32
A Formal Methodology for Smart Assembly Design A Presentation by Kris Downey – Graduate Student Alan Parkinson – Faculty Member 15 June 2000 Acknowledgements to NSF Grant 0084880

A Formal Methodology for Smart Assembly Design A Presentation by Kris Downey – Graduate Student Alan Parkinson – Faculty Member 15 June 2000 Acknowledgements

  • View
    212

  • Download
    0

Embed Size (px)

Citation preview

Page 1: A Formal Methodology for Smart Assembly Design A Presentation by Kris Downey – Graduate Student Alan Parkinson – Faculty Member 15 June 2000 Acknowledgements

A Formal Methodology for Smart Assembly Design

A Presentation by

Kris Downey – Graduate Student

Alan Parkinson – Faculty Member

15 June 2000

Acknowledgements to NSF Grant 0084880

Page 2: A Formal Methodology for Smart Assembly Design A Presentation by Kris Downey – Graduate Student Alan Parkinson – Faculty Member 15 June 2000 Acknowledgements

Presentation Outline

Introduction Research ObjectivesCurrent design techniques and analysis

methodsCase studyCurrent status of researchConclusions

Page 3: A Formal Methodology for Smart Assembly Design A Presentation by Kris Downey – Graduate Student Alan Parkinson – Faculty Member 15 June 2000 Acknowledgements

Robust Design

Design that works properly when subjected to variation

Current robust design methodsFocus on key characteristicsFMEADOE by TaguchiSix sigma analysisOptimization techniques

Page 4: A Formal Methodology for Smart Assembly Design A Presentation by Kris Downey – Graduate Student Alan Parkinson – Faculty Member 15 June 2000 Acknowledgements

Smart Assembly

A smart assembly has features, not otherwise required by the function of the design, which allow the design to absorb or cancel out the effects of variation [Parkinson, 2000]

Page 5: A Formal Methodology for Smart Assembly Design A Presentation by Kris Downey – Graduate Student Alan Parkinson – Faculty Member 15 June 2000 Acknowledgements

Smart Assembly

Examples of smart assembly featuresSlotted holesSprings used for positioningScrew locatorsSliding locatorsShims

Page 6: A Formal Methodology for Smart Assembly Design A Presentation by Kris Downey – Graduate Student Alan Parkinson – Faculty Member 15 June 2000 Acknowledgements

Smart AssemblyTypes of smart assemblies

Shim

Screw

Inclined Plane

Passive

Page 7: A Formal Methodology for Smart Assembly Design A Presentation by Kris Downey – Graduate Student Alan Parkinson – Faculty Member 15 June 2000 Acknowledgements

Smart AssemblyTypes of smart assemblies

Passive

Can Opener Support

Page 8: A Formal Methodology for Smart Assembly Design A Presentation by Kris Downey – Graduate Student Alan Parkinson – Faculty Member 15 June 2000 Acknowledgements

Smart AssemblyTypes of smart assemblies

Passive

Car frame

Page 9: A Formal Methodology for Smart Assembly Design A Presentation by Kris Downey – Graduate Student Alan Parkinson – Faculty Member 15 June 2000 Acknowledgements

Smart AssemblyTypes of smart assemblies

Active

Scissors Gear

Page 10: A Formal Methodology for Smart Assembly Design A Presentation by Kris Downey – Graduate Student Alan Parkinson – Faculty Member 15 June 2000 Acknowledgements

Smart AssemblyTypes of smart assemblies

Active

Garage Door Roller Bearing

Page 11: A Formal Methodology for Smart Assembly Design A Presentation by Kris Downey – Graduate Student Alan Parkinson – Faculty Member 15 June 2000 Acknowledgements

Research Objectives

Create a smart feature implementation methodologyPrinciples developed for methodologyProven analysis methods implemented into

methodology

Apply methodology to case studies

Page 12: A Formal Methodology for Smart Assembly Design A Presentation by Kris Downey – Graduate Student Alan Parkinson – Faculty Member 15 June 2000 Acknowledgements

