, emulsified products, blended products, single chemical-based products (solvents, refrigerants, polymers), and, an option to create templates for new products. The developed computer aided framework is generic and applicable to a wide range of problems. The VPPD-Lab works in the same way as a typical process simulator.

2. Framework The framework is divided into 2 problems (as shown in Fig. 1); (1) product design, and (2) product evaluation. Each problem follows instruction steps which is integrated with VPPD-Lab knowledge base containing the links for the tools (as shown in Fig. 2). 2.1. Product design Step A1: Define problem Tool 1 starts with selection of sub problem (molecular or mixture design), number and type of phases, and available template/ new template corresponding to the user’s product design problem. Templates are stored in VPPD-Lab KB. Each template guides the user to identify product needs and translate needs to target properties and property values using Tool 1 Problem definition Step A2: Go to template Tool 2: Template guides the user to follow the work-flow for the product design problem such as solvent design, blend design, and emulsion design. Each task in the work-flow is connected with the necessary toolboxes (Figure 2). Tool 3: Chemical database store the experimental property values and needed property values calculated using Property model library. Therefore, it can calculate 55 pure component properties and 10 functional properties for a very wide range of organic chemicals. In addition, it

VPPD Lab -The Chemical Product Simulator

1417

has data for more than 24,000 compounds and contain more than 50,000 entries of various types of mixture data and several thermodynamic models to calculate mixture properties. Tool 4: Routine/Algorithm contains mixture calculation models (linear and non- linear mixing models), thermodynamic models (UNIFAC, NRTL, UNIQUAC, and PC-SAFT). Step A3: Go to product evaluation See 2.2. Product evaluation Step A4: Further develop selected product-process After evaluation of the performances of product in step A3, the performances of the process need to be evaluated. The framework has options to link with the process synthesis and evaluation tools such as PROCAFD (computer aided flowsheet design), SustainPro (process sustainability analysis tool), ECON (economic analysis tool), and LCSoft (Life cycle assessment tool). Problem

Product Design Step A1: Define problem 1.1 Select sub problem 1.2 Select number of pahse 1.3 Select/generate template 1.4 Select Need 1.5 Set target property

Step

Product Evaluation Step B1: Define problem 1.1 Select sub problem 1.2 Select number of pahse Step B2: Select compound/ mixture

Step A2: Go to template Step A3: Go to product evaluation

VPPD-Lab KB (Database/ Algoritm/workflow/tools)

Step B3: Select thermodynamic models

Step A4: Further develope selected product-process

Step B4: Verification

Tool1: Problem Definition

Tool 3: Chemical database and property model

Tool 2: Template Tool 4: Routine/Algorithm Tool 5: Product evaluation

Integrated tools

Figure 1.VPPD-Lab Framework 2.2. Product evaluation Step B1: Problem definition The selected product’s performances need to be evaluated. Experimental verification toolbox helps user to design the experimental tests that may be needed in addition to the molecular and blended products. Product evaluation model toolbox contains product specific performance models (uptake of active ingredient for pesticide, and active ingredient delivery on surface) to evaluate their performances. Depending on the type of the product, these can be model-based, experimental based or a combination of both. Step B2: Select compound/mixture This step connects with the chemical database. The user selects the product that can be a single compound or mixture. Step B3: Select thermodynamic model This step the thermodynamic model that suitable for the product is selected. Note that, the primary, secondary, and functional properties needed for product performance

1418

S. Kalakul et al.

calculations are retrieved from chemical database using the suggested property models by VPPD-KB. Step B4: Verification This step calculates product performances or behaviors of the product. Tool 2: Template/Work flow

Tool 1: Problem Definition

Sub problem

Molecular Design

1 phase

New template

Need

Tas k 1 Tas k 2 Tas k 3 Tas k 4

Translate needs to targets properties

Specify target property values

> 1 phase

Formulation

P1, N1

T1, P1, N1

...

...

P1, Ni

Tj, P1, N1

Structural constraints

AIs / MiIs identification

AIs / MiIs identification

Pure component property constraints

Retrive chemical database/ property models

Mixture property constraints

Formulate MINLP/NLP

Select needs

Tool 3: Chemical database Blends

Target property

New template

Tas k 5

Select number of phase

Mixture Design

1 phase

Solvent design

Select subproblem

Emulsions New template

Antioxidant

Surfactant

Polymer

Aroma

Solvents

Lubricant

Emulsions

Lipids

Color compounds

Gasoline

Common chemicals

Property model library Structural constraints

AIs / MiIs identification

Retrive chemical database/ property models

Pure component property constraints

Retrive chemical database/ property models

Solvent mixture design

Stability check

Mixture property constraints

Pure component property constraints

Additive identification

Formulate MINLP

Miscibility constraint

Miscibility constraint

Formulate MINLP/NLP

Formulate MINLP/NLP

Primary property

Secondary property

Functional property

Mixture property

Tool 4: Routine/Algorithm Mixture calculations ...

Stability check

Thermodynamic calculations

Tool 5: Product evaluation Experimental verification Product evaluation model Integrated Tool ProCAMD: Computer-aided molecular design tool ProPred: Property prediction tool MOT: Modeling tool MATLAB PROCAFD

SustainPro: Process sustainability analysis tool

Computer-aided flowsheet design tool

ECON: Economics analysis tool

LCSoft: Life cycle assessment tool

Figure 2.The structure of VPPD-Lab knowledge base (VPPD-Lab KB)

3. Case study The application of the software is highlighted through the tailor-made fuel blends of jetfuels. 3.1. Step A1: Define problem The problem is product design, the sub-problem is blends. The user needs for jet-fuels in terms of target properties, and target property values are identified as shown in Table 1. 3.2. Step A2: Go to template VPPD-Lab KB guides the user to follow the work-flow of blend design template starting from Task 1 to Task 5.

VPPD Lab -The Chemical Product Simulator

1419

Table 1.Product needs and translated target properties Need

Target Property

Target Value

Ability to be burned

Reid vapor pressure (RVP)

RVP > 1 kPa at 310.95 K

Engine efficiency

Higher heating value (HHV) Dynamic visocosity (Ș

HHV > 6125 kJ/mol Ș< 6.8 cP at 253.15 K

Consistency of fuel flow

Melting point (Tm)

775 <ȡ< 840 kg/m3 at 288.15

Flammability

Flash point (Tf)

Tf >310.95 K

Stability

Gibbs energy of mixing (¨*mix) Toxicity (LC50) Carbon footprint (CF)

¨*mix< 0 െ݈‫ܥܮ݃݋‬ହ଴ (݉‫݈݋‬/‫ < )ܮ‬3.6

Enviromental impacts

CF < 1.8 CO2 eq.

Task 1: AIs/MiIs identification The case study of binary mixtures is highlighted where active ingredient (AIs) is the mixture of conventional jetA-1 which is considered as 1 component and The Additives (Minor ingredient or MiIs) are selected from 221 chemicals from different molecular types such as alkanes, cycloalkanes, aromatics, olefins, alcohols with low carbon numbers, ethers, ketones, acid and furan derivatives. Task 2: Retrieve chemical databases/property models Primary properties related to target property calculations; molecular weight (0Z  critical temperature (7F  critical pressure 3F  acentric factor Ȧ constant of the Modified Rackett equation (ZRA /&50, heat of combustion ǻ+c  521 and Wt2are retrieved from the chemical databases. The functional properties of pure chemicals; ȡ Dynamic viscosity ȝ  and vapor pressure (Pvap are calculated. Task 3: Stability check The UNIFAC-LLE group representations of MIs and MiIs are retrieved to calculate the liquid miscibility at 298.15 K (ambient temperature). The stability test is performed using the STABILITY tool (Conte et al., 2011) to check the miscibility of the blends. The blends that are totally immiscible are rejected.50 binary mixtures were found to be partially miscible with jet-fuels. 7DVNIRUPXODWH0,1/3 The tailor-made blend design problems of jet fuels are formulated as MINLP problem. The objective is to minimize fuel consumption subject to product stability and target properties. The results are the formulation of blended jet-fuels and calculated target properties as shown in Figure 3. 3.3. Step A3: Go to product evaluation The blended jet-fuels from Task 4 need to satisfy the desired target properties and aviation standards. Experimental verification toolbox in Tool 5: Product evaluation suggests to perform experimental tests to verify Ș7P7Iȡ/+9593. Furthermore, GLVWLOODWLRQSURILOHVDQG-)727¨3DW‫ל‬C will be tested in order to ensure that the final blends meet the aviation fuel standards based on these properties. 3.4. Step A4: Further develop selected product-process The process to produce the jet-fuel does not consider in this work since mixing of chemicals in order to obtain blends is a routine mixing operation.

1420

S. Kalakul et al.

Figure 3.Results for jet-fuels blend problem from VPPD Lab

4. Conclusion A computer-aided framework for design of chemical products has been developed and implemented into VPPD-Lab. Knowledge base system has been employed to manage data-flow, making it more flexible and capable of solving a wide range of product design problems. The use of the framework has been highlighted through the case study involving a tailor-made design of jet-fuels blend. The results show that adding additive improves jet-fuel properties and reduce the consumption of conventional gasoline. VPPD Lab as the new product simulator that works in the same way as a typical process simulator, it can guide a user to design product, study product behaviors, systematically formulate and robustly solve various type product design problems. Future work is to integrate the software with the process design software in order to systematically solve product-process design problems.

References B. Smith, M.G. Ierapepritou, 2009, Framework for consumer-integrated optimal product design planning, Expert System with Applications, 35, 1-2, 338-349. M. Hill, 2004, Product and process design for structrued products, AICHE Journal, 50, 8, 16561661. R. Gani, 2004, Chemical product design: challenges and opportunities, Computers and Chemical Engineering, 28, 2441-2457. M. Mattei, N. A. Yunus, S. Kalakul, G. M. Kontogeorgis, J. M. Woodley, K. V. Gernaey, and R. Gani, 2014, The Virtual Product-Process Design Laboratory for Structured Chemical Product Design and Analysis, Computer Aided Chemical Engineering, 33, 61-66. ICAS Documentation, Internal report, 2003, CAPEC, KT-DTU, Lyngby, Denmark. S. Kalakul, P. Malakul, K. Siemanond, R. Gani, 2014, Integration of life cycle assessment software with tools for economic and sustainablity analyses and process simulation for sustainable process design, Journal of Cleaner Production, 71, 98-109. E. Conte, and R. Gani, 2011, Chemical-Based Formulation Design: Virtual Experimentations, Computer Aided Chemical Engineering, 29, 1588-1592. E. Conte, R. Gani, Y. S. Cheng, and K. M. Ng, 2012, Design of Formulated Products:Experimental Component. AIChE Journal, 58, 173-189.