David Gao recently received funding from the Department of Energy through the Argonne National Laboratory to test all the possible ways a vehicle's powertrain can be configured and sized to get the best fuel economy and performance. His goal is to develop the most efficient methods for finding the optimal design for hybrid powertrains using a sophisticated toolkit in an automated way.
The Powertrain System Analysis Toolkit, or PSAT, developed by the Argonne National Laboratory, simulates fuel economy and performance in a realistic manner using an unrivaled number of configurations of available power sources: conventional combustion, battery, fuel cell, and different hybrids. The greatest challenge is to find the optimal component sizes and control parameters.
"The challenge requires that we find new ways to optimize hybrid vehicles using PSAT, and our discoveries will be included in future versions of PSAT as a result of this research project," explained Gao, an assistant professor in TTU's Electrical and Computer Engineering Department who has his base in the Center for Energy Systems Research. "We'll investigate different optimization methods for this application."
"But hybrid electric vehicles design optimization can take hundreds of hours running PSAT on just one computer," he continued. "Our goal is to dramatically reduce that time to just a few hours by using multiple computers. It's known as distributed computing, and we have a lab with eight nodes running a program called Matlab Distributed Computing Engine which allows us to work faster."
The vehicles that most of us drive everyday are mechanically driven by a combustion engine. A hybrid electric vehicle (HEV) is a complex electro-mechanical-chemical system that involves two or more energy sources. In a HEV, there are one or more components you won't find in mechanical vehicles, including advanced batteries, ultracapacitors, fuel cells, electric traction machines and electronic continuously variable transmissions. The inherent advantages of HEVs are their increased fuel economy, reduced harmful emissions and better vehicle performance.
It is practically impossible to manually size specific components and test them because the interactions are so complex. Computer simulations using parallel computing, which requires multiple computers, is the most feasible solution.
"A human being cannot do this tradeoff," said Gao. "Manual tweaking doesn't work. It's too complicated because of the additional electrical parts."
Gao says the goal is to find the best marriage of parts that produce maximum fuel economy, minimum emissions and maximum performance.
"It's still a great challenge to make a hybrid vehicle have the desired drivability that consumers demand," said Gao. "It has to have an acceptable 0-60 acceleration time or we haven't found the whole solution."