Studierende finden an der ETH Zürich ein Umfeld, das eigenständiges Denken fördert, Forschende ein Klima, das zu Spitzenleistungen inspiriert.
We invite applications for a PhD position that is part of a larger research project with the goal of demonstrating data-driven process control methods in a industrial manufacturing plant.
The successful candidates will join the Automatic Control Laboratory at the Department of Information Technology and Electrical Engineering as well as Inspire, the technology transfer unit at ETH Zurich under the supervision of Prof. John Lygeros and Dr. Alisa Rupenyan.
Industrial processes can be represented as cascaded structure where the outputs of one sub-stage in the process chain are the inputs of the next. Process control is built on the availability of these inputs and outputs, and on understanding the relationships between them, then implementing a control strategy to achieve process stability or certain quality characteristics. A process model based on the physics of the system is sometimes available, but its direct use could be limited to a certain class of problems, as it often involves complex, nonlinear physical dynamics models.
We propose to implement data-driven process modeling for the different sub-stages of an industrial process by establishing the input-output dynamic relationships at the substages of the process for use cases, where geometry, material, and type of process are specified. For each sub-stage, the input-output relationship can be learnt using data-driven techniques such as deep neural networks or gaussian process regression. In some cases, standard data-fitting techniques might be sufficient, or a simplified first principles model can be complemented by learning the model mismatch from data. Once the relationship between inputs and outputs is available at each stage, optimization-based techniques as model predictive control in combination with iterative learning, or extremum-seeking control can be applied.
This project offers the opportunity to be driving both method development, exploring various ways to integrate learning of the process parameters with optimal control methods, and practical implementation of the devised methods in an industrial environment.
The associated interdisciplinary PhD position available from October 2019, comprises several of the following activities:
This type of project is a collaboration with an industrial partner, and will be completed with the continuous support of a team of industry R&D experts, engineers, and doctoral students with background in mechanical, electrical, or chemical engineering.
We are looking for a candidate with a Master’s degree from a recognised university; strong analytical skills and background in systems and control, process engineering, or mechanical engineering. Experience with data modeling / machine learning, as well as C++/Labview programming could be of benefit.
We look forward to receiving your online application with the following documents: CV, statement of objectives or research interests, three contacts for reference, transcripts of all degrees, and one publication / thesis. Please note that we exclusively accept applications submitted through our online application portal. Applications via email or postal services will not be considered.
Further information about the Automatic Control Lab can be found on our website , and for Inspire on . For questions about the position, please contact Mrs. Baumann, Tel +41 44 633 85 09 or email baumasab@control.ee.ethz.ch (no applications).
29-02-2024
Bitte sage uns, wo du ähnliche Stellenanzeigen suchst und vergiss nicht deine E-Mail Adresse anzugeben!