1 Computational Methodology for the Prediction of Functional Requirement Variations Across the...

25
1 Computational Methodology for the Prediction of Functional Requirement Variations Across the Product Life-Cycle Guillaume Mandil Alain Desrochers Alain Rivière

Transcript of 1 Computational Methodology for the Prediction of Functional Requirement Variations Across the...

Page 1: 1 Computational Methodology for the Prediction of Functional Requirement Variations Across the Product Life-Cycle Guillaume Mandil Alain Desrochers Alain.

1

Computational Methodology for the Prediction of Functional Requirement Variations Across the Product Life-Cycle

Guillaume Mandil

Alain Desrochers

Alain Rivière

Page 2: 1 Computational Methodology for the Prediction of Functional Requirement Variations Across the Product Life-Cycle Guillaume Mandil Alain Desrochers Alain.

2

Problem

Parts of Mechanisms are generally specified for the assembly stage of their life cycle

Useful values of Functional Requirements are usually taken under other conditions

How to link theses values ?

Page 3: 1 Computational Methodology for the Prediction of Functional Requirement Variations Across the Product Life-Cycle Guillaume Mandil Alain Desrochers Alain.

3

Definitions and concepts

Nominal dimension : A Tolerance : [al;au] Mean dimension :

Dimension chain : j1 = B-A Calculation of functional

requirement values

aualA

1j

bublB

A = A +au + al

2

Page 4: 1 Computational Methodology for the Prediction of Functional Requirement Variations Across the Product Life-Cycle Guillaume Mandil Alain Desrochers Alain.

4

Sources of functional requirement variations

Uncertainties due to Tolerances stack-up : analysis of tolerance zones made thanks to existing techniques

Changing environment (variation of mechanical load or temperature) : Elastic deformation of parts.

Page 5: 1 Computational Methodology for the Prediction of Functional Requirement Variations Across the Product Life-Cycle Guillaume Mandil Alain Desrochers Alain.

5

Functional requirements variations across the life-cycle Elastic deformation

Variation of tolerance zone width is insignificant relatively to mean dimension variation.

Aalau AAalau

)1(1 Sj)1(SA )2(SA

)2(SB)1(SB

)2(1 Sj

Page 6: 1 Computational Methodology for the Prediction of Functional Requirement Variations Across the Product Life-Cycle Guillaume Mandil Alain Desrochers Alain.

6

Functional requirements variation across life-cycle Width of uncertainty for Functional

Requirement is not varying along life cycle Loads variations affect the mean value of

the Functional Requirement

Tolerance analysis/synthesis made once at the initial stage.

Variations due to the changing environment are evaluated on the mean values

Page 7: 1 Computational Methodology for the Prediction of Functional Requirement Variations Across the Product Life-Cycle Guillaume Mandil Alain Desrochers Alain.

7

Functional requirements variation across life-cycle

Life-cycle stage

Stage “Si”

Stage “Sf”

Value ofFunctional Requirement

- 0 +

Interference possible motion

j1

Mean value

j1

Page 8: 1 Computational Methodology for the Prediction of Functional Requirement Variations Across the Product Life-Cycle Guillaume Mandil Alain Desrochers Alain.

8

Three kind of calculations

Dependent of known variables Dimension driven

What will be the value of a given functional requirement after thermal dilatation of the parts?

Functional requirement drivenWhich dimension has to be chosen in order to obtain a given value of a functional requirement after thermal dilatation?

Geometry drivenWhich loads are acceptable in order to ensure the respect of a functional requirement at two stages Si and Sf of the life-cycle?

Page 9: 1 Computational Methodology for the Prediction of Functional Requirement Variations Across the Product Life-Cycle Guillaume Mandil Alain Desrochers Alain.

Example of FR management along the product life cycle

9

Check of initial Design (Dimension Driven)

Meet Final Functional Requirements

Product validated

no

Yes

Redesign to fit final requiments (Functional Requirement Driven)

Meet Initial Functional Requirements

Yes

no

Calculation of Acceptable Load variation (Geometry Driven)

Page 10: 1 Computational Methodology for the Prediction of Functional Requirement Variations Across the Product Life-Cycle Guillaume Mandil Alain Desrochers Alain.

10

Simple application case :Geometry

e1 e2 e3

j1 j3j2

b2 b3b1

Presentation of the case studied and corresponding dimensions chains

j1= b1− e1

j2 = e2 − b2

j3 = b2 + b3 − e2 − e3

⎨ ⎪

⎩ ⎪

Page 11: 1 Computational Methodology for the Prediction of Functional Requirement Variations Across the Product Life-Cycle Guillaume Mandil Alain Desrochers Alain.

