PhD position: Hybrid process modeling combining mechanistic transport equations with machine learning for thermodynamic equilibria

Jülich Research Centre (FZJ)

Jülich, Germany

Work group:

IBG-1 – Biotechnologie



Area of research:

PHD Thesis



Job description:

The Helmholtz School for Data Science in Life, Earth and Energy (HDS-LEE) provides an interdisciplinary environment for educating the next generation of data scientists in close contact to domain-specific knowledge and research. All three domains – life & medical sciences, earth sciences, and energy systems/materials – are characterized by the generation of huge heterogeneously structured data sets, which have to be evaluated in order to obtain a holistic understanding of very complex systems. Visit HDS-LEE at: www.hds-lee.deThis HDS-LEE PhD position will be located at the Institute of Biotechnology (IBG-1). The institute investigates how microorganisms and isolated enzymes can be used to produce a variety of products from renewable raw materials. Moreover, IBG-1 hosts the Chromatography Analysis and Design Toolkit (CADET), which is the world leading open source software in the field. Find more Information about IBG-1 www.fz-juelich.de/ibg/ibg-1 and CADET github.com/modsim/cadet


Your Job:


You will apply different machine learning methods for quantitatively describing thermodynamic equilibria within PDAE based models of transport processes in chromatography devices. Technical challenges include the training of these black box models with and without propagating derivatives through the existing PDAE solver. The proposed approach addresses one of the most pressing problem in industrial application of chromatography modeling today, i.e. predictive simulation of processes with unknown adsorption mechanism. You will



  • replace mechanistic adsorption models by machine learning, such as Gaussian process regression, support vector machines, random decision forests, and neural networks (in no specific order)

  • embed these methods in the existing open source PDAE solver for chromatography models

  • improve the performance of the already implemented tangent linear algorithms using adjoint methods

  • evaluate the practical applicability of the proposed hybrid mechanistic/ machine learning approach for quantitatively predicting chromatography processes using carefully selected case studies with increasing complexity

Your Profile:

  • M. Sc. degree in physics, mathematics, biotechnology or a related field
  • Good knowledge of machine learning methods
  • Sound experience in at least one programming language (preferably Python and C++)
  • Experience in open source software development is of great advantage
  • Excellent knowledge of written and oral English: TOEFL or equivalent evidence of English-speaking skills
  • Excellent organizational skills and ability to work independently
  • Strong communication skills and capacity to strengthen a highly international and interdisciplinary team
  • High level of scholarship as indicated by bachelor and master study transcripts and two reference letters

Our Offer:

  • Outstanding scientific and technical infrastructure – ideal conditions for successfully completing a doctoral degree
  • A highly motivated group as well as an international and interdisciplinary working environment at one of Europe’s largest research establishments
  • Continuous scientific mentoring by your scientific advisor
  • Chance of participating in (international) conferences
  • Unique HDS-LEE graduate school program
  • Further development of your personal strengths via a comprehensive training program
  • A work contract for the duration of 3 years
  • Pay in line with 100 % of group 13 of the Collective Agreement for the Public Service (TVöD-Bund)

Forschungszentrum Jülich promotes equal opportunities and diversity in its employment relations.We also welcome applications from disabled persons.


Please apply via recruiter’s website.

Quote Reference: 2019D-240