ETD

• battery ageing metrics
• battery electrical model
• battery thermal model
• dual extended kalman filter
• experimental charaterization
• LiPo battery
• parameter identification
• state of healt estimation
• state of health observer

Electrification, as a part of the industrial development aimed at achieving the goals set by the Paris Agreement, poses a series of challenges to every field of engineering. The transport sector, and aeronautics in particular, cannot escape them. This gives rise to the increasing focus on electrical energy sources, foremost among which are Li-Ion batteries. These devices present a range of properties requiring careful investigations to achieve their successful application in flying vehicles. The development of dynamic models of Li-Ion batteries represents a critical aspect, especially because the device response is inherently nonlinear and strongly depends on ageing. As an example, a battery is considered at End-Of-Life (EOL) when its capacity lowers to 80% with respect to the initial service conditions. If the capacity is not correctly monitored, this means that 20% of the battery weight, though useless, is transported, with consequent reduction of flight endurance. For this reason, it is crucial to develop battery management system capable of estimating battery states (charge, temperature, age).
In this MSc Thesis, starting from the results of previous research activities carried out at the University of Pisa, the parametric identification of the nonlinear electro-thermal dynamic model of a commercial Li-Po battery (MaxAmps, 6 cells, 22.2 V, 2 Ah) is obtained, by carrying out an extensive test campaign in which HPPC identification tests and flight mission simulations are alternated in order to include ageing effects. The experimentally-validated model is thus used to develop and validate observers of the battery states. As a relevant outcome of the work, a capacity-based SOH observer has been developed, using a Kalman Filter idea in the field of parameter estimation. This estimation technique is compared with the experimental results and evidence as a validation procedure that tries to assert the validity in the implementation.