11

Simple application case :Hypothesis Stepping from “Si” to “Sf” life-cycle stages Wheel shaft made of Aluminium Frame made of steel Dimension known at 20°C Deformation of parts due to thermal

dilatation only Design variables : dimensions of the frame

Page 12: 1 Computational Methodology for the Prediction of Functional Requirement Variations Across the Product Life-Cycle Guillaume Mandil Alain Desrochers Alain.

12

Uncertainties onFunctional Requirements

For all dimensions tolerance zones are 0.2mm width

Uncertainties on Functional Requirements are deduced thanks to dimension chains relationj1 has a 0.4mm width uncertainty zonej2 has a 0.4mm width uncertainty zonej3 has a 0.8mm width uncertainty zone

Page 13: 1 Computational Methodology for the Prediction of Functional Requirement Variations Across the Product Life-Cycle Guillaume Mandil Alain Desrochers Alain.

13

Calculation 1Dimension driven

Life-cycle stage

Stage “Si” @ 20°C

Stage “Sf”@ 50°C

Value of j1- 0 +

Inte

rfere

nce

Mot

ion

Poss

ible

Value of j2- 0 +

Inte

rfere

nce

Mot

ion

Poss

ible

Value of j3- 0 +

Inte

rfere

nce

Mot

ion

Poss

ible

0.05

0.450.25

0.079

0.4790.279

0.2

0.60.4

0.610

1.0100.810

0.05

0.850.45

-0.431

0.369-0.031

Page 14: 1 Computational Methodology for the Prediction of Functional Requirement Variations Across the Product Life-Cycle Guillaume Mandil Alain Desrochers Alain.

14

Calculation 2Functional Requirement driven

Life-cycle stage

Stage “Si” @ 50°C

Stage “Sf”@ 20°C

Value of j1- 0 +

Inte

rfere

nce

Mot

ion

Poss

ible

Value of j2- 0 +

Inte

rfere

nce

Mot

ion

Poss

ible

Value of j3- 0 +

Inte

rfere

nce

Mot

ion

Poss

ible

0.05

0.450.25

0.071

0.4710.271

0.1

0.50.3

-0.310

0.09-0.110

0.1

0.90.5

0.581

1.3810.981

Page 15: 1 Computational Methodology for the Prediction of Functional Requirement Variations Across the Product Life-Cycle Guillaume Mandil Alain Desrochers Alain.

15

Calculation 3Geometry driven

Life-cycle stage

Stage “Si” @ 20°C

Stage “Sf”

Allowable final temperature

Value of j1- 0 +

Inte

rfere

nce

Mot

ion

Poss

ible

Value of j2- 0 +

Inte

rfere

nce

Mot

ion

Poss

ible

Value of j3- 0 +

Inte

rfere

nce

Mot

ion

Poss

ible

0.1

0.50.3

0.05

0.450.25

0.1

0.50.3

0.2

0.60.4

0.1

0.90.5

0.05

0.850.45

91.0°C 25.9°C 22.8°C

Page 16: 1 Computational Methodology for the Prediction of Functional Requirement Variations Across the Product Life-Cycle Guillaume Mandil Alain Desrochers Alain.

16

ConclusionPerspectives High-level management of Functional

Requirement along life-cycle.

Need of an extension to 3D Results from Finite Elements calculation

used to quantify dimension variations Use of TTRS/MGRE to obtain relations

between individual dimension.

Page 17: 1 Computational Methodology for the Prediction of Functional Requirement Variations Across the Product Life-Cycle Guillaume Mandil Alain Desrochers Alain.

Further Work

Use of a parametric representation (vectorisation) for 3D mechanism. [Serré] (TTRS / MGRE)

Dimension chains viewed as loop of vectors

Deformations viewed as variations of representation vectors

17

Page 18: 1 Computational Methodology for the Prediction of Functional Requirement Variations Across the Product Life-Cycle Guillaume Mandil Alain Desrochers Alain.

Further Work

Use of a deformed mesh to deformed BRep transfer. [Louhichi]

Association of the deformed mechanism to an ideal and FR compatible “neighbour”.

Calculation of the distance between deformed and associated parameterisation vectors

deduction of minimal functional requirement [Serré]

18

Page 19: 1 Computational Methodology for the Prediction of Functional Requirement Variations Across the Product Life-Cycle Guillaume Mandil Alain Desrochers Alain.

