Displacement-based pushover analysis of a masonry arch bridge

  • Last Post 19 September 2019
Semih posted this 18 January 2019

Dear Community,

I am studying the nonlinear seismic behavior of a masonry arch bridge by means of a static pushover analysis. 

I used macromodelling approach in which the masonry is modeled as a homogenous continuum material. Different structural parts of the bridge (spandrels, backings, arches, piers, fill) are modeled seperately and connected together with bonded contacts. Frictional contacts are assigned between the fill and the spandrel walls. I used SOLID65 element to model the masonry and bilinear isotropic model for the fill. As for the BC's, I have compression only supports with very high contact stiffness and zero displacements in horizontal directions so as to prevent early cracking in the piers.


Currently, I apply the force in the transverse direction of the bridge using inertial acceleration increased gradually (so, it's a mass proportional pushover analysis) and obtain the base reaction in the transverse direction & top displacement values until the non-convergence occur (see an example below). I am fine with the crack pattern etc. However, I can't obtain the degradation in the pushover curve with this method. Instead of increasing the force (inertial acceleration in this case) I need to apply displacement in the transverse direction of the bridge (although I have my doubts whether I can get the degradation with solid65 elements). 

Pushover Curve

Here comes my problem:

The behavior of the bridge is not straightforward to define i.e., mid part deforms more compared to the sides and the top part of the bridge somewhat rotates slightly as you can see in the figure below. 

total deformation at the last converged step
I am using ANSYS 19.2. 

Any help on how to apply such displacement or improve the model is appreciated. 

Kind regards.

Order By: Standard | Newest | Votes
peteroznewman posted this 18 January 2019

When you say "the degradation in the pushover curve" do you mean you want to see the negative slope on the curve you plotted, past the convergence limit of the current acceleration-based input?


  • Liked by
  • jj77
Semih posted this 18 January 2019

Yes, exactly.

akhemka posted this 21 January 2019

On rotational behavior - it looks compression only support will allow for rotation - the bottom of three pillars can have zero displacement to avoid this rotation.



Ashish Khemka

Semih posted this 21 January 2019

Hi Peter, 

Have you got any thoughts on how to obtain the part of the curve with negative slope?



peteroznewman posted this 21 January 2019

Hi Semih,

Under Analysis Settings, you can try to use Stabilization. That adds a fictitious force to keep the structure in equilibrium while continuing past the peak of a force-displacement curve.

Regards, Peter

  • Liked by
  • jj77
Semih posted this 23 January 2019

Hi again Peter,


Stabilization helps the analysis converge a bit further, thanks! 

Do you know how I can turn on the arc-length control? Do I need a command snippet for that?

Also, I just can't picture in my mind how to past the peak of the force-displ curve and obtain the part with negative slope using this continuously increasing acceleration-based input. I would really appreciate if you have any explanation for that. 

Kind regards.



sujitbhandari posted this 02 February 2019

Hi Semih,

I was also working on pushover of a masonry wall. I was curious about what degradation/ plasticity model you used for the modeling of material nonlinearity. I couldn't get the solution to converge for my model using the William-Warnke model defined in concrete plasticity and bilinear isotropic hardening.

I was wondering if you could help me on this regards.

Thanks in advance,



Semih posted this 03 February 2019

Hi Sujit,

Are you modeling a single wall panel? If yes, then is it in-plane or out-of-plane pushover analysis?



sujitbhandari posted this 17 February 2019

Hi Semih,

Sorry for replying so late. I didn't quite saw your reply.

I am modeling in-plane pushover of a multi-leaf (with three different layers with frictional contact in-between). I have used different models since then, including Druckar-Prager Concrete, Microplane model. However, still I am not satisfied with what I have got! The parameters for microplane model are very hard to get for masonry, and it doesn't behave as expected. Also, DP concrete model shows very early degredation in pushover curve in my model. I am quite stuck, and going through all of literature and stuffs I could find.

I don't get how did you make the solution to converge in SOLID65 in your case. I can't just go beyond the elastic region in my case.



Semih posted this 19 February 2019

Hi Sujit,

No worries. 

I had started modeling a wall too, but didn't continue since I don't have the experimental data yet. It is a bit tricky with Solid65 elements and I am not sure if it works well with "small" structures like masonry wall panels, because cracks propagate fast and cause premature failure. In terms of the bridge model, the bridge's thickness is 4.3m. 

I recommend you to turn off crushing and use a shear transfer coefficient greater than 0.2 in Solid65, do mesh refinement where necessary (as Solid65 is a smeared crack model and is mesh sensitive) and use displacement-based loading with very small increments. Turn Modified NR and stabilization ON, and check strain energy vs. stabilization energy to be sure that stabilization is not spoiling your results.

Are you comparing your results with experimental data? DP model with linear softening is similar to Solid65, and I think you can modify the DP parameters to match your results with the experimental data. 

I hope this helps. 

Please let me know if you figure out what to use for your wall model, I must model my wall panel experiment at some point too. 




seker posted this 27 February 2019

Mrb Semih Bey;

Bende tarihi bir köprü üzerinde aynı analizleri yapmak istiyorum. Fakat bir türlü başarılı olamadım. Pushover analizi nasıl yapılır, crack nasıl görüntülenir, solid 65 eleman nasıl tanımlanır gibi sorularıma yardımcı olabilirmisiniz? elinizde basit bir ansys 16 workbench modeli varsa ve bana gönderebilirseniz, görsel olarak sorunu daha kolay halledebilirim. yardımcı olabilirseniz çok sevinirim. iyi çalışmalar. kolay gelsin dileklerimle.

email: senol.seker@amasya.edu.tr

d.g. posted this 16 June 2019

Hi Semih,

I hope you have resolved your doubts. I. too, am trying to study how to do a pushover analysis of masonry, but I do not understand how to set parameters, expecially for the definition of the material. Please, can you sent to me an example file? I work with workbench 18.0, so the project must be saved in this version. Thank you. 

abdalla posted this 19 September 2019

Dear my colleagues

Please, I need to know what is the best element which can accurately model the masonry wall under in-plane & out-plane loads in Workbench? Because the behaviour of an in-plane masonry wall completely differs from its behaviour out-of-plane.