logo SBA

ETD

Archivio digitale delle tesi discusse presso l’Università di Pisa

Tesi etd-01032022-104626


Tipo di tesi
Tesi di dottorato di ricerca
Autore
CARLONI, ANDREA
URN
etd-01032022-104626
Titolo
Wireless charging for drones and low-cost, open-source battery characterization instruments: advantages and design improvements.
Settore scientifico disciplinare
ING-INF/01
Corso di studi
INGEGNERIA DELL'INFORMAZIONE
Relatori
tutor Prof. Roncella, Roberto
commissario Prof. Saponara, Sergio
commissario Prof. Zamboni, Walter
commissario Prof. Fiori, Gianluca
commissario Conti, Massimo
commissario Prof. Gladwin, Dan
Parole chiave
  • drones
  • open-source
  • python
  • raspberry pi
  • Resonant inductive power transfer
  • single lithium-ion cell characterization
  • wireless charging
Data inizio appello
11/01/2022
Consultabilità
Completa
Riassunto
Resonant inductive-coupled power transfer is a very appealing technique for the battery recharge of autonomous devices like surveillance drones. The charger consists of two magnetically coupled circuits: the primary and the secondary. The primary is usually placed into the landing platform and the secondary is always mounted in the device to be recharged. Generally, the charger must charge the battery as fast as possible to reduce the no-flight gap. Furthermore, the design of the secondary circuit mainly focuses on lightness to improve the drone endurance. The latter consists of a full-bridge rectifier and an LC-filter. As a general rule of thumb, the filter cut-off frequency is usually set very low to charge the battery with DC quantities by choosing the capacity and inductance as big as possible to improve the filtering effects. However, these approaches may increment the size and weight onboard the drone. Moreover, the battery is generally assumed as constant voltage generator, but it is not ideal, and it shows a stray series resistance and inductance. If opportunely characterized, the designer can exploit these quantities to reduce the electrical components on the secondary circuit.

For these reasons, Iwe characterized and measured the equivalent lumped parameter of a real Li-ion battery by laboratory instruments. Then, I performed a simulation analysis, which use the measured battery stray inductance instead of the LC-filter external inductor and investigate the capacitor sizing effects on the battery charging power. It was found that the capacity modifies the charging power to the battery, and a value that maximizes the power exists but at the expense of a non-optimal transfer efficiency and an increased ripple current on the battery terminals. That value sets the LC-filter resonant frequency close to the double of the system excitation.

Successively, a theoretical study on the load composed of the LC-filter capacitor and the battery demonstrated that the above constraint is insufficient to achieve the power increment for any battery. It is found that there is a minimum frequency value below which the power transfer gain is not possible because the diode bridge rectifier does not work in discontinuous-mode, a necessary condition to improve the power transfer. The frequency transition point can be calculated, and the gain in power transfer can be obtained for any battery when its equivalent circuit parameters are known. Another simulation analysis was performed to confirm the theoretical findings, which exhibit a good agreement with the analytical investigation. A reasonable design recommendation is proposed to trade-off the gain in power transfer and the amplitude of the oscillating components of the wireless charger output. The recommendations are used to develop a wireless power transfer prototype to charge a drone. The experimental results confirm the simulation and analytical predictions with good agreement. Moreover, it demonstrates the feasibility of using the stray battery inductance instead of an external component to save further weight onboard the drone.

The design recommendation provides at the end of Chapter 4 and implemented in Chapter 5 allows the WPT system to charge a battery with a power 30% higher than the same system designed by using the general method adopted in the scientific literature. In this case, the selection of the WPT excitation frequency strictly depends on the battery parameters, such as the series-resistance. Therefore, the battery needs to be characterized before designing the WPT system. Furthermore, a diagnostic system should be required to perform regular maintenance operations on the battery. The maintenance keeps it in an optimal state and prevents severe fault during drone operations. However, this need is not limited to the drone. It should be required in many lithium applications. It is worth remarking the knowledge of the equivalent battery parameters is extremely important to design the system described above. Indeed, the technology improvements and cost reduction allow electrochemical energy storage systems based on Lithium-ion cells to massively be used in medium-power applications in the next decade, where the low system cost is the major constraint. Battery pack maintenance services are expected to be required more often in the future. For this reason, a low-cost instrumentation able to characterize the cells of a battery pack is needed. Several works use low-cost programmable units as Li-ion cell tester, but they are generally based on proprietary-software running on a personal computer. Therefore, an open-source software architecture based on Python language to control common low-cost commercial laboratory instruments is developed. The Python software application is executed on a Raspberry Pi board, which represents the control block of the hardware architecture, instead of a personal computer. The good results obtained during the validation process demonstrate that the proposed cell station tester features measurement accuracy and precision are suitable for the characterization of Li-ion cells. Finally, as a simple example of application, the state of health of twenty 40 Ah LiFePO4 cells belonging to a battery pack used in an e-motorcycle was successfully determined.
File