Need a book? Engineering books recommendations...

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

Re: Firewall Design Philosophy - - Was "The weakest link"

[Subject Prev][Subject Next][Thread Prev][Thread Next]
Daryl (& others):

I have to say that in reality, this is NOT a new issue.  It has been
around for quite a while.  Firewalls (i.e. that must stay structurally
standing if either side collapses) have been in the BOCA code since at
least the early 90s if not late 80s.  My first major project on which I
was the "project engineer" (i.e. doing most of the design with close
monitoring by my boss) out of school in the early 90s was a good sized
medical office building.  Due to its size, construction type/fire
protection system, and fire truck access, it was determined by the
architect that we had to put in a joint in the building and at that joint
do a firewall.  The result was that I designed a double firewall system.

I have encountered other similar such situations over the years.  Usually,
such projects involve an addition to an existing building but there are
times where it is all new construction.  Had one instance where there was
again a building seperation joint, but this was largely due to the two
different funcitons of the two part of the structure...front part was
office and the back part was a large garage area for testing.

In general, it has been my experience that a double wall system is
generally the "best" and easiest solution.  Doing a cantilevered system
(either where the wall cantilevers vertically or where piers cantilever
vertically with the wall spanning horizontally between piers...assuming
masonry wall construction) usually gets somewhat impractical pretty quick
if the building is more than one taller than normal story.  And I have
never really "bought" the idea of the fusible (i.e. melting clips) link

The biggest issue I have encountered is what lateral load should the
firewall be designed to resist.  It certainly needs to be designed to
resist typical internal pressures (i.e. 5 to 10 psf typically), but some
argue that it must resist full wind pressures...after all if one side of
the structure collapses, then that firewall is now technically an external
wall.  I have been in some long winded debates on this issue.  In the end,
my experience has been that the arguement of just using an internal
pressure is what makes the most I believe that is what the
NCMA Tek Notes recommends, if I recall correctly.  But, even 5 to 10 psf
can make something like a cantilevered wall pretty impractical pretty
quick if has any real height.

So, from my experience, I have learned that in general the best solution
is to have a double wall.  Of course, it is generally NOT a perfect
solution as you now have to accommidate a thicker wall, figure out how to
handle penetrations (i.e. do you need double doors?), increased cost due
to more material labor (i.e. if masonry, you are now laying two walls...on
the plus side, it might be two walls of 6 or 8 inch block rather than one
wall of 12 inch block with a crap load of reinforcement), etc.


Adrian, MI

On Sun, 30 Apr 2006, Daryl Richardson wrote:

> MessageKevin, and fellow engineers,
>         Sorry to be coming into this thread so late but I was otherwise engaged.
>         About five years ago I was faced with the task of designing a fire wall between an existing building and a new building.  I agonized my way through many of the same thoughts as I see appearing in the 17 other posts recently submitted.  The City planning department and the Fire Marshal's office were of very little help other than to tell me that "It has to be designed by a Professional Engineer."  I did some research; and, of course, my research raised more questions than it answered.
>         In reviewing the material in Commentary M, in the Structural Commentaries for the National Building Code of Canada, and finding some of the material in that publication not making a lot of sense to me I reached the conclusion that there is a gap in the field of building technology.  Perhaps WE should do something to fill that gap (If WE don't who will?).
>         Following are some thoughts I have on the problem.
> Purpose of a Firewall
>         The general purposes of a firewall are
>   1.. It must withstand a specified fire for a specified length of time.
>   2.. It must remain standing throughout the fire and for some specified time afterwards even if the building which is on fire undergoes a complete structural failure.
>   3.. It must resist certain specified loads which may result from the fire or from other causes closely related to the fire.
> Types of Firewall
>         Commentary M of the Canadian NBC defines four types of firewall.  There may be others; but this is a good start.
>   1.. Double wall, which is, in fact, two separate walls each designed for at lease half of the combined required fire rating.
>   2.. Cantilevered firewall, which is a free-standing wall not connected to the building on either side of the wall.
>   3.. Tied firewall, where the structures on either side of the firewall are tied to each other but not to the firewall.
>   4.. Weak Link firewall, where the structures on each side of the firewall are both tied to the firewall with a "week ling" that is expected to fail in the event of a fire or structural collapse due to a fire and the failing structure is allowed to fall away without bringing down the firewall.
>         Of these, the last two seem to me as well as to others who have posted recently to be unreliable due to the fact that the "weak link" may not fail if the fire is not close enough cause failure of the link.
>         On further reflection the "size of the wall may make a difference to how it will perform.  For instance, a "long" wall may have only a "short" portion under "attack" by fire at any one time.  To extrapolate this a little further, for a "long" weak link system to fail structurally the failing structure would have to pull down not only the weak link which was supported to fail but didn't; it would also have to pull down the similar link on the other side of the wall PLUS the inherent strength of the wall itself which is connected to other parts of the wall not under duress.  This would seem to make the tied firewall and/or the weak link firewall where instantaneous, complete failure is more probable but a good detail where a progressive collapse failure is more probable.
> Situation Classification of Firewalls
>         Here I am thinking of firewalls which have other uses than just serving as a fire separation.  These include the following.
>   1.. Load bearing vs, non load bearing.
>   2.. Shear walls vs. non shear walls.
>   3.. Damaged vs. undamaged.  Here I am thinking of earthquake country.  I know that fire and earthquake are unlikely to happen at the same time; but an earthquake could cause a fire.  The performance of damaged firewalls should probably not be ignored.  Unreinforced masonry, which might be satisfactory in one location might be completely unacceptable in another.
>         There are some thoughts I have which may be useful for starters.
> Regards,
> H. Daryl Richardson
>   ----- Original Message -----
>   From: Kevin Below
>   To: seaint(--nospam--at)
>   Sent: Wednesday, April 26, 2006 1:23 PM
>   Subject: "The weakest link"
>   In wood structures separated by a masonry firewall, the attachments of the wood to the masonry wall should be designed so as not to cause collapse of the wall if the wood structure on one side burns and collapses.
>   In the Canadian Commentary M to the NBC 1995, they show a system using a "liaison faible" which I guess is translated from "weak link".  The idea is to provide lateral support to the masonry wall at each floor level, and to resist certain lateral loads such as seismic and wind, as well as loads which may occur during the fire, such as the force of the water jet.  But in the case of fire, to be weak enough to break under the force of the falling structure.
>   There is not much detail as to how to implement this.
>   Does anyone have any examples ?
>   Kevin

******* ****** ******* ******** ******* ******* ******* ***
*   Read list FAQ at:
*   This email was sent to you via Structural Engineers 
*   Association of Southern California (SEAOSC) server. To 
*   subscribe (no fee) or UnSubscribe, please go to:
*   Questions to seaint-ad(--nospam--at) 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: 
******* ****** ****** ****** ******* ****** ****** ********