par Goldar Davila, Alejandro ;Romagnoli, Raffaele ;Couto Mendonca, Luis Daniel ;Romero, Alberto;Kinnaert, Michel ;Garone, Emanuele
Référence Computers & chemical engineering, 141, 4
Publication Publié, 2020-07-01
Référence Computers & chemical engineering, 141, 4
Publication Publié, 2020-07-01
Article révisé par les pairs
Résumé : | In this work, we propose and demonstrate two low-computational MPC-based strategies to charge low-capacity commercial Li-ion batteries in the shortest time possible, while limiting the occurrence of degradation mechanisms. To do so, we use a low-complexity electrochemical model (the Equivalent Hydraulic Model, EHM). To reduce the computational burden of the resulting non-convex optimization problem, two approximations of the admissible region are considered: a time-invariant convexification, and a time-varying convex embedding. These approximations are experimentally tested on commercial LCO batteries (Turnigy 160mAh), under controlled ambient temperatures (10, 20, and 30 ∘C) and a non-controlled scenario. They are compared with the Constant Current-Constant Voltage (CCCV) protocol. Experimental results show that the proposed schemes achieve relevant reductions of the charging time with respect to the CCCV in thermal non-controlled and controlled environments. Results also highlight that the time-varying convex embedding does not significantly impact the computational time while further reducing the charging time. |