|2012-2017||EPSRC Leadership Fellowship|
|2011||Henry E. Armstrong Memorial Lecture of the Society of Chemical Industry
||Philip Leverhulme Prize for Engineering|
||Visiting Professor, Department of Chemistry, Warwick University
||PhD in Chemical Engineering, Princeton University
||MEng Chemical Engineering, Imperial College London|
Solving nonlinear bilinear problems to global optimality is a long standing challenge in optimisation theory. In a recent two-part paper, Xenia Kleniati and Claire Adjiman introduce the Branch-and-Sandwich algorithm, which is guaranteed to solve a broad class of bilevel problems to global optimality. To find out more about the theoretical developments that led to the algorithm and its application to over 30 problems, read the two open-access papers: Part I: theoretical developments and Part II: convergence analysis and numerical results. The work has recently been extended to mixed-integer problems and an even broader class of nonlinear problems, as discussed in this open-access paper in honour of Professor Ignacio Grossmann.
What is the best solvent for a reaction? The rate of a reaction can change by several orders of magnitude depending on the solvent used. Given the large number of possible solvents, how can the "right" one be chosen rationally? We have recently proposed an approach based on integrating quantum mechanical calculations and molecular designed and have shown that this can lead to an increase in reaction rate of 40%. This work was published in Nature Chemistry.See also Nature Chemistry News & Views by Truhlar, C&EN news item, The Chemical Engineer news item.
If you like this work, a postdoctoral position is now open in this area. Applications from outstanding candidates are invited for an exciting project at the interface of chemical engineering and chemistry. Closing date: 19 January 2015.
FIND OUT MORE
For an introduction to the group's work, watch Claire's inaugural lecture entitled "Molecules on best behaviour: The engineering of molecular systems"
Recent advances in Molecular Systems Engineering, from molecules to processes, edited by Claire Adjiman and Amparo Galindo, has been published by Wiley-VCH. Cover design based on an idea by Jens Schreckenberg, a member for the Molecular Systems Engineering group, and pictures from the book.
The molecular frontier: extending the boundaries of process design
Dr Adjiman's group brings molecular-level decisions into process design through fundamental advances in modelling and optimisation methods, and follows this through to implementation and application.
In designing new processes e.g., chemical processes, energy conversion devices – the molecules or materials required for processing (such as solvents or catalysts) are often chosen first. This is usually done from a limited set of choices (or design space) because evaluating many options, computationally or experimentally, is costly and time-consuming. Once this choice is made, the process topology and operating conditions can be chosen. This sequential decision making can lead to poor performance as it limits the design space and it ignores the intrinsic links between molecules and process. A simple question such as what is the best solvent for a reaction in a pharmaceutical manufacturing process cannot be answered in isolation. The answer depends on the reactor temperature and pressure, and on what follows in the process. If it is another reaction, it may be best to find a single solvent that works reasonably well for both reactions, to avoid expensive steps such as “swapping” one solvent for another, which often requires an energy-intensive separation.
By extending the boundary of process design to include molecular or material level decisions, real benefits can be accrued: better process economics, and lower environmental impact through increased material and energy efficiency. Further, by using optimisation to pose and solve extended design problems, the explicit evaluation of every design is avoided, making it possible to consider large design spaces: thousands of molecules, countless mixtures, wide temperature/pressure ranges, unusual process topologies. A few promising designs can be assessed experimentally to verify the results of model-based design. This general methodology, integrated molecular and process design, has been singled out for its potential to enhance innovation and competitiveness.
Explore the links below, research pages and publications to learn more.
et al., 2013, Computer-aided molecular design of solvents for accelerated reaction kinetics, Nature Chemistry, Vol:5, ISSN:1755-4330, Pages:952-957
et al., 2011, Integrated solvent and process design using a SAFT-VR thermodynamic description: High-pressure separation of carbon dioxide and methane, Computers & Chemical Engineering, Vol:35, ISSN:0098-1354, Pages:474-491
et al., 2010, A duality-based optimisation approach for the reliable solution of (P, T) phase equilibrium in volume-composition space, Fluid Phase Equilibria, Vol:299, ISSN:0378-3812, Pages:1-23
et al., 2009, Fluid phase stability and equilibrium calculations in binary mixtures Part I: Theoretical development for non-azeotropic mixtures, Fluid Phase Equilibria, Vol:275, ISSN:0378-3812, Pages:79-94