From Molecule to Lab to Plant: Computer Aided Process Design

Academics: Prof Alastair Florence (Strathclyde), Dr Blair Johnston (Strathclyde)
Researchers: Dr Cameron Brown (Strathclyde), Dr Tomas Harrington (Cambridge), Bruce Wareham (Strathclyde)

Computer Aided Process Design (CAPD) and simulation tools have been successfully implemented in the chemical and oil industries since the early 60s to accelerate development and optimise the design and operation of process. In recent years the use of process models in the pharmaceutical industry has increased dramatically and covers all aspects of the discovery and development processes - with the expectation of similar benefits as those seen in the chemical and oil industries. Through its current members and UK Research Partnership Investment Fund (RPIF) CMAC has recently expanded its CAPD capabilities in a number of areas in addition to the existing capabilities of Structure prediction (CCDC – Mercury and BIOVIA - Materials Studio) and crystallizer modelling (PSE - gCrystal).

Prediction of fluid properties:

Conductor like Screening Model for Real Solvents (COSMO-RS) is an ab initio quantum chemistry and thermodynamic method, and software, originally developed by Eckbert and Klamt. Using only the molecular structures of solvents and solutes, the model can be used to estimate activity coefficients required for solvent screening and solubility prediction. Models can also be used to construct multi-component phase diagrams and to predict a range of other physicochemical properties. COSMO-RS model data will be fused with data from other computational techniques in data workflows which underpin operational and control workflows allowing for informed decisions to be made before and during any “wet” laboratory processes. COSMO-RS is also being used to construct a CMAC solvent database to allow for simple, virtual solvent screening by all researchers involved in the project.

Related projects: Bruce Wareham and Dr Blair Johnston: Automated workflows for rational solvate and solvent system selection: in silico models for solubility predictions and quantitative structure - particle attribute relationships from molecular simulation and knowledge-based methods.

capdL2.jpg

Multiphysics simulations:

COMSOL Multiphysics allows for the modelling and simulation of complex systems through the layering and interaction of multiple physics interfaces:

  • Computational fluid dynamics. Enables the modelling of most aspects of fluid flow, including: compressible, non-isothermal, non-Newtonian, two-phase and porous media flows. All in laminar and turbulent flow regimes

  • Mixers. Allows for simulations with rotating machinery acting upon fluids. This includes laminar and turbulent flow, incompressible and weakly compressible flow, as well as non-Newtonian flow. Simulations can be in 2D and 3D, time-dependent models that account for a full description of the rotation of the impeller, or through using the frozen-rotor approximation. Additional physics interfaces can include terms and equations to describe the effects of temperature, reacting species, and free surfaces on the fluid flow equations

  • Chemical reactions: Define material transport in dilute and concentrated solutions or mixtures through convection, diffusion, and ionic migration of any number of chemical species. These can be connected to definitions of reversible, irreversible, and equilibrium reaction kinetics that can be described by the Arrhenius equation, or any arbitrary rate law, where the effects of concentration and temperature on the kinetics can be included.

  • Heat transfer: Tools to study the mechanisms of heat transfer – conduction, convection, and radiation. Can be readily interfaces other modules such as fluid dynamics and chemical reactions

Currently, COMSOL is being used to develop CFD models for the array of lab and pilot scale crystallizers within CMAC to serve as a comparison of each crystallizers mixing performance.

Related projects: Dr Cameron Brown – CharacterisatiIOn and ModellIng of Crystallisers (COMIC)


Stirred vessel mixing parameters

With stirred vessel being commonplace in the chemical industry numerous models of their mixing patterns have been developed. Visimix turbulent provides an easy interface for calculation of these models based on classical approaches to hydrodynamics by Kolmogoroff, Hinze and Levich. This provides the necessary process parameters for analysis, scale-up and optimization of mixing vessels with a range of impeller types and configurations. Visimix turbulent is presently being used in conjunction with COMSOL multiphysics in developing comparison tools for crystallizer performance.

Related projects: Dr Cameron Brown – CharacterisatiIOn and ModellIng of Crystallisers (COMIC)


Process flowsheeting and simulation

SuperPro Designer is a flowsheeting environment which facilitates the modelling, evaluation and optimization of integrated processes. With the combination of manufacturing and environmental models, SuperPro Designer provides an assessment of end-of-pipe treatment processes, project economics and environmental impact. With over 140 unit operations, equipment sizing and costing and batch scheduling, SuperPro Designer offers the opportunity to take data from current crystallizer modelling tools and extrapolate that to a full production facility. In turn, it can then be used to develop a series of future state supply network design scenarios, based on the emerging process technology options, in order to inform the business case for moving towards continuous processing. Presently, SuperPro Designer is being utilised to develop flowsheets for the existing and potential future production routes for paracetamol as part of the CMAC Phase II integration project.

capdL6.jpg