Current Design Techniques and Analysis Methods

Exact constraint designVery close relationship to smart

assembly designScrew theory

Constraint information inferred from analysis

Tolerance analysisCritical dimension information provided

Page 13: A Formal Methodology for Smart Assembly Design A Presentation by Kris Downey – Graduate Student Alan Parkinson – Faculty Member 15 June 2000 Acknowledgements

Exact Constraint Design

Degrees of freedom

6 in 3D space3 in 2D space

Page 14: A Formal Methodology for Smart Assembly Design A Presentation by Kris Downey – Graduate Student Alan Parkinson – Faculty Member 15 June 2000 Acknowledgements

Exact Constraint Design

ConstraintA mechanical connection between objects that

reduces the degrees of freedom of each object

2D constraints 3D constraints

Page 15: A Formal Methodology for Smart Assembly Design A Presentation by Kris Downey – Graduate Student Alan Parkinson – Faculty Member 15 June 2000 Acknowledgements

Exact Constraint DesignExactly one constraint for each degree

of freedom

No two constraints collinear No four constraints in a single plane No three constraints parallel No three constraints intersect at a point

No four constraints are parallel No four constraints intersect at a point No four constraints in the same plane

Rules for exact constraint design:

2D space 3D space

More mathematically-based theory is needed in this area

Taken from [Blanding, 1999] and [Skakoon, 2000]

Page 16: A Formal Methodology for Smart Assembly Design A Presentation by Kris Downey – Graduate Student Alan Parkinson – Faculty Member 15 June 2000 Acknowledgements

Exact Constraint DesignExactly one constraint for each degree

of freedom

Exactly constrained

Overconstrained

Underconstrained

Page 17: A Formal Methodology for Smart Assembly Design A Presentation by Kris Downey – Graduate Student Alan Parkinson – Faculty Member 15 June 2000 Acknowledgements

Exact Constraint Design

Conclusions:Exact constraint design very closely related to

smart assemblySmart assembly provides solution to

overconstrained designs

Page 18: A Formal Methodology for Smart Assembly Design A Presentation by Kris Downey – Graduate Student Alan Parkinson – Faculty Member 15 June 2000 Acknowledgements

Screw Theory

Allows determination of over- or underconstrained designs

Motion analysisPerformed on individual part of assemblyAssumes parts do not break contactEach joint type has distinct screwmatrices

Taken from [Ball, 1900], [Roth 1966], [Konkar, 1993], and [Adams, 1998]

Page 19: A Formal Methodology for Smart Assembly Design A Presentation by Kris Downey – Graduate Student Alan Parkinson – Faculty Member 15 June 2000 Acknowledgements

Screw Theory

Steps to constraint analysisTranslate joints into twistsFind reciprocal wrench of each twistmatrixUnion of wrenchesFind reciprocal twist of wrenchResultant twistmatrix

Page 20: A Formal Methodology for Smart Assembly Design A Presentation by Kris Downey – Graduate Student Alan Parkinson – Faculty Member 15 June 2000 Acknowledgements

Screw Theory Interpretation of resultant twistmatrix

and wrenchmatrixTwist rows = degrees of freedomWrench rows = overconstraints

Motions

ConstraintsY-Translation

Z-Translation

X-Rotation

Y-Rotation

Z-Rotation

Page 21: A Formal Methodology for Smart Assembly Design A Presentation by Kris Downey – Graduate Student Alan Parkinson – Faculty Member 15 June 2000 Acknowledgements

Screw Theory

Conclusions:Screw theory can determine if a design is over-

or underconstrainedLocation of idle degrees of freedom identified Information useful to smart assembly design

Page 22: A Formal Methodology for Smart Assembly Design A Presentation by Kris Downey – Graduate Student Alan Parkinson – Faculty Member 15 June 2000 Acknowledgements

