BAT-MAN 2nd life: Data acquisition system for Lithium-ion cells SoH and SoC estimation

Lithium-ion batteries play a vital role in the rapidly expanding electric vehicle market. However, the challenges associated with battery use, reuse, and disposal remain significant. The BAT-MAN 2nd Life project aims to extend BAT-MAN, a device designed by Brain Technologies for diagnosing lead-acid batteries, to characterize lithium-ion batteries. State of Health (SoH) and State of Charge (SoC) are crucial metrics used to evaluate the battery's behavior for its future life and can be obtained by employing specific algorithms. This thesis aims to prototype a device for acquiring the electrical parameters needed to characterize lithium-ion cells. The system was designed to validate the theoretical algorithms and serve as a starting point for an application-specific device capable of autonomously managing battery diagnosis and characterization. After an initial research phase, a discharge system was designed to discharge the battery at a selected current that remains constant throughout the test, irrespective of battery variations and environmental changes. Three discharge modes were implemented: constant, step, and random. The constant mode is used to determine the actual battery capacity by discharging at low currents to avoid stressing the battery and calculating the total depleted charge. The step mode enables the characterization of battery behavior under higher current stress, while the random mode adds additional test points to the model. The system was designed to prevent current spikes that could harm its components or the battery in the case of state transitions or faults, employing real-time battery voltage monitoring for safety purposes. The system behavior is managed by a microcontroller, with data being collected and sent via UART to a PC for post-processing, generating plots, and arranging the data in CSV format to feed algorithms (developed separately) thanks to a Python script. The system's firmware was developed on the STM32 platform, implementing a finite state machine on FreeRTOS. The entire test process is automated due to its time-consuming nature, with each discharge cycle preceded by a charge performed using a commercial charger. The system prototype was implemented on a stripboard and extensively tested in the laboratory. The results aligned with the expected models. Furthermore, the system provides ample room for improvement, both in terms of hardware, such as the development of a PCB or an ASIC, and firmware capabilities, as it has the potential to track current profiles in addition to constant current discharging.