Return to index: [Subject] [Thread] [Date] [Author]

Re: Software Developers participation in list discussions

[Subject Prev][Subject Next][Thread Prev][Thread Next]
Thanks to Dennis Wish for his invitation to software vendors to participate
more fully in this list server.  There are several of us here at Keymark
who lurk, and the lurking has been invaluable.  It is not our intention to
take undue advantage of this forum, but there are several questions we
would like to pose to this audience.  In the process we hope to clarify the
abilities and limitations of our KeyLat software.

It is possible within the KeyLat system to model many different plate
heights within a single level.  The software recognizes different wall
heights for each shear wall defined within that level, and will adjust the
stiffness of the wall accordingly (for rigid diaphragm theory).  The
current version of KeyLat takes the varying plate height information found
in the 3D model of the structure and assigns a single elevation for the
diaphragm for each level.  The user has control of the assigned elevation.

The difficulty is related to the definition of the diaphragm.  For example,
a stepped diaphragm is modeled as a continuous diaphragm.  The "step" is
ignored and forces are transmitted to the shear walls as though the
diaphragm had no discontinuity.  Recognizing this limitation, KeyLat
calculates and reports all forces in the diaphragm so that the design
professional has enough information to design an appropriate load path to
transmit these forces from one step to the other.

Question to the audience:
If we were to perform an analysis of lateral forces (using either rigid or
flexible diaphragm theory), what would we do differently if we were to
consider stepped diaphragms?  Perhaps it does not change the distribution
of forces at all -- if we stay within the strict confines of the theories
involved.  In reality, the distribution of forces for this case is complex
in the extreme, perhaps involving warping of the diaphragm and bending of
the members supporting the plywood panel.  Two important issues that result
from the use of stepped diaphragms are as follows:

1) How to transmit chord and drag forces past the discontinuity.
2) How to get a stepped diaphragm to transmit shear past the discontinuity.

We are not sure how to programmatically assist the design professional with
these design issues.  Your suggestions are welcome!

Within the current version of KeyLat, the user can create a model of a
building that has skewed (i.e. non-orthogonal) wings.  With some
straightforward modeling techniques it is possible to analyze this type of
structure for flexible or rigid analysis.  

What KeyLat does NOT currently handle is one or more non-orthogonal walls
in an otherwise orthogonal building.  Imagine a rectangular structure with
several skewed interior walls intended to carry shear.  Or perhaps a
would-be 90-degree re-entrant corner that has been replaced with two
135-degree corners.  These types of walls are not explicitly supported
within the current version of KeyLat.

Within the KeyLat software, walls that are not orthogonal can be defined as
a shear wall.  We do this by allowing the user to set an angular tolerance.
 Any wall whose orientation defines an angle with one of the principle axes
that is less than this angular tolerance can be selected as a shear wall.
However, it will be treated in the analysis (flexible or rigid) as an
orthogonal wall.

Question to the audience:
Exactly how does one implement non-orthogonal walls into the flexible
diaphragm theory?  We at Keymark are not convinced that such walls have
meaning within a strict interpretation of flexible theory.  It is much more
straightforward to utilize non-orthogonal wall in the rigid theory.  We
intend to support the use of non-orthogonal walls in the rigid diaphragm
theory as soon as possible.  

I would welcome other questions about the KeyLat software, although this
list is probably best-served by specific technical discussions about the
program.  More general questions can be directed to our KeyLat email
account: "keylat(--nospam--at)" or to our new website discussion forum at
"".  Please bear in mind our discussion forum is less than a
week old and we are still working out the bugs.


John M. Drozdek, E.I.T.
Project Engineer for KeyLat
Keymark Enterprises, Inc.

At 01:15 PM 2/29/2000 -0800, you wrote:
>   There are a large number of  software developers who lurk among our
>discussions and who refrain from  participating in the discussion out of
>respect to the goals of this list.     I want to invite all  software
>developers to participate in the discussions in an advisory position.  Our
>intent is not to sell or promote software (although users often make 
>recommendations). However, it is best to receive answers to specific
>questions  by the developer rather than through speculation.    In most
>cases, questions  are best asked of the developers through their own tech
>support sites.  Realistically, this does not happen as software is a common
>issue which is  discussed on this list. Therefore, I think it certainly
>appropriate for a  developer to jump in and answer questions as well as
>discuss special features of  their software which can both help existing
>users gain the most of their  products AND offer information to those
>evaluating software to help them draw  their own conclusions. All other
>types of promotions are best left to the  software developers website or
>advertisements.   Regards, Dennis S. Wish, PE Structural Engineering
>Consultant (208) 361-5447 E-Fax  Attachment Converted: