Nowadays, batteries are always more important in our lives especially in automotive field where they are the key point of innovation. In this scenario it is necessary to check how good the battery is. The aim of this project is to evaluate the State of Charge (SoC) and the State of Health (SoH) of a lead acid 12V battery; the first one indicates the remaining charge of a battery while the second is a measure of its aging. The two parameters are strictly related: an old battery has a low SoH, this implies a derating of its capacity and a consequent different behaviour of SoC. In order to compute these figures of merit, most commercial devices measure the voltage of the battery during the different operating stages. In the developed device both battery voltage and sinking/charging current are measured to obtain a better estimation of the desired parameters. This introduces huge problems related to the high currents involved; the technical solutions to these problems are dealt with in this elaborate. The whole project has been realized in cooperation with brainTechnologies Srl and SIVE SpA and sponsored by MESAP. The device consists in a PCB having at its core a 32 bit microcontroller, programmed using a real time operating system, and some properly chosen sensors connected to it permitting to collect the required data. The use of a RTOS allows to take advantage of multiprocessing computation, which plays a primary importance role in an application where different tasks have to be synchronised. The load current is transduced in a voltage through a shunt resistor and then, together with the battery voltage, it is acquired using a sensing device equipped with a 16 bit delta-sigma ADC. Such a high number of bits is necessary in order to cover a high current range while keeping also a good resolution. PCB and hood temperature are acquired too: the first one allows to compensate the thermal diffusion of the components, the second one is taken into account by the SoH and SoC algorithm. These data are stored in a flash memory and sent via Bluetooth to an Android device in order to be collected and plotted. The SoH and SoC values can be evaluated by the microcontroller (edge computing) or by a server by means of a neural network: the pros and cons of both the solutions are deepened in the elaborate. Firstly, the device has been tested in laboratory in order to verify the fulfilment of the requirements, then it has been field-tested to check the algorithm correctness. The followed approach and the developed technology can be applied not only to lead acid batteries but also to more modern ones, such as electric vehicles’ Li-Po batteries, simply by modifying the ratings of the devices and changing the algorithms.

AUTHORS: Di Fazio, Davide Fazi

ADVISORS: Eros Gian Alessandro Pasero, Giovanni Guida, Jacopo Ferretti, Vincenzo Randazzo

DGREE COURSE: Master’s degree in Electronic Engineering

ACADEMIC YEAR: 2019/2020