In hybrid vehicles, onboard algorithms are aimed at splitting power between the various energy sources in order to minimize fuel consumption and/or pollutant emissions. This real-time management requires a dynamic optimization approach (or optimal control).
As part of a thesis project(1, 2), IFPEN developed an energy management optimization method incorporating the internal temperatures of the engine and/or the gas after-treatment system, – which represent dynamic states not taken into account by state-of-the-art optimized management systemsa.
Initially, based on prior knowledge of the driving cycle, management laws were determined using the Pontryagin Minimum Principleb. Based on the numerical results, a compromise was established between the optimality of the management system and the complexity of the model used. For each of the different cases studied, the impact of the simplification implemented was estimated using regular perturbation theory for optimal control.
During a second phase, the ECMS method was enhanced by incorporating new thermal dynamics. This led to the development of sub-optimal strategies — not dependent on knowledge of the driving cycle, — which were then validated numerically and experimentally.
The method developed could be extended to other optimal control problems, either multi-state or, more generally, complex.
a- based on the ECMS (Equivalent Consumption Minimization Strategy) method in which the unknown equivalent cost of electricity (adjoint state) is calculated as a function of the measured battery charge (state).
b- mathematical principle used to specify the conditions required for optimality.
(1) D. Maamria, A. Sciarretta, F. Chaplais, N. Petit (2017), in Proc. of the IFAC World Congress, Toulouse, France, 9-14 Jul. 2017.
(2) D. Maamria, F. Chaplais, N. Petit, A. Sciarretta (2015), Comparison of several strategies for HEV energy management system including engine and catalyst temperatures - Proc. of the American Control Conf. (ACC), Chicago, IL, 1-3 Jul. 2015.
>> DOI: 10.1109/ACC.2015.7171183
Scientific contact: Antonio Sciarretta