masc seminar: validation of volumetric contact dynamics models

MotivationModel

ExperimentsConclusions

Validation of Volumetric Contact DynamicsModels

Mike Boos

May 11, 2011

Mike Boos Validation of Volumetric Contact Dynamics Models

MotivationModel


Outline

1 Motivation

2 ModelVolumetric modelNormal forcesFriction forces

3 ExperimentsNormal forcesFriction forcesExperimental apparatus

4 Conclusions


MotivationModel


Motivation

Figure: Dextre at the tip of Canadarm2 [1].


MotivationModel


Contact Models

Electrical Connectors

Alignment Sleeve

Alignment Pins

Micro Fixture

Coarse Alignment Bumper

36"28"

12"

Battery WorksiteBattery

Worksite

SPDM OTCM

Figure: ISS battery box [1].

Point contact models

Small contact patches only

Simple, convex geometries

No rolling resistance,spinning friction torque

FEM

Too complex for real-time


MotivationModel


Contact Models

Falling ISS battery box:real-time

Point contact models

Small contact patches only

Simple, convex geometries

No rolling resistance,spinning friction torque

FEM

Too complex for real-time


MotivationModel


Goals

1 Experimentally validate the volumetric contact dynamicsmodel for hard-on-hard (metal) contact

2 Demonstrate parameter identification for the model


MotivationModel


Volumetric modelNormal forcesFriction forces

Outline

1 Motivation



4 Conclusions


MotivationModel



Volumetric model

fN

kv

Figure: The modified Winkler elastic foundation model [1].


MotivationModel



Volumetric properties

nj

Kw

fs,j(s)

fs,i(s)

Contact Surface S

Contact Plate

ni

s

Bi

Bj

Figure: The contact surface between two deformable bodies [1].

Volumetric properties

V - volume of interference Js - surface-inertia tensorn - contact normal Jv - volume-inertia tensor


MotivationModel



Normal forces

Normal force

fn = kvV (1 + avcn)nn

V

S

vcn


MotivationModel



Rolling resistance

Rolling resistance torque

τ r = kv aJs · ωtn

V

S

ωt


MotivationModel



Friction

f t Friction force

τ s Spinning friction torque

7-parameter model

Bristle stiffness and damping (σo,σ1)

Slip-stick effects (µS, µC, vS)

Dwell-time dependency (τdw)

Viscous damping (σ2)

fN

Contact sites

Figure: Surface asperities(‘bristles’) in contact [1].


MotivationModel



The Contensou effect

Translational friction forcestend to ‘cancel out’ as angularvelocity increases.

C

A Bv

ωvA

vB

vC

vDD

ω r

ω r

ω r

ω r

Figure: v << ωr [2]


MotivationModel





C

AB

v

ωvA

vB

vC

vDD

ω r

ω r

ω r

ω r

Figure: v >> ωr [2]


MotivationModel





C

A B

v

ωvA

vB

vC

vDD

ω r

ω r

ω r

ω r

Figure: v ' ωr [2]


MotivationModel



Volumetric contact model

Ball-table simulation: real-time


MotivationModel


Normal forcesFriction forcesExperimental apparatus

Outline

1 Motivation



4 Conclusions


MotivationModel



Contact geometries

Focus on simple geometric pairs:

Cylinder-on-plane

Sphere-on-plane


MotivationModel



Normal force experiments

fN = kvV

Volumetric stiffness

Increase force on payload quasi-statically

Measure normal forces and displacements(to calculate volume of interference)

Damping

Drive the payload into contact plate at setvelocities

Measure forces and displacements overtime


MotivationModel



Normal force experiments

fN = kvV (1 + avcn)

Volumetric stiffness

Increase force on payload quasi-statically

Measure normal forces and displacements(to calculate volume of interference)

Damping

Drive the payload into contact plate at setvelocities

Measure forces and displacements overtime


MotivationModel



Translation

Static friction and bristle dynamics

1 Begin with payload at rest

2 Slowly increase force until slipping occurs

Peak force can be used to estimate µS:

ft

t

f µN S

Also, σo =µSδz


MotivationModel



Translation

ft ≈ fn(µC + σ2vt)

Coulomb friction and viscous damping

Drive payload at different constantvelocities

ft

f µN C

vt

f σN 2


MotivationModel



Translation

τdw ≈tpeak − tstopln( µS−µC

µS−µpeak )

Dwell-time dependency

Static friction: Bonds between surfacesform over time when at rest.

1 Drive payload at a constant velocity

2 Bring to a stop for a set period of time

3 Slowly increase force until slippingoccurs

4 Repeat, increasing the dwell time,until no change in peak force detectedbetween iterations


MotivationModel



Rotation

Repeat translation experiments, rotatinginstead of translating

Compare parameters for translation androtation


MotivationModel



Combined translation and rotation

Contensou effect

1 Drive at constant tangential velocity withincreasing angular velocity

2 Drive at constant angular velocity withincreasing tangential velocity

3 Using parameters identified in previousexperiments, model the friction forces tocompare with observed Contensou effect


MotivationModel



Normal force configuration


MotivationModel



Friction configuration


MotivationModel



Apparatus


MotivationModel


Outline

1 Motivation



4 Conclusions


MotivationModel


Conclusions

Volumetric contact dynamics model discussed

Experimental procedure developed for parameter identificationand validation

Design of experimental apparatus


MotivationModel


References

Y. Gonthier.Contact Dynamics Modelling for Robotic Task Simulation.PhD Thesis, University of Waterloo, 2007.

Y. Gonthier, J. McPhee, C. Lange.On the Implementation of Coulomb Friction in aVolumetric-Based Model for Contact Dynamics.Proceedings of IDETC’07, 2007.


MotivationModel


Research supported by


MotivationModel


Questions


masc seminar: validation of volumetric contact dynamics models

Technology

friction torquefemtoo

convex geometriesno

interference js surface

parameter identification

inertia tensorn

normal jv volume

hard metal contact2

damping o