Modeling of asphalt and experiments with a discrete particles method

(1)

V. Magnanimo, H. ter Huerne, S. Luding & T. Ormel

Tire & Road Consortium, CTW, University of Twente

Asphalt durability and self healing

modeling with DEM approach

(2)

Materials with Microstructure

Single particle Contacts Many particle simulation Continuum Theory

Flow in a silo

Salt

Grain

(3)

Single particle Contacts Many particle simulation Continuum Theory

Asphalt samples

Materials with Microstructure

Can we model this material behaviour giving

(4)

1. loading

Luding, S. 2008. Granular Matter 10.

(5)

1. loading

PLASTIC loading

stiffness: k

₁

Elasto-plastic adhesive contact model

(6)

1. loading 2. unloading *

elastic un/re-loading

stiffness: k

PLASTIC loading

stiffness: k

₁

Elasto-plastic adhesive contact model

(7)

1. loading 2. unloading 3. re-loading

PLASTIC loading

stiffness: k

₁ *

elastic un/re-loading

stiffness: k

Elasto-plastic adhesive contact model

(8)

1. loading 2. unloading 3. re-loading 4. tensile failure *

elastic un/re-loading

stiffness: k

tensile force

PLASTIC loading

stiffness: k

₁

Elasto-plastic adhesive contact model

(9)

1. loading 2. unloading 3. re-loading 4. tensile failure

transition to ELASTIC

stiffness: k

₂

max. tensile force

ELASTIC un/re-loading

stiffness: k

₂

Elasto-plastic adhesive contact model

(10)

Discrete element model (DEM) of Asphalt

• First approach simple 3d model • Spherical particles

• Not modeling mastic as particles (Bitumen, fillers, fine aggregate) • Mastic present in contact model (the way particles interact)

• 18 model parameters in total, important parameters:

Normal Contact force:

Loading/ unloading stiffness (K1/ K2) Phi (parameter accounts for of plasticity)

Tangential Contact Force

(11)

Animation Uniaxial Loading

Color scale =

kinetic energy (motion)

(12)

Experimental Uniaxial loading

• Mold with asphalt • Uniaxial compaction • Measuring Force and

displacement • 2 Types of bitumen (link to parameters) Mould Stamp Asphalt Force measuring Bottom Plate

(13)

Samples after Experimental Uniaxial loading

Oil sample

Bitumen sample

(14)

Results (Experimental)

0 0.05 0.1 0.15 0.2 0 2 4 6 x 10 6



_vol [-]



[ P a ] Oil sample

Bitumen sample

Porous

asphalt

Uniaxial

(15)

Results (Experimental + DEM)

Porous asphalt

0 0.05 0.1 0.15 0.2 0 2 4 6 x 10 6



_vol [-]



[ P a ] k 1 = 10 =0.1 =0.2 =0.3 =0.4 Oil Bit

Phi=0.1

K1/K2 = 0.1

Loading stiffness

K1 to low

(16)

Results (Experimental+DEM)

Changing k1 or μ.

Both give good fit!

0

0.05

0.1

0.15

0.2

0

2

4

6 x 10

6  vol

[-]



[

P

a

]

Oil

k

1

=20



=0.17

k

1

=22



=0.1

(17)

1. Preparation 2. HIGH pressure 3. Relaxation 4. Compression 5. Tension 6. Healing

healing (tension, DEM)

(18)

Conclusions DEM

• Simple DEM Model fits the experiments • Modeling of self healing is possible

• Parameters influencing the fit: -Friction (Scaling the curves)

-Phi (Length of region 1, based on experiments) -K1 Dominant parameter

(19)

Continuum model

• Goal is to link DEM (micro-scale) to continuum model (Cam Clay) which can be used on the macro-scale

• Cam Clay model is implemented in abaqus, needs calibration:

• - Hydrostatic compression tests:

• Hardening behavior

• - Triaxial tests (not performed yet):

• Slope critical state line (M) • Cap shape (β)

(20)

(21)

Example Cam clay

0 0.05 0.1 0.15 0 1 2 3 4 5 6 x 10 6  vol [-]  [ P a ] Bitumen DEM

Cam Clay (Abaqus)

Comparison of

compaction:

1. Experiments, 2. DEM,

3. Cam Clay (FEM).

Hardening: prescribed

stress belonging to

(22)

Conclusions

• Cam Clay model is capable modelling the from the

experiments

• Better fit with results from triaxial tests? (not

performed yet)

• The linking between micro- (DEM) and macro-scale

(continuum) is realised.

(23)