Hybrid vehicle technology has become widely accepted due to its ability to reduce emissions and energy consumptions over conventional vehicles. Current optimization strategy for a parallel hybrid requires a lot of computational time and relies heavily on the drive cycle to accurately represent the driving conditions in the future. In this paper, a vehicle simulation model and a method to optimize parallel strategy to minimize energy consumption while having a fast solving time and being drive cycle independent is developed.
The proposed methodology differs from the widely used dynamic programming approach for optimizing parallel strategy, which relies on accurately representing the actual driving behavior, by using the efficiencies of the engine, electric machines, inverters and the battery to predict the impact of current operating point on the overall powertrain system efficiency. The approach takes into account both the forward and reverse losses due to load-balancing the engine in parallel operation using the electric machines. This information is precalculated to generate an operating map indicating the optimal load-balancing torque to maximize system efficiency across the operating range, enabling online use of this strategy for any given vehicle architecture on an arbitrary drive cycle.
A P2 parallel architecture is selected with specific components and a MATLAB Simulink model is built to validate proposed method. The results show that by using the proposed parallel strategy, propulsion system efficiency is not improved compared with using engine-only method due to the high electrical system losses. Parameter studies show a significant improvement in propulsion system efficiency by using the proposed parallel strategy when electric machine and inverter combined efficiency are greater than 93%. A modified testing method is also developed to yield a more consistent results over testing methods using a single drive cycle.