Validation & Verification of automated non- linear ... · aethereng.com Validation & Verification of automated non-linear modelling for the seismic response of cross laminated timber

aethereng . com

Validation & Verification of automated non-linear modelling for the seismic response of

cross laminated timber structures

Lorenzo Riparbelli

Ioannis P. Christovasilis

- Par i s , 20 .03 .2018 -

aethereng . com

The Big Picture The Professional use

ImportCAD Geom.

Create ModelGeometry and Mesh

Perform Linear Analyses

To determine sections,To design for regulatory codes

Perform NON-LINEAR analyses.

OptimizationPerformance-based DesignResilience-based Design

All with AUTOMATED Procedures

aethereng . com

Our Strange Friend:Wood

▪ Composite layout

▪ Different bending stiffness in the two maindirections and planar stability

▪ Use both as walls and floors

CLT Panel

Anisotropy

aethereng . com

Global Behaviour dominated byContact and Discrete Connections

Seismic BehaviourRocking (rotation) and Sliding for each panel

Monolateral contact

Tension-onlyhold-down

Shear-only discrete + Friction

aethereng . com

Global Non-linear BehaviourConnectors:o No Resistance in compressiono Pinching Response in tension/shearPanels:o Monolateral contact

Uplift

Link Video: https://www.dropbox.com/s/7y6nez1yewim94d/Deformed_C4-Kobe-082.mp4?dl=0

aethereng . com

Modelling Approach

aethereng . com

Modelling Approach

Wall Face Wall-to-Wall Contact Face

Wall-to-Floor Contact Face

Floor Face

aethereng . com

Modelling Approach

Wall Assembly

Floor Assembly

Wall Assembly

Wall Assembly

Wall Assembly

Coincident nodes

1st generation model

2nd generation model

▪ LIASON GROUP on DX, DY, DZ

▪ Contact face material properties

based on connections

▪ Contact Zones

▪ Formulation CONTINUOUS

aethereng . com

AB

AB

Angle Bracket (AB)

K_T_D_L springs with shear stiffness

Connections

aethereng . com

WFC

Connections

WC / WFC / FCK_T_D_L springs with axial and shear stiffness

WC

WC

aethereng . com

HD

Connections

Hold-down (HD)K_T_D_L springs

with axial stiffness

HD

aethereng . com

Connections

Static non-linear analysis

▪ K_T_D_L spring with axial stiffness and

kinematic hardening (DIS_ECRO_CINE)

▪ CABLE element

Dynamic non-linear analysis

▪ K_T_D_L spring with axial stiffness and

kinematic hardening (DIS_ECRO_CINE)

▪ K_T_D_L spring with axial stiffness and

asymmetric hardening (ECRO_ASYM_LINE)

HD as two elements in series

aethereng . com

Sofie Project Real Scale Test on CLT Structure

CNR-IVALSA(Italian National Research Council – Trees and Timber Institute)

Scientific Director: Prof. Ing. Ario Ceccotti

Funded by: Provincia Autonomadi Trento

In collaboration with:

aethereng . com

The model

o 449 faces (61 CLT panels)

o 204 common edges with contact&friction

o 1543 discrete elements, thatrepresent 9 different types of connectors

o Panels are connected with 211 different connections

aethereng . com

The model

All groups and contact zonescreated automatically

aethereng . com

Validation Process

o Linear Elastic Model

o Non-linear with contact without friction

o Non-linear with contact and friction =0.2

o Non-linear with contact and friction =0.4

Subjected to

o Kobe JMA motion scaled to 0.15g

o Cyclic Pushover analysis

Validation based on:

o Hysteresis loops

o Time-history of base shear and drift

o Vibration periods

aethereng . com

Linear model

modeltest

+16.8

+34.7

+115.7

+98.0

aethereng . com

Non-linear modelContact without friction | µ = 0.0

modeltest

+13.3

+40.2

+139.0

+130.5

aethereng . com

modeltest

-4.7

-2.0

+37.2

+17.3

Non-linear modelContact with friction | µ = 0.2

aethereng . com

modeltest

-0.6

+3.3

+24.7

+9.3

Non-linear modelContact with friction | µ = 0.4

aethereng . com

Phenomenological Difference for low acceleration

With LiasonsNo-Separation

With Contact &Friction

aethereng . com

Comparison

model-µ=0.0test

model-µ=0.4linear model

modelµ=0.0

modelµ=0.2

modelµ=0.4

Max Shear

+16.8% +13.3% -4.7% -0.6%

Min Shear

+34.7% +40.2% -2.0% +3.3%

Max Drift

+115.7% +139.0% +37.2% +24.7%

Min Drift

+98.0% +130.5% +17.3% +9.3%

aethereng . com

Comparison

model-µ=0.0linear model

model-µ=0.4

0.40F

0.48F

0.12F

0.45M

0.45M

0.10M

aethereng . com

Comparison

model-µ=0.0linear model

model-µ=0.4

Link Video: https://www.dropbox.com/s/bvjdi3hmn5j6t1q/Cyclic_C7_High.mp4?dl=0

aethereng . com

Comparison

0.40F

0.48F

0.12F

0.45M

0.45M

0.10Mlinear model

modelµ=0.0

modelµ=0.2

modelµ=0.4

K1[kN/mm]

86.5 84.0 116.0 120.0

K2[kN/mm]

42.9 39.5 56.4 59.4

K3[kN/mm]

15.2 14.1 20.5 22.0

K1

K2

K3

T1 model[sec]

0.2080.213

0.214 0.180 0.176

T1 test[sec]

0.183 0.183 0.183 0.183

Perc. Diff. +13.7%+16.4%

+16.9% -1.6% -3.8%

And for Professional Application??

aethereng . com

Fields of Verification

1. Combined axial (compression or tension) and moment-induced stresses in the longitudinal direction of the panel (grain direction of external layer);

2. Shear stresses for perpendicular-to-plane shear forces in the longitudinal direction of the panel (grain direction of external layer);

3. Combined axial (compression or tension) and moment-induced stresses in the transverse direction of the panel (perpendicular-to-grain direction of external layer);

4. Shear stresses for perpendicular-to-plane shear forces in the transverse direction of the panel (perpendicular-to-grain direction of external layer);

5. In-plane shear stresses as maximum between gross, net and torsional shear.

aethereng . com

Evolutionary Field of Verification (Transient)

Link Video: https://www.dropbox.com/s/7y6nez1yewim94d/Deformed_C4-Kobe-082.mp4?dl=0

Link Video: https://www.dropbox.com/s/bnwcd9tfkxqq5g1/Trans_C4-Kobe-082.mp4?dl=0

aethereng . com

Evolutionary Field of Verification (Cumulative)

Link Video: https://www.dropbox.com/s/n6k9e1row5m517c/Cumul_C4-Kobe-082.mp4?dl=0

Link Video: https://www.dropbox.com/s/7y6nez1yewim94d/Deformed_C4-Kobe-082.mp4?dl=0

aethereng . com

Conclusions

o Automated Structural Modelling for designing complex structures, usingcontact and friction, is indeed possible, thanks to the extraordinarycapabilities of code_aster

o Results of these non-linear analyses are better than linear also for accelerations assimilable to Serviceability Limit States, given the factthat the dynamic characteristics of the structure vary also for low accellerations

o These modelling methodologies can be extended to other structuralsystems (for structures where the performance or resilience is relevant)

Thank You!!