PhD Research Project: Energy storage systems behaviour prediction via machine learning algorithms

Energy storage systems are one of the most promising technologies in the market for a wide range of applications. Electromobility is one of the applications requiring these kind of systems, for instance in electric vehicles and electric trains, which are autonomous systems and require a certain amount of energy available without any grid connection. However, other electromobility applications like elevators are also including energy storage in their designs even if they have a grid connection to improve the performance of their products in many different ways. Besides, energy storage systems are also included in grid applications to improve its performance against sudden non-desired events. Electrochemical batteries and more concretely lithium ion batteries are one of the most promising technologies within energy storage although material science researchers are looking for better performance technologies according to degradation and energy density.

One of the most important drawback of this technology is the heterogeneous behaviour of the battery cells depending on the chemistry, technology, cell format, manufacturer or even production and storage ambient conditions. Moreover, during the whole life of the battery cell, the cycling performance and ambient conditions affect the degradation mechanisms and this degradation affects the electric performance of the cell. Thus, machine learning algorithms are presented like a promising solution to overcome the problem of the heterogeneous performances. Nowadays, one of the most common solution is to cycle the cell under certain conditions to evaluate its behaviour during its life. This solution requires a huge amount of resources in equipment to test the cells and in researchers to analyse this data.

Machine learning is an area of study that tries to apply algorithms on a set of data samples to discover patterns of interest and can help a lot in the described context reducing the amount of experimental test done and making easier changing the technology used for manufacturing the electrochemical cell. Nowadays, many applications in any field of technology measures all the relevant variables during long time periods and one of the most promising fields in the future will be the use of this set of data samples for improving these applications.

The objective of this thesis is to predict the behaviour of electrochemical cells based energy storage systems using machine learning algorithms, optimizing their performance for the target application.

 

Position details

Organization: Mondragon Unibertsitatea. Faculty of Engineering.

Research area: Drive Systems Applied to Traction and the Generation of Electric Energy

Researcher profile: First Stage Researcher (R1).

Type of contract: Research contract - Temporary.

Job status: Full time.

Location: Arrasate-Mondragon, Gipuzkoa, SPAIN

PhD supervisor: Unai Iraola, MONDRAGON University (uiraola@mondragon.edu)

PhD co-supervisor: Iosu Aizpuru, MONDRAGON University (iaizpuru@mondragon.edu)

Funding information

https://www.mondragon.edu/en/research-transfer/engineering-technology/aids-scholarships