Compound Lever Based Optimal Configuration Selection of Compound-Split Hybrid Vehicles

Finding optimal split hybrid configurations through exhaustive search is almost intractable, mainly due to the huge design space, e.g. 252 compound split configurations using two planetary gear sets (PG). Thus, a systematic exhaustive design methodology is required to find optimal configurations. While most of the prior studies proposed methodologies that assess the performance within the physical design space, i.e. based on the powertrain configurations, this paper proposes a compound lever-based comprehensive design methodology. The (virtual) compound lever is an attractive design tool defined by two design variables, i.e. α and β, that omits the redundancy existing within the physical design space, thus, reduces the computational load. The proposed method explores the entire (virtual) compound lever design space to find optimal compound split configurations with outstanding fuel economy and acceleration performance. First, two performance metrics, i.e. fuel economy and acceleration performance, are assessed within the entire compound lever design space, and optimal compound lever designs, i.e. optimal α and β, are selected. These selected compound lever designs, however, lack the physical parameters (e.g. PGs connections and gear ratios) required for further design steps. Therefore, a conversion model and conversion equations are developed in order to convert the selected optimal compound lever designs into feasible powertrain configurations. The developed compound-lever design methodology was applied to a few optimal designs on the Pareto front, and multiple novel compound split configurations with outstanding fuel economy and acceleration performance were found.
SAE International
Issue Date

SAE World Congress Experience, WCX 2017

Appears in Collection
GT-Conference Papers(학술회의논문)
Files in This Item
There are no files associated with this item.
  • Hit : 109
  • Download : 0
  • Cited 0 times in thomson ci


  • mendeley


rss_1.0 rss_2.0 atom_1.0