Tolerance Analysis

Vector loop analysis (DLM)Sensitivity matrix

Results provide useful information regarding critical part dimensions

80% of variation in angle is attributed to dimension a

Conclusion: Sensitivity matrix determines preferred location of smart features

Page 23: A Formal Methodology for Smart Assembly Design A Presentation by Kris Downey – Graduate Student Alan Parkinson – Faculty Member 15 June 2000 Acknowledgements

Case Study

Baffle designOverconstrained in z-direction

Taken from [Kriegel, 1994]

Taken from Kodak design problem

Page 24: A Formal Methodology for Smart Assembly Design A Presentation by Kris Downey – Graduate Student Alan Parkinson – Faculty Member 15 June 2000 Acknowledgements

Case Study

Results:Less deflectionDoes not address

constraint problemTabs tear from baffle

Taken from [Kriegel, 1994]

Solution #1Reinforce baffle

and frame

Page 25: A Formal Methodology for Smart Assembly Design A Presentation by Kris Downey – Graduate Student Alan Parkinson – Faculty Member 15 June 2000 Acknowledgements

Case Study

Solution #2Double screw Smart feature

Results:Variation absorbedNo deflectionExpensive parts Increased assembly

costs

Taken from [Kriegel, 1994]

Page 26: A Formal Methodology for Smart Assembly Design A Presentation by Kris Downey – Graduate Student Alan Parkinson – Faculty Member 15 June 2000 Acknowledgements

Case Study

Solution #3Double-slotted tabs Smart feature

Results:Variation absorbedNo deflection “No cost” solutionMinimal assembly

costs

Taken from [Kriegel, 1994]

Page 27: A Formal Methodology for Smart Assembly Design A Presentation by Kris Downey – Graduate Student Alan Parkinson – Faculty Member 15 June 2000 Acknowledgements

Case Study

Conclusions:Worst case tolerance analysis contributed to

need of smart featuresOverconstrained design was not initially

recognizedSmart features allowed variation absorptionSome smart features are more expensive than

others

Page 28: A Formal Methodology for Smart Assembly Design A Presentation by Kris Downey – Graduate Student Alan Parkinson – Faculty Member 15 June 2000 Acknowledgements

Current Status of Research

Principles of smart assembly being considered and explored Implementation as nesting forcesAbsorption of tolerancesElimination of overconstraintUse in designs where redundant constraints

are necessary

Page 29: A Formal Methodology for Smart Assembly Design A Presentation by Kris Downey – Graduate Student Alan Parkinson – Faculty Member 15 June 2000 Acknowledgements

Smart Assembly PrinciplesSmart assembly features as nesting forces

Eliminate mechanical play and assembly stresses

Rigid constraint

Smart feature

Page 30: A Formal Methodology for Smart Assembly Design A Presentation by Kris Downey – Graduate Student Alan Parkinson – Faculty Member 15 June 2000 Acknowledgements

Smart Assembly PrinciplesTolerance absorption

When other methods are not sufficient

Part A

Part B

Part A

Part B

Total gap No gap

Spring

Page 31: A Formal Methodology for Smart Assembly Design A Presentation by Kris Downey – Graduate Student Alan Parkinson – Faculty Member 15 June 2000 Acknowledgements

Smart Assembly PrinciplesRedundant constraints become smart features

Eliminate assembly stresses

Page 32: A Formal Methodology for Smart Assembly Design A Presentation by Kris Downey – Graduate Student Alan Parkinson – Faculty Member 15 June 2000 Acknowledgements

ConclusionsResearch Objectives:

Create a smart feature implementation methodology

Principles inferred from existing designs: Smart features replace overconstraints Smart features absorb tolerances in exactly

constrained designs Smart features implemented as nesting forces

Proven analysis methods adapted for methodology Exact constraint design, screw theory and

tolerance analysis

Apply methodology to case studies Indirect absorption of variation