Advanced modeling and development of mathematical methods for real-time diagnosis on lithium-ion batteries and state prediction

The capability to assess and monitor the state of charge (SOC) and state of health (SOH) of lithium-based cells is a highly demanded feature for advanced battery management systems. The main purpose of this thesis is to develop advanced models and methodologies to have a real-time diagnostic of battery state. In this regard, electrical, electrochemical, and aging characteristics have been analyzed, in order to develop accurate and low-computational cost models in the time and frequency domain. In particular, a novel electrical lithium-based battery performance model with aging characteristics has been defined and validated through experimental tests. The virtual model developed and implemented in Simulink platform has an interesting peculiarity. Model parameters have been evaluated from experimental tests performed on several end-of-life (EOL) automotive cells, at different SOHs. Nowadays, there is a lack of reliable and accurate battery models to assess the applicability of EOL batteries, giving them a second-life in stationary applications. Consequently, the battery model developed in this work could simulate the performance of a real second-life battery system. Moreover, this work has been presented a set of algorithms for the estimation of SOC, specifically deployed for lithium-ferrum-phosphate (LFP) batteries. The algorithms proposed are founded on state-observers model-based approach. Especially for LFP batteries, key factor is the introduction of a hysteresis property inside the battery voltage model for an adequate treatment of SOC estimation error. Finally, an innovative SOH diagnosis method for lithium-based cells is proposed. Battery SOH can be detected by exploiting impedance measurements obtained by fast active electrochemical impedance spectroscopy test. Key factors are the following: first, the excitation on the battery has been generated by a novel electronic prototype. This prototype is made up of cheap facilities and can perform EIS tests with reduced time duration and low energy consumption. Second, the experimental EIS test has been shown the possibility to determine frequency points in which the impedance measurements dramatically change due to different cell SOH. As a consequence, these peculiar frequencies can be adopted as a reference for cluster separation and SOH determination.