Response-2000 Bugs

Response-2000 has been in development for over 4 years. It consists of over 56,500 lines of C++ code so far. Unfortunately, the complexity does lead to a few bugs here and there that have yet to be found. This page considers itself primarily with things that Response-2000 doesn't do yet but will consider soon. It also includes things that Response-2000 doesn't do as well as it will soon.

Inactive Menus

On some older systems, mainly Windows NT with Service Pack 3 or earlier and Windows 95, some menus in the program may not run properly. This is a result of Response-2000 relying on a part of the operating system that has been changed by Microsoft frequently over the years. If this error occurs, download the program "50COMUPD.EXE" from the microsoft support site at the following address. This will update the file comctl32.dll to the lastest version and may improve the operation of other programs besides Response-2000. The download is free and small available at:

http://support.microsoft.com/support/kb/articles/q186/1/76.aspb

Things that Response-2000 is poor at now:

• Analyses at high shear to moment ratios. (results noisy and slow). This is generally only a problem when the M/V ratio is less that about the beam depth.
• Unreinforced concrete. Poor answers are obtained for sections without any steel at all.
• Behaviour with shear and high moment not optimal. When running an analysis that is governed by the longitudinal crack check (red line on the principal tension plot cuts in diagonally), the analysis generally stops prematurely. The load at failure is well modelled, but the ductility at failure is underestimated.
• Axial load application far from gross centroid. Response-2000 allows axial load to be applied at any depth in the cross section. Analyses have been found to be unstable if it is applied too far away, however. If default location of axial load application is used, results are generally very stable.
• Noisy M-V, N-V interaction diagrams. In some situations, an underestimation of the shear strength happens and shows up as "spikes" in the interaction diagrams. Generally does not happen for "Full Sectional Response" analyses.
• Problem with elastic-to-rupture materials. For materials that display linear elastic behaviour to the point of rupture, Response-2000 will produce very conservative results. The problem is that the program assumes that the strain at a crack must be able to be twice the average strain for the crack check. To account for this, increase the strain at rupture for the material to twice the measured value and at the same ultimate stress. That is, give the material a "yield plateau" that reaches to twice the yield strain.

Things that Response-2000 doesn't do at all yet:

• Confinement. (though this can be simulated with the segmental concrete model)
• Rotate cross section for biaxial moment case
• Hollow cross sections. Response-2000 can implicitly consider the behaviour of hollow sections simulated by moving all wall material to centre. For example a hollow box becomes an I beam. A double Tee beam becomes a single T etc. Response-2000 will eventually be able to print out the cross section with the hole shown however.
• Footing style UDL not yet implemented
• Cyclic behaviour of materials