Toolkit for Tomorrow's Car
A new weapon in the war on runaway fuel consumption and the pollution it causes
by Scott T. McBroom
To help meet performance, emissions and fuel economy goals for the 21st century, engineers in the SwRI Engine and Vehicle Research Division are developing computer software tools that simulate advanced vehicle powertrains. This effort is part of the Partnership for a New Generation of Vehicles, or PNGV.
The PNGV Challenge
A major goal of the PNGV program, and the targeted use of the software being developed at SwRI, is development by 2004 of a production prototype mid-size family sedan with a fuel economy of up to 80 miles per gallon. That is three times the fuel efficiency of today's comparable class of vehicles such as the Ford Taurus, Chrysler Concorde, and Chevrolet Lumina. The goal further challenges engineers to maintain or improve current levels of performance, size, utility, and cost of ownership, and the new vehicle must meet or exceed federal safety and emissions requirements.
While there are many ways to achieve 80 mpg, the hybrid electric vehicle is seen as one of the more promising alternatives. A hybrid vehicle is one in which two sources of energy are converted to power the vehicle (see sidebar). Hybrid vehicles offer a number of advantages: they can recover energy normally dissipated by the brakes, reduce gaseous emissions by virtue of their smaller heat engines, and improve fuel economy with more efficient electric drives.
Another alternative, the conventional powertrain consisting of a reciprocating internal combustion engine and transmission, averages only 27 mpg in today's mid-size sedans. Though fuel efficiency gains have been made over the last two decades, just 15 percent of the energy from the gasoline in the tank is used to move the vehicle. Considerable improvements in engine efficiency and significant vehicle weight reduction would be required for a conventionally powered sedan to meet the up to 80-mpg goal. Many are confident, however, that such improvements can be made.
The breadth of existing and conceptual technologies being considered and the anticipated cost of research and development to build prototype cars of the future require that a comprehensive analytical capability be designed to select and integrate the most appropriate technologies. The United States Council for Automotive Research (USCAR) hopes to meet this challenge with the PNGV Systems Analysis Toolkit.
The PNGV Toolkit
The PNGV Systems Analysis Toolkit will be used to evaluate powertrain configurations and components such as advanced heat engines and energy storage devices in conventional and hybrid electric vehicles, two options determined by PNGV participants to be among the most capable of satisfying program goals. It provides an opportunity for component experts to see how their technologies interact in the context of whole and varied vehicle systems.
The Toolkit allows analysts to conduct trade-off studies for performance considerations such as 0-60 mph acceleration, fuel economy, and emissions, as well as for nonperformance considerations such as cost and reliability. The immediate results of the trade-off studies include projection of component power and energy requirements, identification of areas where energy is being lost, and management of power and energy, all of which will aid in determining which components and systems can help the car of the future achieve up to 80 mpg.
The first phase of Toolkit development was completed in December 1996. The next phase is scheduled for completion in August 1997, with planned enhancements beyond 1997 to include porting the Windows™-based program to client workstations or networks, allowing users to simultaneously perform virtual design prototyping, vehicle simulation, and visualization on a single platform; permitting the addition of custom component models and data so as to perform a wider range of simulations; and upgrading the program to provide for high-fidelity, dynamic analyses.
Three sections make up the Toolkit architecture. The graphical user interface allows vehicle definition, results reporting, and static modeling, including cost and reliability analyses. Vehicle power train models developed with the MATLAB/ Simulink™ programming language are used to assess performance, emissions, and fuel economy through a program developed at SwRI called APACE™ (see following section). Finally, a software database is provided for storage of a component library, vehicle configurations, and simulation results.
Vehicle systems analysis begins by creating computer models that simulate the major effects of desired powertrain components. The program can, for example, calculate how much energy is left in the battery given a number of different driving scenarios. The analysis also allows integration of components into appropriate configurations and produces results related to key vehicle characteristics. In addition to increased fuel efficiency, key desired characteristics for the PNGV prototype vehicle include acceleration of 0-60 mph in 12 seconds, continuous driving on a 6.0-percent grade at 55 mph, and a range of 380 miles.
It is anticipated that the Toolkit will be used by PNGV participants to determine the direction and magnitude of research programs necessary to remain on schedule while meeting technical requirements. Before delivering a prototype in 2004, participants must produce a concept vehicle by the year 2000. To accomplish this task, the most promising technologies will be identified with the aid of an Institute-developed program that assesses vehicle and component performance.
APACE™
As a means to narrow the broad field of candidate technologies, Institute engineers have created a new simulation program known as Advanced Powertrain Assessment, Comparison and Evaluation, or APACE™, for integration into the PNGV Toolkit. APACE™ models the emissions, performance, and fuel efficiency of conventional and hybrid electric vehicles and their components. Component models are programmed in MATLAB/Simulink™. Components completed to date include:
- Direct Injected Spark Ignited Engine
- Direct Injected Compression Ignited Engine
- Turbo-Alternator (turbine engine integrated with high-speed alternator)
- Permanent Magnet Motor and Generator
- AC Induction Motor
- AC Synchronous Generator
- Advanced Batteries
- Lead Acid Battery
- Electromechanical Flywheel Battery
- Ultra-Capacitor
- Manual Transmission
- Automatic Transmission
- Continuously Variable Transmission
- System Controls
To implement the models, APACE™ relies on time-based integration for performance prediction. This "forward-looking" approach simulates the behavior of actual control systems and enlists control loops to set and correct the behavior of the system. For example, in a conventional vehicle model, when the software element simulating the driver wishes to achieve a desired speed from rest, an acceleration command is produced. The command is received by the power controller, which then issues a throttle command to the engine. If by the next simulation time step the vehicle has not achieved the desired speed, the acceleration command is increased. MATLAB/Simulink™ uses a variable time step integrator that decreases the time step until the difference between desired and actual vehicle state falls within a specified tolerance.
The forward-looking technique allows development of realistic control algorithms that can be used in hardware-in-the-loop analyses, in which the computer model (software) controls a vehicle component (hardware) operating on a test stand. This reduces experimental prototyping by permitting evaluation of component performance as part of a simulated system.
Conclusions
Though the PNGV Toolkit has the immediate task of supporting development of an up to 80-mpg production prototype by the year 2004, its future use may be much broader. Using the Toolkit, it may one day be possible for automotive engineers to design a variety of new vehicles entirely by computer -- selecting a powertrain, sizing components, and evaluating performance, cost, and reliability through a user-friendly interface that automates the process. With automakers targeting a 28-month concept-to-production turnaround time, digital prototyping and simulation will save time and resources by reducing the many stages of hardware development and testing.
Acknowledgments
Collaborating with Southwest Research Institute on PNGV Toolkit development is TASC, Inc., with assistance from Reality Graphics, Oakland University in Rochester, Michigan, and the University of Michigan in Ann Arbor. The first phase of this effort was supported by NASA; current work is sponsored by the U.S. Army National Automotive Center. USCAR provides technical direction, and technical support is provided by the Ford Motor Company, General Motors Corporation, Chrysler Corporation, National Renewable Energy Laboratory, Idaho National Engineering Laboratory, Argonne National Laboratory, and Lawrence Livermore National Laboratory.
Development of the APACE™ program would not have been possible without the contributions of several SwRI staff members, among them Joseph Baumgartner, John Bishop, David Buntin, Joseph Grogan, Don Mowbray, Cherian Olikara, Jayant Sarlashkar, and Dr. Robert Thring.
Published in the Summer 1996 issue of Technology Today®, published by Southwest Research Institute. For more information, contact Joe Fohn.
