News

The student Clara Rojas García obtained an EXCELLENT

---

• Thesis title: Parameter estimation of physics-based models for lithium-ion battery accounting for electrochemical and thermal effects

Court:

• Presidency: Mikaël Cugnet (CEA-French Alternative Enenergies and Atomic Energy Commission)
• Vocal: Javier Rodríguez Aseguinolaza (UPV/EHU)
• Vocal: Emilie Bekaert (CIC energiGUNE)
• Vocal: Elixabete Ayerbe Olano (CIDETEC)
• Secretary: Iker Lopetegui Tapia (Mondragon Unibertsitatea)

Abstract:

Energy storage systems (EES) play a key role in facing the actual energy transition. To supply the current energy demands with renewable sources and ensure the autonomy of the mobility sector, EES integration will be essential. Lithium-ion batteries (LIBs), in particular, are the leading technology due to its higher energy density and longer lifespan.

The characterisation and modelling of LIBs is a key field of study to develop efficient control strategies that will enhance the battery performance. For the next-generation of battery management systems (BMS), physics-based models (PBM), unlike other empirical or behavioural models, give relevant information about the physico-chemical processes occurring inside the cell. Model predictions heavily depend on a proper parametrisation, and then, many efforts are focus on find the best approach to estimate the parameters unequivocally. However, the high number of model parameters, the interdependencies between them, and the observability issues that the model itself may present, make the parametrisation a challenging task.

This thesis focuses on path the way towards accurate parameter estimation of an electrochemical-thermal model. Considering the wide variety of methods proposed to date, the objective is to design an efficient and fast technique that minimise the cost of the parameterisation without sacrificing accuracy. To this end, a mixed methodology is designed, combining invasive and non-invasive techniques. The methodology is first applied in a virtual environment where parameters are known. An exhaustive analysis of the methodology itself is then carried out to assess the suitability of the final parameter set.

In the real environment, a commercial cell is parameterised with the optimal experimental design and minimal invasive techniques. Additionally, extra variables no contemplated up to date, such as the generated heat, are included to optimise the model's thermal parameters. The assessment of the final model against experimental tests demonstrate that parameters can be accurately estimated by a mixed methodology. This research opens the way towards high-fidelity modelling for control and management of lithium-ion batteries in real-time applications.