PhD Position in Data Driven Fault Diagnosis for Electric Vehicles

The introduction of heavy electric vehicles is an important step for reducing greenhouse emissions. Still, for these vehicles to be truly sustainable, we need to optimize the way we operate their electric batteries. This requires on-board diagnosis and prognosis algorithms that can maximize the battery lifespan and then allow to repurpose them in less demanding applications, such as grid energy storage.

While advanced diagnosis algorithms are model-based, their design, tuning and validation require knowledge of first principle equations and considerable amounts of data. Often, more than what is available: especially in real-world operating conditions, where continuous sampling over large fleet of vehicles is prohibitive.

On one hand, there is a need for system identification approaches that can deal with large, sparse or irregularly sampled datasets. On the other hand, there is a need for pure data-driven, machine learning approaches to simplify model identification from such datasets.

You will carry out research as part of the project SPARSITY [http://www.dcsc.tudelft.nl/~riccardoferrar/projects/funded_projects/6_SP... , which is an academic-industrial collaboration between the Delft Center for Systems and Control (TU Delft, The netherlands) and Volvo Group, a world-leading automotive company based in Gothenburg (Sweden). You will be supervised by Dr. Riccardo Ferrari [http://www.dcsc.tudelft.nl/~riccardoferrar] , who is leading a young, dynamic and diverse group of researchers focusing on fault tolerant control, with applications in automotive and aerospace diagnosis, in safety and security of industrial control systems and in renewable energy generation. You will be co-supervised and mentored by Prof. Michel Verhaegen [https://www.dcsc.tudelft.nl/~mverhaegen] , a renown expert in system identification methods and an ERC grantee.

You will focus on research topics like:

• extending state-of-the-art system identification algorithms to use sparse, large datasets;
• developing pure data driven, machine learning methods with lower complexity for obtaining prediction models from such datasets — e.g. using compressive sensing for neural networks
• applying diagnosis and prognosis methods to automotive components, such as lithium-ion batteries.

You will have the chance to test your algorithm on real use cases and datasets, thus benefitting from TUD collaboration with Volvo.

The department Delft Center for Systems and Control (DCSC) of the faculty Mechanical, Maritime and Materials Engineering, coordinates the education and research activities in systems and control at Delft University of Technology. The Centers' research mission is to conduct fundamental research in systems dynamics and control, involving dynamic modelling, advanced control theory, optimisation and signal analysis. The research is motivated by advanced technology development in physical imaging systems, renewable energy, robotics and transportation systems.