Research in Product and Process Design
The development of systematic model-based methodologies for the rational design of processes and products continues to be an important area of activity within CPSE

Product and Process Design area encompasses a growing range of scales, from nanoscale models for materials selection, to mesoscale models for the design of processes for specific tasks, as well as, overall process models for integrated plant design. The approaches we develop are generic and we aim to enable engineers to meet the constraints and objectives imposed by today’s business environment, in particular, in the field of sustainable development. We consider not only economic aspects, but also environmental, safety and health factors and this is exemplified both in theoretical work and in technology development projects. Part of the work is focused on early process development. Other projects are applicable to later development stages, and focus on detailed design for separation, reaction, reactive separation, or operability issues such as controllability or maintenance. Modelling spans a range of scales and includes the design of devices where material issues play an important role such as solid oxide fuel cells, the level of process units, and entire process systems. This wide-ranging activity engages researchers along four main themes, and interacts very strongly with the Molecular Systems Engineering activity. 

Materials design for process synthesis 

The design of products and processing materials, such as solvents and catalysts, is tackled in a holistic manner, from the development of new property prediction techniques and modelling techniques to the use of these techniques in the design of environmentally benign yet functional systems. Problem formulation and the interplay between the mathematical form of the model and the optimisation techniques are key issues addressed in this area. The design problems we address have an increasing number of degrees of freedom: in addition to “standard” variables such as temperature and pressure, molecular structure, microstructure or formulation variables are considered explicitly in the problem formulation. Novel designs suggested by the modelling methodology are verified through targeted experiments in collaboration with experimentalists. The results of the experimental exercise are then fed back to the model, resulting in an iterative modelling-experimentation strategy.

Design of novel manufacturing processes

Models and techniques are developed for the design of state-ofthe- art processes with a particular focus on fine chemicals and polymers. Separation and/or reaction systems with increasingly complex interactions are considered, which require the combination of detailed models and state-of-the-art numerical techniques.

Integrated process synthesis

The interactions of design and operability are used to create processes with better overall performance. Diverse tools such as life-cycle analysis, computational fluid dynamics and process modelling are combined to enable the consideration of multiple decision criteria.

Technology transfer

This activity is focused on facilitating the transfer of our more mature technologies to industrial partners.

 

New and noteworthy achievements in the past year:

Advanced process modelling of chromatography for industrial bioseparations

Edward Close & Eva Sørensen

Therapeutic proteins are well established as a clinically and commercially important class of therapeutics, and have played a key role in major advances in the treatment of various disorders and diseases over the last quarter century such as cancer and autoimmune diseases. Developing and operating biopharmaceutical processes is, however, extremely challenging and is traditionally almost completely reliant on an extensive experimental effort conducted at great effort, time and cost. In collaboration with Pfizer and the Department of Biochemical Engineering at UCL, we have been developing systematic approaches to bioprocess design by integrating high fidelity predictive models of chromatographic bioseparations and stochastic simulation methodologies with traditional experimentally based industrial bioprocess development platforms. The objective has been to accelerate the development of industrial purification processes and to increase process robustness, whilst deriving fundamental knowledge and process understanding.

We have developed detailed models of key purification steps in the production of multiple commercial therapeutic proteins with annual revenues totalling over $1 billion. These models have been used in various applications across the development timeline. During early phase development, models have provided a link between high throughput ultra-scale down experimentation and laboratory scale scouting column experiments, identifying robust operating parameter ranges for challenging separations and directing optimal performance, whilst decreasing total development times. For candidate molecules nearing commercialisation, models have been used to generate probabilistic design spaces in order to resolve process performance issues and reduce risk by quantifying the impact of this process variability on the design space. The successful application of process systems engineering approaches to industrial bioseparations in this work provides a basis for the next generation purification process development.

Integrated process and solvent design for carbon capture

Jakob Burger (University of Kaiserslautern), Smitha Gopinath, Vasileios Papaioannou, Amparo Galindo, George Jackson, Claire Adjiman

We have been developing novel methodologies for the integrated design of processes and solvents by combining advances in the prediction of physical properties (see Molecular Systems Engineering) with computer-aided molecular design approaches. A specific challenge in this context is to handle the numerical complexities that arise from the simultaneous variation of the solvent and the process conditions, which can lead to large changes in the fluid phase behaviour within the process units. To address this issue, we have developed a hierarchical approach, in which an idealised process is first investigated using multi-objective optimisation. The best solutions (solvent molecular structure and process operating variables) that arise from this step are then used as starting points for a more reliable optimisation, in which a detailed process model is considered.

This hierarchical approach has been applied to the separation of carbon dioxide from methane via physical absorption. Natural gas streams often contain a large proportion of carbon dioxide which must be removed from the stream to increase its value. They are produced at a high pressure (typically a few MPa), which makes physical absorption a viable option. Using a simple absorption/desorption process consisting of one absorber and one flash unit, the SAFT-γ Mie equation of state was used in a predictive manner to model the impact of using different solvents on process performance (methane purity, recovery, net present value of the process). The class of solvent considered consisted of n-alkanes and linear ethers (methyl ethers, symmetric ethers, oxyethylene ethers and oxymethylene ethers). The hierarchical design approach allowed the identified of tetra(oxymethylene) dimethyl ether as the best solvent for this process, with a net present value 60% larger than that for the best n-alkane solvent.

Figure 1.
Figure1. Solutions of the multi-objective optimisation used to determine suitable starting points for integrated process and solvent design. Solutions on the Pareto front are shown, based on three objective functions: solvent flowrate (vertical axis), methane production rate (horizontal axis) and solvent loss (colour scales