Using fuzzy logic to get miles out of hybrids and electric cars

Folks have a distressing tendency to believe that things once true will always remain true.

In the energy field, that kind of thinking is being turned on its head all the time.

(Photo: A deep cycle battery and a car, neither representative of the batteries and cars discussed in this story, but there just for illustrative purposes.)

New research, new technologies, new applications and changes in the price of oil—they all work to change the mix of what works, how well and how much better or cheaper than the next thing.

University of Hawai'i researchers Bor Yann Liaw and Matthew Dubarry, who are both engineers and materials scientists, are working on improving our understanding of how electric vehicles work, and how to make the battery banks last longer.

The two work with the Electrochemical Power Systems Laboratory at the Hawai'i Natural Energy Institute, which is in the university's School of Ocean and Earth Science and Technology. Their most recent paper, “From driving cycle analysis to understanding battery performance in real-life electric hybrid vehicle operation,” was published in the Journal of Power Sources.

Liaw and Dubarry have been working with a fleet of 15 Hyundai Santa Fe electric SUVs operated by Hickam Air Force Base, Hawaiian Electric, the City and County of Honolulu and the Hawai'i Electric Vehicle Demonstration Project, studying how people use the cars, and how ways of using them affect their performance.

The cars were used for errands and commuting, some at highway speeds and others—notably the military base cars—kept operating at lower speed limits. How to combine information from all those different driving styles into something that can help improve performance is the challenge, the researchers said.

“Conducting driving cycle analysis with trip data collected from (electric and hybrid vehicles) is very difficult and challenging,” they wrote.

The researchers combined something called fuzzy logic and pattern recognition to track the behavior of the vehicles. Fuzzy logic is a system for getting useful information out of complex systems that don't produce simple yes-no or true-false answers.

They used equipment tracking speed and time to measure how the cars were being driven, and looked at the pattern of battery discharge.

“Our ultimate goal is to use real-life data and laboratory testing to establish a realistic model for battery performance and life prediction,” they wrote.

In an email, Liaw conceded that some scientists question the viability of fuzzy logic in such issues, and he said his and Dubarry's work has not yet yielded specific recommendations for improving battery performance. But he said he is convinced the process will work.

“Yes, we have found an interesting way to quantify such influences in using the battery for portable applications by the users, but we need more efforts to make such understanding useful for implementation into practical devices to improve the battery life and efficiency," Liaw said.

It is the kind of work that, if successful, will yield improved electric vehicle performance—perhaps refuting the assumption that electric cars can never be use for more than short-hop driving.

And it's an example of how new research has the potential to significantly change our understanding of whether what used to be true, still is true.

© 2007 Jan W. TenBruggencate