proper modeling of integrated - automotive research...

1

arc

Automotive Research Center Modeling of Integrated Vehicle Powertrain Systems

Geoff Rideout

Graduate Student Research Assistant

Automated Modeling LaboratoryUniversity of Michigan

Proper Modeling of IntegratedProper Modeling of IntegratedVehicle SystemsVehicle Systems

2

arc


¥ Jeff Stein, Professor UM

¥ Geoff Rideout, Graduate Student UM

¥ Loucas Louca, Research Fellow UM

¥ Dan Grohnke, Senior Development Engineer NAVISTAR

¥ Polat Sendur, Graduate Student UM

Modeling of Vehicle Modeling of Vehicle DrivelineDrivelineSystems for an IntegratedSystems for an Integrated

EnvironmentEnvironment

Project team - Driveline Systems:

3

arc


OutlineOutline

¥Motivation for integrated simulation tool - VESIM

¥Integrated simulation environment

¥Subsystem model implementation tools

¥Description of powertrain module models

¥Simulation results

¥Future work

4

arc


MotivationMotivation

¥ Integrated Ground Vehicle Simulation:

- Facilitates concurrent engineering of military,passenger and commercial vehicles

- Enables rapid study of alternate vehicle configurationsÈ system design and optimization studies

¥ driveability, fuel economy, emissions, etc.

¥Modular environment ensures long-termrelevance of simulation tool

5

arc


Goal: Integration ofGoal: Integration ofVariable Complexity ModelsVariable Complexity Models

t

RPM

MULTI-CYLINDERDIESEL ENGINE

EXHAUSTMANIFOLD

INTAKEMANIFOLD

INTER-COOLER

COMPRESSOR TURBINE

WA

ST

EG

AT

E

FUELSYSTEM

Air

FuelExhaustgas

W.

EMPIRICAL POINT-MASS

THERMODYNAMIC

MULTI-BODY

SIMPLIFIED

HIGH-FIDELITY

VEHICLE DYNAMICSENGINE

6

arc


Background and CurrentBackground and Current State-of-the-Art State-of-the-Art

¥ Integrated simulation packages have been developed to:- Study transient operation of engines and vehicles

- Predict dynamic response

- Assess alternate system configurations

¥ Existing software packages allow building-block modelconstruction

¥ Models have been developed for each subsystem, butintegration effort is fairly new

We have yet to fully tap the potential of such a tool.

7

arc


Integrated Modeling ToolsIntegrated Modeling Tools

MATLAB-SIMULINK

¥ Block diagram graphical user interface- Simulink block libraries

- Matlab programming language

- C or Fortran source code

¥ Common solver used for engine, driveline, vehicle

8

arc


20SIM

¥Bond graph model graphical user interface

¥Common formalism for different energy domains- Systematic generation of state equations

¥Bond graph and block diagram elements can beintegrated

¥Generates source code for Matlab C-MEX files

Integrated Modeling ToolsIntegrated Modeling Tools

9

arc


DrivelineDriveline System Schematic System Schematic

AAAAAAAAAA

AAAAAAAA

AAAAAAAA

AAAAAAAA

AAAAAAAA

AA

AA

AAA

AAA

AAA

AAA

Trns

Tr-C

D-F

AAAAAAAAAAAAAAAAAA

D-FR

D-R

IA-D

T C

MODULAR STRUCTURE

¥ flexibility in choosing vehicleconfiguration

10

arc


Torque ConverterTorque Converter

¥3 element torqueconverter (pump,turbine & stator)

¥Static or dynamicmodel?

TURBINE

STATOR CONCAVE SIDE

STATOR CONVEX SIDE

PUMP DRIVE HUB

IMPELLER

IMPELLER ROTATION DIRECTION

(J.C. WHITE/FORD MOTOR CO.)

11

arc


Torque Converter ModelTorque Converter Model

¥ Static Model Methodology (e.g., Salaani and Heydinger,1998)

¥ Implementation- 20SIM -> C code -> Matlab/Simulink

- Direct block diagram entry to Matlab/Simulink

Engine Speed

Turbine Speed Speed Ratio

Engine Speed

Capacity Factor

2

Torque Ratio

Capacity Factor

TurbineTorque(to trans.)

PumpTorque(to engine)

12

arc


Automatic TransmissionAutomatic Transmission

¥Modeling Strategies- Input-output model based on experimental data- Physical component-level model

INPUT TORQUE

OUTPUT SHAFT SPEED

TURBINE SPEED

TORQUE TO DIFFERENTIAL

TRANSMISSION

TURBINE SPEED TORQUE TO TRANS.

