Need a book? Engineering books recommendations...

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

RE: Calculations / BIM

[Subject Prev][Subject Next][Thread Prev][Thread Next]
 

Current Project as an example. Seemingly simple carport (not quite is 3D).
Owners original drawings showed rafters from 150x50x5 RHS, with a 7m radius
of curvature. In the mean time he has found a bargain and got some 100x50x4
RHS with a 5m radius. So had to build a new multiframe model 3D.

If the entire thing could be done in Excel then simply change the radius and
all calculations updated: that's all I have to do to get the new wind loads.
But I have to manually transfer into multiframe, before which get the new
geometry from ACAD.

If I had a BIM system then I would expect that all I have to change is the
radius, and the rest of the dimension and geometry would automatically
update inaccordance with defined relationships, and that the loads on the
structure can also be automatically adjusted. Then I can size the members
and connection components and I don't have to do anything outside the BIM
application.

As it is I have to get the results from multiframe, and then conduct the
member checks and connection design in Excel or with pencil and paper.
Member checks outside multiframe because it cannot form a design member for
the curved beam, to use the AS4100 steel designer module: not that stops it
from checking it indicating its perfectly fine.

For simple stuff can build an entire parametric model in Excel (A simplified
BIM, which can automatically generate drawings in ACAD if desired). But when
comes to 3D structures, or complex 2D frameworks, then that linear
calculation presentation starts to become cumbersome. Just want to get the
appropriate results of the frame analysis into Excel, so that continue with
the rest of the assessment. For current situation I haven't got the required
automation in place.

What want to be able to do, is define a model by a few basic parameters, and
then experiment with changes in those parameters, and assess the sensitivity
of the results to those changes in parameters. At the moment Excel and
similar calculation workbooks appear to be the best option, though without
any parametric graphics: though getting better at developing XY charts for
parametric sketches.

The real requirement of BIM is a universal data model, which any software
can extract appropriate information from. Which is perceptibly different
than exporting from and importing to. Standards like STEP and cimSteel have
not yet been fully adopted: and typically provided as export/import
facilities rather than the normal way of doing things.

Consequently may aswell use Excel or MS Access as a central data repository,
defining own data model, and use these tools to automate the tasks that
normally carry out: which may be part or the whole of a building project.

If primary contribution is calculations and other businesses provide
drawings of proposals to be assessed, then Excel is all that really need to
create the parametric whole structure model. With occassional need for
software like multiframe, CFS etc...

There is no real benefit in analysis of a 3D model, if series of 2D frames,
or it is simple beam construction. BIM seems to force unnecessary repetition
of calculations by pushing the 3D analysis procedure.

My impression is also BIM doesn't support assemblies and subassembly
definitions. So it determines the number of columns, by counting individual
columns, rather than counting common plane frames and multiplying out by
columns in plane frame.

If it supports assemblies then should be able to export the plane frame,
analyse, modify member size, import back into BIM main model, then have all
like frames automatically update to the new member size. 

Different people however have different concepts of the abstract assemblies
making up a project. Construction may have the view that the building is a
collection of columns to which rafters are added, so they want all the
columns first. Another may view as series of plane frames, each of which is
assembled in place first. The manufacturing/construction process thus has an
influence on the product-structure-tree. Computer science refers to data as
raw material, and information as the data presented in such a manner as to
aid decision making. The data in the BIM, needs to be sorted and structured
differently in order to generate the information required by the different
people involved in getting from concept, through design, procurement,
fabrication, and constuction, aswell as final inservice operation and
maintenance.

Even with something as simple as an indented bill of materials (BOM), I can
structure the realtionship between single level BOM's differently than
someone else: thus defining a different product-structure-tree, and
consequently different relationships between the component parts. This in
turn changes the sequencing of the production and procurement process. So
that having a centralised data model itself is not a benefit if that data is
not correctly structured to meet the needs of all persons in a project.
Manufacturing tends to benefit better because all tasks are generally
completed by the one organisation, with only component parts bought in: that
is there is less separation of the design and production processes and so
potential for better integration than in building and construction industry.

