Upload
others
View
6
Download
0
Embed Size (px)
Citation preview
1
Statistical Optimization and Structural Analysis: Design of handmade custom Snowboards
Benoit CAILLAUD
4th Workshop on Structural Analysis of Lightweight Structures Innsbruck, 02.06.2016
2
Introduction
Materials & Structure, Geometry & Shape
1. Calibration of test feedback
Elaboration of questionnaires Overview of first results
2. Correlation of test results with board parameters
Direct Sensitivity Study and ranking of significant parameters Application: optimization of custom handmade snowboards
3. Optimization under Structural analysis
Ranking of significant parameters according to certain target outputs Board parameters – Analysis results – Test results
Conclusions
Agenda
3
±45° fabric UD Roving Input parameters:
- Wood species combinations - Thickness profile - Reinforcement & Matrix types - Layup (orientation, thickness, stacking sequence) - Material properties (E Modulus {El Et Glt}, Poisson ratios) - Tensile, Compessive, Flexural Strength {σe}, {σr} - Density, Toughness, Stability
UD Lin fiber Basalte UD Roving
Wood
Sidewalls
Composite reinforcements
Edges Layup
Thickness profile
Materials & Structure
4
Input parameters:
- L, Lm, Lcontact
- Lnose, Ltail, Lcamber
- Wnose, Wwaist, Wtail
- Stance - Setback - Hnose, Htail, Hcamber
- Curvatures: sidecut, camber, rockers
Dimensions
Camber profile
Sidecut shape
Inserts positions
0
10
20
30
40
-800 -600 -400 -200 0 200 400 600 800
Sidecut Radius [m]
Geometry & Shape
Portfolio
5
6
Calibration of test feedback
→ 25 output parameters including:
- Board behaviour (discipline related) - Board aspect (graphics, weight) - Terrain conditions (weather, snow quality)
→ 35 feedback results gathered over Winter 15/16
7
Calibration of test feedback
I’m gonna die I’m loving it Boring A lot of fun!
Very slippery Very grippy Too stiff Too soft
8
Sinks easily Floats easily
Calibration of test feedback
Too short Too long
- Quantify parameter influence (slope of fitted numerical model) → compare effects of different parameters input parameters must be standardized to their practical variation range
9
Correlation of test results
Direct Sensitivity Study:
- Compute Correlation Coefficients between two sets of {Inputs} and {Outputs} → determine which parameters are significant to which outputs
0
1
2
3
4
5
6
0 5 10 15 20
Taper [mm]
0
1
2
3
4
5
6
0 10 20 30 40 50
Bindings Setback [mm]
0
1
2
3
4
5
6
0 20 40 60
Nose Rocker [mm]
Nose float = f(Taper) Nose float = f(bindings Setback) Nose float = f(Nose Rocker)
CC = 0.258 Slope = 46,6.10-3
CC = 0.135 Slope = 8,3.10-3
CC = 0.061 Slope = 4,3.10-3
10
Correlation of test results
0
1
2
3
4
5
6
2.0E+05 4.0E+05 6.0E+05 8.0E+05 1.0E+06
"Ove
rall
Stif
fne
ss"
fee
dbac
k
Average Bending Stiffness Dx [N.mm]
"Overall stiffness" = f(Dx)
CC = 0.584
Feedback parameter « Global Stiffness » D
x [N
.mm
]
« Too soft » « Too stiff »
0
200000
400000
600000
800000
1000000
1200000
1400000
0.32 0.37 0.42 0.47
Dx
[N.m
m]
UD Fiberglass, layer thickness t0 [mm]
Dx = f(t0)
400000
500000
600000
700000
800000
900000
1000000
1100000
1200000
5.0 6.0 7.0 8.0
Dx
[N.m
m]
Core thickness tc [mm]
Dx = f(tc)
11
Structural Analysis Optimization
Sensitivity Study: computation of average bending Stiffnesses for different layups
Stacking sequence (symmetric)
Material Orientation Thickness (mean value)
Fiberglass ±45° t45 = 0,4mm
Fiberglass 0° t0 = 0,4mm
Wood core 0° tcore = 6,8mm
Fiberglass 0° t0 = 0,4mm
Fiberglass ±45° t45 = 0,4mm
Input parameters 3 (tcore, t0, t45)
Variation interval ±15%
Number of samples 2197 (all combinations) Theory of thin laminates, [ABD]-1 matrix
Average bending stiffness Dx [N.mm]
CC = 0.980 Slope = 2,6.105
CC = 0.130 Slope = 9,2.105
600000
650000
700000
750000
800000
850000
900000
2.8 2.85 2.9 2.95 3 3.05 3.1
Dx
[N.m
m]
M (0,35m²) [kg]
Optimization of structural thickness according to output Mass [kg]
Variation of CORE thickness only
Variation of UD PLY thickness only
Variation of 45° PLY thickness only
12
Structural Analysis Optimization
- Define optimization target parameter (Mass, Material Cost, Time…) - Assess variation of single parameters around the deterministic sample - Sort the parameters by influence ranking
Deterministic sample
13
Conclusions
Design of a custom snowboard:
1) Define the board specifications: for whom? for which activity?
2) Pick up the best fit among all existing shapes, based on rider profile and feedback results
3) Run sensitivity study with the chosen existing board as "mean value"
4) Optimization of geometrical and structural parameters using the study results:
- sort significant input parameter by influence level (descending) - define priorities and « best compromise »
Further steps: - Expand number of samples and feedback results - Structural analysis via FE modelization: consideration of field inputs and selected load cases
- Qualification of material data - Consideration of physical inputs correlation - Fitting of nonlinear numerical models
Thanks!
www.baguetteboards.com