A general model to assess fuel consumption of HEV architectures using Dynamic Programming

Designing a hybrid vehicle powertrain remains a complex task. To explore the entire design space, tools are developed to search the architecture topology level. Considering that a methodology could generate all the possible graphs of architectures using a certain amount of components, the issue is then the assessment of these architectures. This paper proposes an automatic assessment method to tackle this issue. A specifically developed general hybrid model is developed. It calculates the fuel consumption of all the architectures generated by an upper level of architecture generation that provides for each architecture a description of the modes available and the connections between the components. Based on this information and on the component characteristics, a specific Dynamic Programming method allows to find the optimal energy management of the architectures leading to the minimum fuel consumption on a given driving cycle.