Bond_SP01_FAQ

 

Answer to Frequently Asked Questions

  

|Home|Problem statement|Model development|Simulation examples|References|Contact us|FAQ|
Table of contents:
How to model the bond-slip related member end rotation?

Top
Q:  I am modeling a bridge. Can I define BondSlip model at bottom and top of the bent columns?
A:  Rebar anchorage in bridge bent is close to column footings, so you should be able to define a zero-length section element with Bond_SP01 for both ends of your bridge columns. By the way, using the model for beam-column joints needs further study.
How to define a zero-length section element?

Top
Q:  I am modeling 3 D bridge. I defined duplicate nodes 101 and 102. I used equalDOF command to constrain 101 102 with
equalDOF 101 102 1 3 4 6. But the program gave some errors about eigen values. Have I something wrong with that?
A:  If you are not doing any coordinate transformation, direction 1 ~ X, 2 ~ Y, and 3 ~ Z, 4 ~ bending about X, 5 ~ bending about Y (torsion), and 6 ~ bending about Z. In a 3-D analysis, you should constrain translation in X(1), Z(3), and torsion (bending about Y), which is degree 5.
equalDOF 101 102 1 3 5

Q:  How can a zero-length element be assumed to have a unit length?
A:  The assumption is made by the nature of the zero-length section element, which equals the section deformation (curvature and strain) to the element deformation (rotation and elongation). This is because the zero-length section element has one integration point. Usually when we do Moment-curvature analysis, this is not a problem, However, from overall equation of motion, we get rotation instead of curvature. To interpolate the rotation to get curvature for the section, the length has to be assumed, in this case is unit length.
How to define a Bond_SP01 uniaxial material model?

Top
Q:  how can I calculate la? I used Eq.2 in the ACI paper to find la,min and I got 95 mm.
A:  you currently need six parameters (Fy, Sy, Fu, Su, b, and Rc) to define the bond_SP01 material model in the latest version of OpenSees. The rest four parameters are reserved for later upgrades of the model.

Q:  I use SI system throughout the modeling. But in your module, input system is in US system. Please suggest how I use your module?
A:  The equations for bond_Sp01 uses the normalized slips, hence the unit has no effect on the shape of the model curves.  You need six parameters (Fy, Sy, Fu, Su, b, and Rc) to define the bond_SP01 material model in the latest version of OpenSees. Among these parameters, factor b is initial harding ratio and Rc is pinching factor (both are unitless).   The ultimate slip currently is set to be dependent on the yield slip.  Hence, all you need to do is to obtain Sy the yield slip. The equation in the current ACI paper is in SI units as required by ACI, you would have to use Pa for fc’ in the equation to get the correct Sy in mm.  Some changes have been made to the equation such that the equations in Si units and English units can be converted using unit conversion. These equations will be included in the OpenSees manual and the ACI paper.

Q:  This units thing is a mess!!!
A:   use ksi for Fy and Fu in the model parameter.  Use psi in the equation for Sy.  The reason for have a different set of units in the equation for Sy is to have unit conversion correct and viable and Fy/square root of Fc’ unitless in the equation.
How to interpolate the analysis results with zero-length section elements?

Top
Q: I obtained the displacements from the top joints of the columns of the 3 bents. I got less displacements through X direction (along bent width) with Bondslip model than without Bondslip model. But, I found bigger displacements through Z direction (along bridge length) with Bondslip model. For example, my max X disp. is 2.13 inch with Bondslip model ,whereas 2.21 inch without Bondslip model. Max Z disp. is 2.43 inch with Bondslip model while it is 2.35 inch without Bondslip model.
A:   In a dynamic analysis, in which the stiffness of the bridge affects the natural frequency of the structure. this can be found through an eigen value analysis.  With a different natural frequency, the bridge might get hit by different frequency components such that the peak displacement cannot be compared directly. In this case, to compare the analysis with and without the strain penetration, the displacement of the bridge AND the force acting on he mass center of the bridge (-M*Xg(double dot) – M*X(double dot) – C*X(dot)) should be plotted.  Two such curves should be put together  to see the effect of stain penetration.
How to use Bond_SP01 model?

Top
Q: I would like to know in order to run your Bond Model in OpenSees, there is a need to do anything else in addition to save it in the OpenSees
folder. I am trying to run a case with the model and the program does not find the file.
A:  To incorporate Bond_SP01 into your own OpenSees compilation,

  • copy Bond_SP01.cpp and Bond_SP01.h into the ..\OpenSees\SRC\material\uniaxial\ folder
  • get a C++ compiler like Visual studio 2003
  • change TclModelBuilderUniaxialMaterialCommand.cpp like this or directly replace it with the one in the package (ONLY WHEN YOUR C++ complier has the same setting as mine).
  • add a line to classTags.h like I did or directly replace it with the one in the package.
  • compile the opensees for release.

These above steps do not make sense if you are not developing new models. In stead you can directly download the lateset verstion of opensees.ese from the website and do nothing.

|Home|Problem statement|Model development|Simulation examples|References|Contact us|FAQ|