Product defintion may have to be modified by production to assist with
process design. So that the BIM created by the architect may not be of any
use to anybody else: other than as a starting point for restructuring the
data and/or over laying additional data. Whilst the BIM may enable precision
calculation of linear metres of steel required, the steel has to be
purchased in economic order quantities (EOQ). If the offcuts from a stock
length aren't any use, then why cut them off, why pay for scrap?
Manufacturers can make these decisions and adjust the design to suit, the
tendering contract process however, tends to result in paying for scrap:
since contracted to supply what is on the design drawings. Such things can
be played around with if have a complete data model, but whether pictorial
graphics are required is another matter.

Can equally well start with a blank piece of paper and start writing the
BOM, or start drawing. The only way anything gets on the blank paper is from
the visualisation in our heads, and that 3D graphical model can be generated
way faster than the one on the computer. For simple structures can go
straight from imagination to finished structure without any drawings. For
more complex systems, need to coordinate with others, and therefore need
appropraite media for communications. And I think Alex's article also
implied that BIM has not yet achieved appropriate means of communication. At
present BIM basically provides the dimension and geometry, but attaching
additional information to that data is wasteful if it cannot be attached to
classes of object and has to be attached to individual objects. Don't just
want to know the building has rafters, want to classify them into groups,
apply group technology concepts, and identify common parent forms from which
all can be derived. Sure can just churn out CNC code for machine tools, but
still need to properly sequence the production. Making one unique rafter at
a time may cause unnecessary delays in construction. Knowing there is a
common parent and fabricating that parent first, and then reprocessing to
form the unique rafters may be more efficient, and also achieve the required
consistency (lower manufacturing variation).

If BIM is to achieve its full potential then a lot of people need to get
involved building the BIM for a project early on in the process, and meshing
their dependent calculations together. The results of one persons
calculations are the inputs to another persons. If can do this for the use
of Revit or similar, then could equaly well mesh each contributors Excel
calculations together. So that at a meeting, can vary the parameters as each
person argues their case, and immediatley review the calculated effects on
others contributions: no drawings required.

But who can get that level of coordination and sharing? For it places
peoples jobs at risk, for the computer is now doing in minutes what one or
more people may have been taking hours to do. With a fully integrated model,
may collapse back to the need for just one designer: possibly in the form of
the builder or even the owner.

So have two situations:

1) Repetitively manufactured product, where objective is to eliminate the
delays caused by design. For such purpose have custom design tools, with
high level of automation.

2) Custom product, where have to accept delays for design, but still want it
conducted as fast as possible with supply of finished object as soon as
possible. Here tend to use more off-the-shelf design tools, which are
flexible and can accommodate a variety of project types.

Though in both cases Excel may be a suitable starting point. Carport
manufacturers tend to provide sales agents with Excel workbooks, to size up
and cost a customers requirements. Though I believe there is a huge
difference between what the customers want, and what the workbooks are
capable of handling: which results in delays in approval that the customer
wasn't expecting. But providing every sales agent with Revit is not
realistic. Rather the Excel workbooks need more constraints, and the sales
agents need proper training. A small scale customised BIM is what they need:
and such cannot be economically built on the likes of ACAD, Revit with
multiframe or similar as the analysis engine.

Such things lead people towards building custom solutions, because the
monster packages which are available are just not suited to the end-users.




Regards
Conrad Harrison
B.Tech (mfg & mech), MIIE, gradTIEAust
mailto:sch.tectonic(--nospam--at)bigpond.com
Adelaide
South Australia





******* ****** ******* ******** ******* ******* ******* ***
*   Read list FAQ at: http://www.seaint.org/list_FAQ.asp
* 
*   This email was sent to you via Structural Engineers 
*   Association of Southern California (SEAOSC) server. To 
*   subscribe (no fee) or UnSubscribe, please go to:
*
*   http://www.seaint.org/sealist1.asp
*
*   Questions to seaint-ad(--nospam--at)seaint.org. Remember, any email you 
*   send to the list is public domain and may be re-posted 
*   without your permission. Make sure you visit our web 
*   site at: http://www.seaint.org 
******* ****** ****** ****** ******* ****** ****** ********