While fuel cell-powered automobiles look promising, they are still years away from commercial realization. And, hybrids seem to be the practical solution for fuel economy in the near future. Although there are more hybrids on the road today, cost and the current capabilities of battery technology keep them out of the mainstream. However, this could all change based on a research project being undertaken at Kettering University in Flint, Mich.
Supported by a grant of $75,000 from EDA supplier Mentor Graphics, researchers Juan Pimentel and Jim Gover, both professors of electrical and computer engineering at Kettering University, are developing computer models of an entire hybrid powertrain that will allow them to perform a complete analysis of a hybrid vehicle design. Using these models, they can then determine functionality, lifetime, effects of component aging, EMI generation, mechanical and electrical effects of thermal cycling, and cooling requirements and costs, among other things.
Currently, hybrids cost anywhere from $3,500 to $6,000 more than traditional cars. Besides vehicle cost and battery lifetime, battery energy and power density are also issues on the mind of a buyer. As a result of these issues, widespread adoption of hybrid electric vehicles has been delayed.
Because more than 50% of the cost is electrical in hybrid vehicles, the current practice of contracting all electrical systems to suppliers based on performance specs is costly. Plus, it requires them to conduct expensive tests to determine if their supplier has met their specifications. Simulation and modeling can go a long way in lowering the cost of a hybrid car in the future.
The research efforts in this project will focus on several areas of a hybrid powertrain. One such area is power electronics comprising a dc-dc converter and a three-phase inverter that converts dc voltage into a three-phase, frequency-dependent ac voltage. This voltage then provides power to the electric motor that operates the wheels. One starting point for this grant-funded project is to model the inverter and the control system for the ac motor. In fact, for this study, the researchers will use the powertrain of a Ford Escape hybrid, which incorporates a 50 kW high-frequency inverter.
According to the researchers, Gover will work on developing computer models of the power electronics system, and Pimentel will model the control electronics for the inverter. Both will use the IEEE's standard modeling language called very high-speed hardware description language-analog and mixed signal (VHDL-AMS). This work is being supported by Mentor's VHDL-AMS-based SystemVision, a multiphysics simulator. Here, the goal is to calculate the heating rate generated in the inverter semiconductor switches to determine cooling requirements over the full range of powertrain speeds, as well as the performance sensitivity to electrical parameters.
“The VHDL-AMS language is important to this project because it can handle complex systems where there is a mixture of electrical, mechanical, fluidic, thermal and other physical phenomena,” Pimentel said. Furthermore, he stated that the VHDL-AMS language “has acquired worldwide acceptance and is the language of choice for modeling and simulation of multiphysics complex systems.”
Another benefit of this project is that these models could provide automotive designers with the capability of designing a hybrid vehicle powertrain and evaluate it without having to build the hardware, which is extremely expensive.
When this project comes to fruition by year's end, the researchers believe that it will impact the cost of building a hybrid powertrain. Even though the inverter and the control system are just a subset of the complete powertrain. Consequently, other power blocks of the powertrain must also be modeled to realize the full impact. Ultimately, such small incremental steps will go a long way in making hybrids cost efficient.