TORQUE CONVERTER

INPUT TORQUE INPUT SHAFT SPEED

DIFFERENTIAL

13

arc


Input-Output Transmission ModelInput-Output Transmission Model

¥ Based on experimental measurement of input and outputeffective inertias, stiffnesses, and damping

¥ Implemented as discrete components (e.g. Bond Graph)- 20SIM -> C-MEX files -> Matlab/Simulink

¥ Incorporates losses- Gear inefficiency

- Fluid pumping

INPUT INERTIA

OUTPUT INERTIA

GEAR REDUCTION

14

arc


Input-Output Transmission ModelInput-Output Transmission Model

Empirical blending functions are used for shifting.

Speed Ratio

Torque Ratio

Time

Shift Duration

15

arc


Alternative Transmission ModelAlternative Transmission Model

¥ Physical-Based Model (Choand Hedrick, 1989)

¥ Model considers inertias ofgears and compliances ofshafts

¥ Clutches engage/disengage during shiftevent- Torque phase

- Speed phase

¥ Model can switch betweenbond graphs for differentphases

TO 1-2 BAND

TO FINAL DRIVETO FIRST CLUTCH

TO SECOND CLUTCH

TO SECOND CLUTCH

TO 1-2 BAND

16

arc


Shift LogicShift Logic

¥ Decision to shift based on- Transmission output shaft

speed

- Throttle position

¥ Implementation- C-MEX function

È Accepts inputs fromtransmission and driver

È Outputs gear number, torquemultiplication and speedreduction ratios Output Shaft Speed [rad/s]

ThrottlePosition[0-1]

1 2 3 4

17

arc


Differential/Driveshaft ModelsDifferential/Driveshaft Models

¥ Differentials- Bond graph representations

È Conventional bevel-gear

È Worm-gear

¥ Propshafts and Driveshafts- Compliance, inertia, damping

- Axle cooler churning losses

¥ Implementation- 20SIM -> C-MEX files

WHEEL

WHEEL

DIFFERENTIALAUTOMATIC TRANS.

18

arc


Vehicle SpecificationsVehicle Specifications

¥ V8 DI Diesel

¥ Turbocharged, Intercooled

¥ Rated Power: 210 HP@2400 rpm

¥ GVWR: 7950 Kg

¥ 4 Speed Automatic Transmission

¥ Rear Wheel Drive - 4x2

Vehicle/DrivelineEngine

NAVISTAR 4700 Series

19

arc


Validation and Parametric StudyValidation and Parametric Study

¥Vehicle launch performance validation- Ò0-60 mphÓ full throttle acceleration from idle- Engine and vehicle speed compared to experimental data

¥Torque converter stall test validation- Compare torque converter stall speed with test data

¥Shift quality parametric study- Shift duration nominally 0.8 seconds- Vary duration +/- 0.4 seconds- Compare vehicle forward jerk

20

arc


0 5 10 15 20 25 30 350

500

1000

1500

2000

2500

3000

0 5 10 15 20 25 30 350

10

20

30

40

50

60

Model Validation (0-60 MPH)Model Validation (0-60 MPH)

ENG

INE

SPE

ED [

RPM

]

TIME [SEC]

TestVESIM

VEH

ICLE

SPE

ED [

MPH

]

TestVESIM

1st 2nd 3rd 4th Gear

21

arc


0 1 2 3 4 5 60

500

1000

1500

2000

2500

3000

3500

4000

Torque Converter Stall TestTorque Converter Stall Test

TIME [SEC]

ENG

. S

PEED

[RP

M]

TestVESIM

23

arc


Shift Quality - Vehicle JerkShift Quality - Vehicle Jerk

0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8-8

-6

-4

-2

0

2

4

6

8

NORMALIZED TIME [SEC]

FORW

ARD

JER

K [

M/

S^3

]

¥ Consider trade-offbetween ridequality and- Acceleration

- Clutch wear

0.4 sec shift0.8 sec shift1.2 sec shift

24

arc


SummarySummary

¥An inventory of models for different powertraincomponents has been implemented

¥Driveline models have been integrated withengine and vehicle (VESIM)

¥Models have been partially validated with testdata

¥Potential utility of simulation tool has beenexplored through example design studies

25

arc


ConclusionsConclusions

¥Vehicle/engine integration of varying fidelity ispossible

¥Engine/vehicle interactions appear to be importantto vehicle mobility design and evaluation

¥Additional work on driveline models is necessary

26

arc


Future WorkFuture Work

¥Extend usefulness of VESIM as predictive designtool by expanding component model inventory

¥Continue data collection effort for validation

proper modeling of integrated - automotive research...

Documents