1 Topic by engpawlo (edited by engpawlo 2013-02-18 07:32:04)

  • Registered: 2013-02-03
  • Posts: 4

Topic: End anchorage of headed bars, Non-linear springs

Hi,
I'm using the latest version of ATENA 3D-Eng to study the behaviour of GFRP headed bars in beam-column joints under seismic loads where beam longitudinal reinforcement is anchored in the joint using bearing heads (conical shape heads).

From the experimental testing, I have the curve that describes the relationship between the strain in the bar (just before the bearing head) and the end slippage of the head. This relationship is non-linear.

Also, it worth to mention that the head has a bearing capacity limit (less than the bar tensile strength) at which it breaks. Also it is needed to consider that the bearing head is active only when the reinforcement bar is subjected to tensile force since GFRP reinforcement is not active in compression.

Some Ideas and brainstorming that I need you to discuss to reach the best model:
1- Fixing the bar end is not a good idea, since the head has a bearing capacity less than the tensile strength of the bar.

2- Use a non-linear Joint spring, axial stiffness (EA/L) ?? ATENA needs material, area, length. I added spring material that represents the displacement-stress relationship (at bar section just before the bearing head). however, when I add joint spring, it does not accept the area of the bar (2.0E-4 m2 equal to area of 16mm diameter bar) and consider it = 0 and I get a warning message that the area should be in the range of 0.000 and 1E99. I multiplied the area by a factor of 10 then I devided the stress in the spring material by the same factor(10). is this OK?

3- if the end of the bar and the joint spring are connected to the same geometric Joint, Will this simulate the effect of the head to resist the slippage of the bar end?

4- Can I divide the reinforcement bar into 2 parts connected in one geometrical joint?. The bearing head can be presented by a bar 50mm-long (very small length) with a user-defined bond material that is different than the bond material of the main bar.

Could you please help and guide me through these ideas or you are welcome to give other suggestions and recommendations on how to model this end anchorage of the bars?

Quote Post 1

  • dpryl
  • CC software developer
  • Offline
  • Registered: 2007-07-13
  • Posts: 1,107

Re: End anchorage of headed bars, Non-linear springs

Ad 1: Actually, even this simple model can make sense. However, you need to check the head limit stress in postprocessing and ignore all the results from all steps following the one when first time exceeded.

Ad 2: Springs in ATENA are NOT for connecting elements together, but for modelling the response of external members, like the soild under a fundament or similar (spring supports).

Ad 3: No - see 2 above.

Ad 4: You can, but the problem is the bond slip is independent for the 2 bars. In other words, a discontinuity, "jump" of the bond slip can develop at the point - which is not what you need.

An option might be to define a function scaling the bond properties depending on the distance from the bar beginning. However, the ATENA .inp file has to be modified manually to use this option.

Another possible way of modelling would be an elastic volume representing the head, with interfaces to the surrounding concrete (and the bar ending inside the volume, with slip blocked at the end).

Yet another, more simplified, way would be to make the head volume a simple cube (bilinear steel or concrete material model, Poisson's ratio 0), meshed with 1 or 8 finite elements, attached to the surrounding concrete using perfect connections (or maybe 1 surface perfect conn, other 5 no conncetion or similar), defining the material properties such that the response of the cube corresponds to the head response.

dpryl's Website

Quote Post 2

  • Registered: 2013-02-03
  • Posts: 4

Re: End anchorage of headed bars, Non-linear springs

Thanks for the recommendations,
could you please explain the last 2 options in more details to get a clearer vision especially for the last option. I am not sure how to define the material either by using concrete or bilinear steel to reflect the behaviour of the bearing head.
Thanks.

Quote Post 3

  • dpryl
  • CC software developer
  • Offline
  • Registered: 2007-07-13
  • Posts: 1,107

Re: End anchorage of headed bars, Non-linear springs

As you say you have the working diagram of your anchor head, the last option is the first one to try. If the anchor response is similar to bilinear steel-like curve, define a material based on von Mises Steel and assign it to the volume such that it behaves similar to the measurements you have. In other words, calculate the material properties from the measured data and the dimensions of the volume representing the head. Only make the volume a single element to preven unrealisti local platifications around the bar endpoint.

If the response is more similar to the tensile or compressive response or concrete, make the single-element volume from a material based on NLCem2.

dpryl's Website

Quote Post 4

  • Registered: 2013-02-03
  • Posts: 4

Re: End anchorage of headed bars, Non-linear springs

Thanks for the prompt response and the clarification you provided. I'll try it and let you know.

Quote Post 5