In process synthesis, the selection of suitable materials and molecules plays a decisive role in energy and process efficiency. Computer-aided molecular design (CAMD) makes it possible to optimally design not only the operating conditions but also the working medium of a heat pump for the coupled process. The optimization of molecules hereby poses a particular challenge, as it is often based on complex, non-linear intermolecular relationships. This makes a gradient based approach unsuitable in most cases, leaving only gradient-free methods to solve the tasks at hand. This thesis therefore aims to develop a methodology based on Bayesian optimization (BO). This thesis approach enables a systematic search for optimal molecules by combining a graph-based Gaussian process as surrogate for the molecular structure, with conventional mixed-integer optimization. The aim is to develop a robust approach that automates molecular selection in process synthesis, thereby further improving energy integration.
About Division 2.2 at BAM
In division 2.2, research focuses on the development of dynamic process models to describe all system states from start-up to shutdown, supported by an open library of dynamic, pressure-driven models. The division uses machine learning, uncertainty quantification and hybrid modeling methods to drive forward real-time application and plant monitoring. For the comprehensive digitalization of process engineering, the division is also researching information and data modelling as well as process models from engineering to the operation of chemical plants. This is supplemented by the development of methods for the safe and optimal transformation of chemical plants.
Tasks:
• Development of a methodology for molecular optimization based on Bayesian optimization.
• Application of BO to graph-based molecular structures in combination with mixed-integer optimization.
• Validation of the developed methodology using relevant process examples for the optimal selection of working media.
Desirable skills:
• Experience with Python, Matlab or similar.
• Basic knowledge of process modelling and multiphase thermodynamics.
• Experience with optimization methods, in particular Bayesian optimisation and mixed-integer optimization.
Contact persons:
Lukas Scheffold, M.Sc. :
lukas.scheffold@bam.de
Stefan Tönis, M.Sc:
stefan.tönis@bam.de