In Search of Super-Efficient and Fun-to-Drive Hybrid Electric Vehicles

An innovative search method is proposed to find super-efficient and fun-to-drive split hybrid powertrain architectures among all existing designs by using the virtual design space, which dramatically reduces the evaluation time from 114,000 years down to just seven days. A few novel power-split hybrid configurations with world-class fuel economy and acceleration performance (i.e. 27.8% faster than Toyota Prius, and 26.2% faster and 3.3% more efficient than Chevy Volt) were discovered.

Article  |  Spring 2018

Power-split hybrid electric vehicles (PS-HEV) such as Toyota Prius and General Motors Chevy Volt are one of the most promising solutions for the energy and environmental concerns around the globe. They owe their popularity mainly to their superior fuel economy and Plug-in capability. Unlike parallel hybrids, the PS-HEV, which uses two motor/generators (MG), can achieve high-performing full EV mode as well as a very efficient hybrid mode, and thus can be applied to both hybrid electric vehicles and plug-in HEVs.

Despite their high efficiency, if you are looking for a fun-to-drive vehicle, maybe the PS-HEV will not be your first option given the common prejudice that they are super-efficient, but usually ‘not so fun-to-drive’. Yet, one cannot guarantee that the commercialized configurations are the best within the large design space of PS-HEVs as they are just a subset of thousands of candidate configurations. In fact, connecting the four powertrain components (i.e. engine, two MGs, and output drive) to one or more planetary gear sets (PG) results in at least 200,000 designs. Additionally, multiple design parameters such as the PG gear ratios and the final drive gear ratio must also be considered when comparing one configuration to another. Therefore, a comprehensive design search of the entire design space is required to look for the ‘best’ performing power-split configurations.

Evaluating the performance within the physical design spaces (i.e. either the kinematic diagrams or the powertrain configurations) takes approximately 114,000 years, as illustrated in Fig. 1. This becomes intractable due to the large design space. In order to tackle this problem, Professor Dongsuk Kum from the Cho Chun Shik Graduate School of Green Transportation and his research team proposed searching for optimal configurations within the virtual design space. Evaluating the performance within the virtual design space omits the redundancy existing within the physical design space, as multiple physical configurations exist for a single virtual configuration, and thus drastically reduces the computational time to just 7 days (See Fig. 1).

 

The selected optimal virtual configurations are converted back into physically feasible configurations that can be realized, and their performance is verified using a hardware-in-the-loop simulation (See the video).

 

Novel configurations with world-class performance able to compete with the commercialized configurations were revealed by professor Kum and his research team. For instance, Configuration #5-C1, which uses the same components and vehicle specifications as the Prius 3rd generation, has a better acceleration time than both Toyota Prius (green) and Chevy Volt (blue), as illustrated in Fig. 2. In fact, the 0-160km/h acceleration time of configuration #5-C1 decreased by 27.8% (5.9 seconds) compared to Prius 3rd generation while maintaining a similar fuel economy. Configuration #5-C1 abolishes the common prejudice that power-split hybrids are usually ‘not so fun-to-drive’ vehicles, as it is not only super-efficient but also fun-to-drive.

 

In addition to three international journals published with this work, the best configurations revealed throughout this study have been protected through filing and registering seven patents in Korea and abroad. Professor Kum and his research team currently plan to further develop this technology and make further steps towards the real-world application of their PS-HEVs. Their ultimate goal is to manufacture a prototype that would be potentially mass-produced. Will KAIST power-split vehicles have their own share of the automotive market one day?

http://www.kaist.ac.kr/Upl/downfile/S09E04_20180402.mp3