Learning Forum

[htozam]

Respected members, 

In applying displacement on a reinforced concrete model I am recording the correspoinding resistance force amplitude for each of the displacement (0 - 40 mm amplitude) applied, the force - displacement result graph of the analysis is attached.

The question is how can the resistance force curve spike more than once? 

I mean in real-life situation once a resistance force is dropped (say if a sample is cracked or damaged) it will not increase again specially if we continue to apply higher displacement amplitudes on the model. 

Here it looks like the model is regaining strength at 4.e-3 mm (after the force drops from 37500 to approx. 6000 N) and again in 9 mm displacement amplitudes.  

Additionlly, if the number of Sub-steps are increased the spikings increase and the whole results graph changes. For examples attached is the force - displacement curve for 1 Sub - step and 5 Sub-steps manually selected. The number of displacement steps are 14. 

 

[Wenlong]

Hi, 

Before making any comments, it would be nice to how do you model your reinforced concrete? Is the reinforcement defined explicitly? If yes, What kind of contact did you specify between the reinforcement and concrete? What types of failure criteria do you use on the concrete? 

Bests,

Wenlong

 

 

 

 

[htozam]

Hi Wenlong, 

Thank you for reply. 

Yes, the reinforcement is defined explicitly. I used node merging to connect the reinforcement with concrete. 

The concrete properties is defined as following:

 

/PREP7 !N AND MM

 

* GET,NMAT0,MAT,0, NUM, MAX, , 

* GET,NELEM0,ETYPE,0, NUM, MAX,

* GET, Nreal0, RCON, 0, NUM, MAX,

 

 

NELEM1=NELEM0+5

NMAT1=NMAT0+1

Nreal1=Nreal0+5

 

! Define concxrete Element

ET,NELEM1,SOLID65

!*

KEYOPT,NELEM1,1,0

KEYOPT,NELEM1,3,0

KEYOPT,NELEM1,5,0

KEYOPT,NELEM1,6,0

KEYOPT,NELEM1,7,0

KEYOPT,NELEM1,8,0

 

fc=25

ft=fc/10

nu=0.25

E=0.468253*fc/0.000778

!Define concrete Material 

MPTEMP,,,,,,,,

MPTEMP,1,0

MPDATA,EX,NMAT1,,E

MPDATA,PRXY,NMAT1,,nu

 

 

TBDE,CONC,NMAT1,,,

TB,CONC,NMAT1,1,9

TBTEMP,0

TBDATA,,0.2,1,ft,fc,,

 

!Define concrete eral

R,Nreal1, , , , , , , 

RMORE, , , , , , , 

RMORE,0.01, 

 

ALLSEL,ALL

CMSEL,S,Concrete,ELEM

ALLSEL,BELOW,ELEM

EMODIF,ALL,TYPE,NELEM1

EMODIF,ALL,MAT,NMAT1

EMODIF,ALL,REAL,Nreal1,

ALLSEL,ALL

/solu

OUTRES,ALL,ALL

Best regards, 

Roman

 

[htozam]

Hi Wenlong, any comment on this regards will be very highly appreciated. Thank you for your considerations. 

[Wenlong]

Hi Roman,

I would recommend you take very small sub-steps, and use only 1 or 2 iterations per sub-step, because your problem is path-dependent and bifurcation won't help you go back to the previous state. 

You may try the following commands:

NCNV, 0     ! simulation will keep going even without convergence

NEQIT, 2      ! only 2 iteration per sub-step

AUTOS, OFF   ! turn off auto time stepping

NSUB, 1000   ! Define 1000 substeps

 

Best Regards,

Wenlong