This website uses cookies to improve your browsing experience and the continued use of the webpage indicates your consent to ÅF’s use of these cookies. Find out more about how ÅF uses cookies and how you can manage them here: Read more

Available job

Master Thesis - Creating a digital clone of a Process plant using Neural networks

About the company

ÅF Pöyry is an international leader within engineering, design and advisory services. We create solutions to support our customers worldwide to act on sustainability as well as the global trends of urbanisation and digitalisation. We are more than 16,000 devoted experts within the fields of infrastructure, industry and energy operating across the world to create sustainable solutions for the next generation.

Making Future.

About the job

Background:

The field of AI/Machine Learning is developing fast and the technology is maturing bit by bit every day. Traditional simulation tools are still the main approach for process simulation, but other technologies are gaining ground for certain problems. ÅF have identified a few problems where traditional modelling has not yet given satisfactory results. This thesis is a proof of concept where we aim to, together with the student, grow an understanding of the possibilities, as well as the difficulties, of applying deep learning to chemical processes.

Objective:

The objective of the thesis is to use simulated data, instead of real plant data, and train one or more networks to predict a set of outputs depending on certain inputs. The thesis comes in two parts where part one is to train one neural network over the entire plant model, and part two is to train a few neural networks for a number of sub-systems (one or a few unit operations per NN), and then to assemble them into the larger model.

Deliverables:

Create and train one NN for the entire systemCreate and train several NN for sub-systems and assemble them into a complete system modelEvaluating the two approaches in terms of difficulties, effectiveness, data requirements and the amounts of data point.  

Prerequisites:

  • The project should be performed by two students
  • The students should preferably have good knowledge of one of the most commonly employed programming languages in AI such as matlab or python.
  • Knowledge in AI/ML is highly merited (See recommendations below for available courses)
  • At least one of the students should be chemical engineer with focus in process chemistry
  • The thesis will be performed at and supervised by ÅF Gothenburg
  • Examiner is located at Chalmers  

Recommendations:

There are multiple online courses that are available for free. It is recommended that the student look into these and is well familiarised with its content prior to starting the thesis in January 2020.

 

Who are you?

Master Programs:

Students may preferably apply from master’s programs (or similar to) “Innovative and Sustainable Chemical Engineering” (Chalmers) or “Sustainable Energy Systems” (Chalmers) but are welcome to apply from mechanical engineering, physics and data/IT as well. It is also possible, from ÅF’s side, to make a joint thesis across fields, e.g. one student from Chemical Engineering and one from Data.

We offer

We are looking for someone who wants to be part of ÅF’s success story. Are you passionate about technology development? Do you like to work together to find the best solution? Then we can offer you career opportunities in a modern workplace with challenging assignments and exciting projects all over the world.

The ÅF Group is ranked as one of Sweden’s most popular employer among engineers. At ÅF you will be involved in developing innovative and sustainable solutions within infrastructure, energy and industry. We are always looking for the sharpest skills that can create a future society together with us. We hope you will learn as much from us as we will learn from you.

Contact information

Tobias Petersson
+46 105053385
tobias.petersson@afconsult.com

Application period ends 24/11-19 but recruitments may occur earlier!

Apply here