19

Computational Methodology for the Prediction of Functional Requirement Variations Across the Product Life-Cycle

Guillaume Mandil

Alain Desrochers

Alain Rivière

Page 20: 1 Computational Methodology for the Prediction of Functional Requirement Variations Across the Product Life-Cycle Guillaume Mandil Alain Desrochers Alain.

Discussion

20

Modèle Géométrique CAO

Spécifications Fonctionnelles (GPS)

Paramètres de conception

Simulation Éléments Finis

Paramètres de calcul

Maillages résultats

SATT / EGRM

Modèle CAO

Déformé

Vectorisation

Page 21: 1 Computational Methodology for the Prediction of Functional Requirement Variations Across the Product Life-Cycle Guillaume Mandil Alain Desrochers Alain.

21

Page 22: 1 Computational Methodology for the Prediction of Functional Requirement Variations Across the Product Life-Cycle Guillaume Mandil Alain Desrochers Alain.

22

Discussion :Contribution du LISMMA?

Utilisation des relations de dépendance en 3D comme équation pour caractériser des conditions fonctionnelles.

Pour les mécanismes iso-statiques ?

Page 23: 1 Computational Methodology for the Prediction of Functional Requirement Variations Across the Product Life-Cycle Guillaume Mandil Alain Desrochers Alain.

23

Three kind of calculations

Choice of functional requirement under study and extraction of the corresponding dimension chain

Extraction of individual dimensions along dimension chain

Calculation of mechanical deformations

Integration of calculated deformations in the corresponding

mean dimension

Calculation of final functional requirement value with dimensions

updated values

Comparison of the results with designer intent or with specifications

at final stage.

Choice of functional requirement under study and extraction of the

corresponding dimension chain

Specification of initial functional requirement

Distribution of functional requirement : deduction of initial dimension values

Calculation of parts deformations

Deduction of final values for individual dimensions

Optional calculation of functional requirement at final stage

Comparison of the results with designer intent or with specifications

at final stage.

Choice of functional requirement under study and

extraction of the corresponding dimension chain

Specification of initial and final values for functional

requirement

Distribution of functional requirement: deduction of initial

dimension values

Calculation of thermo-mechanical load variations from

initial to final condition

Optional Calculation of final dimensions

Dimension driven Functional requirement driven Geometry driven

Page 24: 1 Computational Methodology for the Prediction of Functional Requirement Variations Across the Product Life-Cycle Guillaume Mandil Alain Desrochers Alain.

24

Calculation 1Dimension driven

Hypothesis

ti = 20°C

tf = 50°C

e1 = at 20°C

e2 = at 20°C

e3 = at 20°C

b1 = at 20°C

b2 = at 20°C

b3 = at 20°C

Results

at 50°C

at 50°C

at 50°C

at 50°C

at 50°C

at 50°C

j1 =[0.079 ; 0.479] mm at 50°C

j2 =[0.610 ; 1.010] mm at 50°C

j3 =[-0.431 ; 0.369] mm at 50°C

What will be the value of a given functional requirement after thermal dilatation of the parts?

mm1,060

mm1,01440

mm1,060

0,160.3 mm

0,11439.7 mm

0,160.8 mm

1 60.043e mm2 1441.028e mm3 60.043e mm1 60.322b mm2 1440.218b mm3 60.822b mm

Page 25: 1 Computational Methodology for the Prediction of Functional Requirement Variations Across the Product Life-Cycle Guillaume Mandil Alain Desrochers Alain.

25

Calculation 2Functional requirement driven

0,160.893 mm

Hypothesis

ti = 50°C

tf = 20°C

j1 =[0.05 ; 0.45] mm at 50°C

j2 =[0.2 ; 0.6] mm at 50°C

j3 =[0.05 ; 0.85] mm at 50°C

e1 = at 50°C

e2 = at 50°C

e3 = at 50°C

b1 = at 50°C

b2 = at 50°C

b3 = at 50°C

Results

at 20°C

at 20°C

at 20°C

at 20°C

at 20°C

at 20°C

j1 =[0.071 ; 0.471] mm at 20°C

j2 =[-0.310 ; 0.09] mm at 20°C

j3 =[0.581 ; 1.381] mm at 20°C

Which dimension has to be chosen in order to obtain a given value of a functional requirement after thermal dilatation?

0,11440.628 mm

0,160.293 mm

0,160.043 mm

0,160.043 mm

0,11441.028 mm

2 1440.109b mm1 60.27b mm

3 60.87b mm

3 60e mm

1 60e mm2 1440e mm

0,160.893 mm