It is true that "connections are much more likely to give trouble than
the members themselves." Once connection strength has been changed, all
structural properties, i.e., stiffness, mass, damping, may change.
Undoubtedly, structure must have a different response. The problem is
that the traditional methods for structural analysis and design cannot
analyze the effects of variable-strength connections, but simply assume
perfect-rigid connections. This not only creates a hidden cause to
structural failure, but also leads to an abnormal design procedure. Many
engineering analysis and design makes assumption at the beginning, and
checks if the assumption is OK at the final design. The traditional
methods for structural analysis and design simply makes an assumption of
perfect-rigid joints, but never check if the perfect-rigidity assumption
is OK at the final design. It is a strange and incomplete procedure.
A new trend is partially rigid analysis, which can answer the questions
the traditional methods fail to solve. Partially rigid analysis is a new
technique. There is a trial edition at
(http://www.Equation.com/IFAS/download.htm). Download at your
convenience (9.7-Mega bytes), if interesting. IFAS is the only
structural package that provides partially rigid static analysis,
partially rigid dynamic analysis, and partially rigid effective-length
analysis (theoretical FEM result).
> From: Christopher Wright[SMTP:chrisw(--nospam--at)skypoint.com]
> Reply To: seaint(--nospam--at)seaint.org
> Sent: Thursday, June 24, 1999 9:42 AM
> To: SEAOC Newsletter
> Subject: Re: 125' long steel truss, connection question
> >If you design according to the above, you can model the joints with
> >degree of fixity (not perfectly fixed though). It is easier to
> >as pinned and have some reserve strength, however.
> Undoubtedly anathema to building designers, but I make it a practice
> use member loads corresponding to complete fixity results for
> connections. My own experience (borne out by the Northridge business)
> that connections are much more likely to give trouble than the members
> themselves. This is where the reserve strength is needed. I think this
> especially true of welds because they don't handle relative
> between connected members very well--the strains are just too large.