This thesis presents the development and validation of an optimized fast-charging strategy for a pack of 4-in-series lithium-ion NMC pouch cells using Model Predictive Control (MPC). A physics-based reduced-order model (ROM) was selected, improved, and rigorously validated both against full-order simulations and experimental data. Additionally, a thermal submodel was developed from scratch to capture internal heating and thermal interactions within a 4-cell pack configuration. These models formed the control framework aimed at minimizing charging time while mitigating aging mechanisms, with particular focus on lithium plating, predicted through anode potential monitoring. The proposed MPC strategy was benchmarked against conventional Multistage Constant Current (MSCC) and Constant Current Constant Voltage (CCCV) protocols, demonstrating superior performance in both charging speed and safety. Furthermore, a complete experimental setup was designed to support future hardware-in-the-loop validation of the control strategy. This work lays the foundation for advanced, real-time, aging-aware battery management systems tailored for practical energy storage applications.
In questa tesi è stato sviluppato e validato un metodo di ricarica rapida ottimizzato per celle pouch agli ioni di litio con chimica NMC, utilizzando come metodo di controllo controllo il Model Predictive Control (MPC). È stato selezionato, migliorato e validato un modello fisico semplificato (Reduced-Order Model, ROM), confrontandolo sia con simulazioni Full-Order che con dati sperimentali. Inoltre, è stato sviluppato da zero un modello termico per stimare il riscaldamento interno e le interazioni termiche all'interno di un pacco composto da 4 celle. Questi modelli costituiscono la base di un sistema di controllo progettato per ridurre i tempi di ricarica, limitando al contempo i fenomeni di invecchiamento della cella, con particolare attenzione alla deposizione di litio (lithium plating), monitorata attraverso il potenziale dell’anodo. Il protocollo MPC proposto è stato confrontato con i metodi convenzionali Multistage Constant Current (MSCC) e Constant Current Constant Voltage (CCCV), dimostrando prestazioni superiori sia in termini di velocità di carica che di sicurezza. Infine, è stato progettato completamente da zero un setup sperimentale per un pacco da 4 celle, al fine di consentire future validazioni hardware-in-the-loop del controllo. Questo lavoro costituisce una base solida per sistemi di gestione batterie avanzati, in tempo reale e consapevoli dell’invecchiamento, applicabili a scenari reali di accumulo energetico.
Development of model predictive control-based fast charging algorithms for a 4-cell pack of li-ion batteries
PASSERA, ALESSANDRO
2024/2025
Abstract
This thesis presents the development and validation of an optimized fast-charging strategy for a pack of 4-in-series lithium-ion NMC pouch cells using Model Predictive Control (MPC). A physics-based reduced-order model (ROM) was selected, improved, and rigorously validated both against full-order simulations and experimental data. Additionally, a thermal submodel was developed from scratch to capture internal heating and thermal interactions within a 4-cell pack configuration. These models formed the control framework aimed at minimizing charging time while mitigating aging mechanisms, with particular focus on lithium plating, predicted through anode potential monitoring. The proposed MPC strategy was benchmarked against conventional Multistage Constant Current (MSCC) and Constant Current Constant Voltage (CCCV) protocols, demonstrating superior performance in both charging speed and safety. Furthermore, a complete experimental setup was designed to support future hardware-in-the-loop validation of the control strategy. This work lays the foundation for advanced, real-time, aging-aware battery management systems tailored for practical energy storage applications.File | Dimensione | Formato | |
---|---|---|---|
2025_07_Passera.pdf
accessibile in internet per tutti
Descrizione: testo tesi passera
Dimensione
5.42 MB
Formato
Adobe PDF
|
5.42 MB | Adobe PDF | Visualizza/Apri |
I documenti in POLITesi sono protetti da copyright e tutti i diritti sono riservati, salvo diversa indicazione.
https://hdl.handle.net/10589/240621