Abstract
The application of optimization methods and algorithms to energy management is crucial when trying to find instantaneous compromises between various energy sources that can provide the power required by a powertrain. Because of the complexity of both the problem and the system structure, it is difficult to determine the optimal strategy in real time (on-line and using the onboard computer). This article tackles the problem of optimizing the power provided by various sources available to meet the power demand from the driver whilst minimizing the total hydrogen consumption during a journey. The real challenge is to find an energy management law applicable in real time on any power profile. This paper presents two new energy management methods: off-line "Dynamic Programming with Improved Constraints (DPIC)" and a real-time optimized decision-maker based on a two-levels optimized Fuzzy Logic (Fuzzy Switching of Fuzzy Rules-FSFR). DPIC produces better results than the classical discrete dynamic programming with state-of-the-art constraints, in terms of execution time and hydrogen consumption. FSFR is a real time energy management algorithm based on fuzzy rules learnt on specific profiles and real-time fuzzy switching of these fuzzy rules. Both methods are evaluated on different types of real world profiles (urban, road and highway profiles), to assess and confirm their effectiveness.