Microelectronics

1
Modeling and Optimization of Vehicle Drivetrain
Dynamic Performance Considering Uncertainty
Zissimos P. Mourelatos Associate
Professor Daniel Wehrwein Graduate Student
Objectives

• Model the powertrain and drivetrain in order to
  simulate dynamic vehicle performance.
• Optimize the drivetrain parameters for better
  fuel economy and wide open throttle acceleration.
• Use probabilistic design optimization to ensure
  performance criteria are met in the presence of
  uncertainty.

Vehicle Model Components
Driveline Model

• Bond Graph Method is used.
• Bond Graphing is an energy base, compact
  modeling technique that is modular and compatible
  with block diagrams.

Driveline Equivalent Block diagram
Design Under Uncertainty

• Deterministic Optimization Results
• Fuel economy increased 16
• 0-90 time improved 4.5
• 0-60 time improved 4.7

Simulation was performed in Simulink using the
equivalent block diagram of the Bond Graph model.
List of Constraints
Problem Formulation
Constraint Description
G1 (Quarter Mile Time) - 16.10
G2 -(0 to 30 Time) 2.4
G3 (0 to 30 Time) - 2.5
G4 (0 to 60 Time) - 7.89
G5 (0 to 90 Time) -18.41
G6 -(Gradeability) 22.256
G7 -(0 to 30 Towing Time) 5.29
G8 (0 to 30 Towing Time) 5.39
G9 (0 to 60 Towing Time) - 17.01
G10 - (Towing Gradeability) 9.57
G11 (Max Engine RPM) - 6000

• Probabilistic Optimization Results
• Fuel economy increased 11
• 0-90 time improved 3.9
• 0-60 time improved 4.5

2. Possibility-based optimization design