Topic: Modeling pull-out test

Dear Sir,

Now I am trying to modeling a pull-out test, in which the steel bar is prestressed by pretension. From your earlier answers, I find you suggested that using cooling to simulate the prestress due to pretension because initial strain (which is the accurate way) doesn't work for reinforcement. So, my questions are:

1) Is this cooling really equivalent to initial strain? Because for pretension, we know how much is the initial strain e0 after we stretch the steel bar. However, at this time concrete has not been cast. After concrete is cast and its strength is fully developed, we release steel bar. The steel bar will contract and the concrete bond will prevent it, thus prestress is introduced to concrete (now, the strain in steel bar is less than e0 and it is unknown and not uniform along its length). However, when I cool the steel bar, the concrete bond has already existed and it will prevent the steel bar from contracting. If I still force steel bar to cool to e0, the strain in steel bar is uniform along its length and the real prestress is higher than pretension.

2) In Atena 3D, could I use 3d model to model the steel bar, i.e., model bar as a cylinder? If I could, what interface material do you think it is suitable to simulate the concrete bond?

Thanks a lot

Re: Modeling pull-out test

Dear Mr. Qiang,
Ad 1. initial strain and thermal-induced strain is exactly the same then applied to an embedded bar in ATENA. The only difference is how you define it (eps or delta T times epst). In both cases you describe, the reinf. strain acts against mature concrete.

Ad 2. you can, but in only makes sense in very special cases (e.g., where the shear resistance of the bar is the main factor). Defining the interface material parameters is one of the issues which cost significant amount of time+work if you have use that modelling.

Re: Modeling pull-out test

Dear dpryl,

Thank you very much for your answers. However, if you could help me clarify the following questions, I will deeply appreciate it.

1) Could you explain to me the time sequence of initial strain and thermal-induced strain? As I understand, when you apply initial strain eps0, it means at t=0, the steel bar has a strain = eps0 and the surrounding concrete has a zero strain. Then ATENA will calculate the reaction at step 1 (t = t1). But for thermal-induced strain, does it follow the same time sequence?

2) If the initial strain is very big, usually 80% of its yield strength, it will usually cause the bound slip which I am focusing on. Thus it means small steps may be needed to simulate the process of prestress release. Do you know how I am able to use multiple steps to do this?

Thank you very much.

Best,

Qiang

Re: Modeling pull-out test

Dear Qiang,

Ad 1: in both cases, you apply "strain which the bar would have without any restrictions". Both of them are normal loads in ATENA (like prescribed forces or displacements), and are applied incrementally. As long as we talk about static analysis, time is not of much importance. Anyway, if you apply one or the other in load steps 1-10, the effect is the same.

Ad 2: define multiple steps with step multiplier 1/No_of_Steps (instead of a single load step with multiplier 1).

Re: Modeling pull-out test

Dear dpryl,

Thank you very much