1. ME6105 05/01/2008 Hydro-Mechanical Hydraulic Drive Train Hydro-Mechanical Hydraulic Hybrid Drive Train Chen Chu Koji Intlekofer Vigneshwar Kalyanasundaram

2. ME6105 05/01/2008 Hydro-Mechanical Hydraulic Drive Train Background The power density of hydraulic pumps/motors and accumulators is greater than electrical components Hydraulic components are inexpensive when compared to electrical components Allows for independent torque control of the each wheel Hydraulic components continue to improve

3. ME6105 05/01/2008 Hydro-Mechanical Hydraulic Drive Train Parallel Hybrid Drive Train

4. ME6105 05/01/2008 Hydro-Mechanical Hydraulic Drive Train Dymola Model

5. ME6105 05/01/2008 Hydro-Mechanical Hydraulic Drive Train Significant Components Engine Torque Output Wheel and Tire Radius / Rolling Resistance Hydraulic Pumps – Leakage, Friction, and Displacement Hydraulic Accumulators – Volume and Pressure Speed and Torque Controller – Gain Values Car Mass

6. ME6105 05/01/2008 Hydro-Mechanical Hydraulic Drive Train Working with Dymola Engine Model

7. ME6105 05/01/2008 Hydro-Mechanical Hydraulic Drive Train Central Composite Results

8. ME6105 05/01/2008 Hydro-Mechanical Hydraulic Drive Train Meeting Criteria – EPA Profile Inaccurate losses result in failed models Successful model achieved, finally!

9. ME6105 05/01/2008 Hydro-Mechanical Hydraulic Drive Train Dealing with parameters Scripts can be used to divide up work when performing central composite to speed up computations.

10. ME6105 05/01/2008 Hydro-Mechanical Hydraulic Drive Train Optimizers Darwin Optimizer –Computationally expensive –4495 runs –Max utility = 0.832757 DOT Optimizer –19 runs –Max utility = 0.832752

11. ME6105 05/01/2008 Hydro-Mechanical Hydraulic Drive Train Elicitations Engine Size Car Mass Rolling Resistance

12. ME6105 05/01/2008 Hydro-Mechanical Hydraulic Drive Train Engine Performance Engine should turn off during braking. Most energy efficient model keeps engine running.

13. ME6105 05/01/2008 Hydro-Mechanical Hydraulic Drive Train Uncertainty - Latin Hypercube Starting ValuesOptimum Value Engine Size1.71.69984 Wheel Radius2.942.92708 Accumulator Size0.0380.037837 Engine Size2.52.50315 Wheel Radius22.0036 Accumulator Size0.0350.0350038 Engine Size1.3361.33707 Wheel Radius22.00403 Accumulator Size0.039743.97494

14. ME6105 05/01/2008 Hydro-Mechanical Hydraulic Drive Train Uncertainty - Latin Hypercube It always ends up where it started! Possible reasons: –Uncertainty in other variable dominates the influence of the variables so that there’s little difference between two sets of LHS. –Several maxima and minima appear when you consider uncertainty. Difficult to explore design space under uncertainty as it costs too much.

15. ME6105 05/01/2008 Hydro-Mechanical Hydraulic Drive Train Conclusions Models can quickly become very complex Issues with trusting models –Will trust a model that does not fail, but may be less accurate –Tendency not to trust models with accurate values but fails repeatedly. Elicitations are hard –Easy to know too much or too little

16. ME6105 05/01/2008 Hydro-Mechanical Hydraulic Drive Train Questions?