Batterie multilivello per la mobilità elettrica e le smart grid: dai modelli di cella al comportamento di sistema

The storage of electric energy is a key process in mobility applications but is also essential in fixed smart grids for the time-shifting of energy production/consumption. The storage unit is usually a battery pack with adequate power factors and with requirements concerning energy density, lightweight and long life cycle. In many applications, the battery is seen as a raw energy source, unsuitable to directly power the load. Thus, a converter/inverter is usually put as an interface between the battery and the final load to assure the required quality of output voltage. The first part of this thesis explores the advantages that can be gained by including the multilevel solutions of inverters inside the battery architecture. Batteries are complex non-linear systems that must be run under the supervision of a Battery Management System (BMS) to assure safe charge/discharge operations. The development of a complete autonomous system requires fine-tuning of many algorithms, which include the SoC/SoH estimation algorithms from voltage and currents signals at the battery terminals. The second part of this thesis aims at developing a Battery Digital Twin (BDT) that receives normalized state of charge, state of heath parameters as inputs, and produces a realistic voltage output signal in response to a load current waveform of any shape. This characteristic allows to use the BDT for many purposes: (i) for the validation of SoC and SoH estimation algorithms (ii) for battery monitoring or diagnostic when run in parallel with the real battery. This BDT is very useful in smart batteries where the single cells inside the pack can be used according to their states (SoC, SoH), possibly leading to a longer lifetime of the battery. Since the BDT allows to simulate a smart battery pack with cells at different charge and ageing states, it can be used for the validation of different SB balancing algorithms, consequently allowing to test the optimization of the lifetime of the SB.