The present paper aims to develop an innovative real-time power management strategy dedicated to the efficient operation of an auxiliary power unit (APU) in a heavy-duty vehicle. Specifically, the APU comprises a Solid Oxide Fuel Cell (SOFC) system and a Lead-Acid battery pack. The power management strategy envisages optimal power sharing between the APU elements and it is defined based on Simultaneous Perturbation Stochastic Approximation (SPSA) principle, pursuing SOFC power profile smoothing in real-time. The SPSA-based algorithm introduced here overcomes real-time operation issues remarked in other implementations (fuzzy logic, genetic algorithms), accounting for a robust and less complex formulation. The power management strategy is implemented in a dynamic model developed in Matlab/Simulink, simulating SOFC-based APU behavior. Simulation outcomes highlight that the proposed strategy allows a global energy saving over 6% if compared to a conventional power management, based on power split control. Moreover, comparing the power profiles corresponding to the battery and the SOFC, it is remarked as SOFC power oscillations evaluated over 1 s timeframe are halved, achieving values lower than 4.5 W/s for more than 80% of the operation time.
Stochastic power management approach for a hybrid solid oxide fuel cell/battery auxiliary power unit for heavy duty vehicle applications
Barelli L.
;Bidini G;
2020
Abstract
The present paper aims to develop an innovative real-time power management strategy dedicated to the efficient operation of an auxiliary power unit (APU) in a heavy-duty vehicle. Specifically, the APU comprises a Solid Oxide Fuel Cell (SOFC) system and a Lead-Acid battery pack. The power management strategy envisages optimal power sharing between the APU elements and it is defined based on Simultaneous Perturbation Stochastic Approximation (SPSA) principle, pursuing SOFC power profile smoothing in real-time. The SPSA-based algorithm introduced here overcomes real-time operation issues remarked in other implementations (fuzzy logic, genetic algorithms), accounting for a robust and less complex formulation. The power management strategy is implemented in a dynamic model developed in Matlab/Simulink, simulating SOFC-based APU behavior. Simulation outcomes highlight that the proposed strategy allows a global energy saving over 6% if compared to a conventional power management, based on power split control. Moreover, comparing the power profiles corresponding to the battery and the SOFC, it is remarked as SOFC power oscillations evaluated over 1 s timeframe are halved, achieving values lower than 4.5 W/s for more than 80% of the operation time.I documenti in IRIS sono protetti da copyright e tutti i diritti sono riservati, salvo diversa indicazione.
