Old Wood Truss Failure 4

[JAE]

I've got a task to study some old (1940's) wood trusses which have been undergoing various levels of distress for the last 20 years or so.

The trusses are bowstring (bottom chord flat/level) spanning 160' with a depth of about 20' at midspan. They are made up of 2-6x12 bottom chords and two similar sized glu-lams at the top chord. The webs are 6" nominal that extend between the chords.

What is happening is the 6 splices that occur across the bottom chord have been failing. These are made up of about 6 feet length of lapped members which are through-bolted with 4" diameter shear plates at each bolt/interface.

It appears that the wood isn't checked or split generally (except for the splits where it failed) but it looked quite dry. No sign of rot. These trusses are about 28 feet above a concrete floor within a building that is very long.

Does wood tend to loose strength over time? This is in a non-air conditioned space so it probably gets a little hot in the summer (midwest US). I know all about duration of load factors and moisture strength loss factors. What I was wondering was, does wood simply get dried out and "brittle" or something after 60 years?

 

[DaveVikingPE]

There should be no problem with "old" wood. The USDA Wood Handbook relates brittleness in wood to exposure to acids. What's in the building might be a good question is there's a possiblity that it held chemicals or one sort or another.

Likewise, the same book states, "In relatively dry and moderate temperature conditions where wood is protected from deteriorating influences such as decay, the mechanical properties of wood show very little change with time."

Was there anything unusually heavy hanging or hung from the bottom cords at the splices? What's the condition of the shear plates and bolts?

 

[JAE]

The bolts are fine, the shear plates are OK. It looks like the wood was split at the splices due to bolt bearing or slippage of the shear plates.

Visually, the wood just looked very dry and dusty and appeared (visually anyway) to be less dense than I would have expected.

No acid that I'm aware of as the trusses are 28' in the air and the entire facility was used in the 40's to build aircraft. Currently a fitness center.

No unusual loading either....there are at least 50 other trusses that are OK.

 

[Aton]

As I understand the situation, only some of the tension chord splices are failing, while others hold well? That could make a certain amount of sense, given the time in which they were engineered. Every piece of wood is different from the next (of course...). Although good quality wood cut from larger-diameter trees may have been more plentiful 60 years ago, the funds and ability to test wood was not as advanced as it is now. Nor had limit states design factors been established (along with the 5th percentile failure probability). In fact, testing wood in tension has always been problematic, for stress concentrations would inevitably develop due to the clenching of the jaws of the testing apparatus - making any tensile data virtually useless. In addition, Young's Modulus was based on small, clear samples, and not on the variety of grades that are tested today. 60 years ago the great possible differences between one piece of wood and another might negate some of the less conservative factors of safety used then. I would guess that - barring the obvious possible causes that you have already eliminated - those failures may simply be due to the cyclical loading of those members that were not quite as able to resist stress as were those that did not fail. It doesn't take much to make a member prone to splitting: an internal pitch pocket, a small spike knot, gnarly reaction grain (from growing at the edge of a forest and being wind-buffeted), even a small gouge during construction - all can contribute to eventual failure. Sustainable, Solar, Environmental, and Structural Engineering - Solutions world-wide.

 

[JAE]

Thanks, Aton and DaveViking

 

[JWJ]

Take a look at the article, "Investigating and Repairing Wood Bowstring Trusses" in the Practice Periodical on Structural Design and Construction, Vol. 1, No. 1 (available through ASCE). I have a book (Structural Renovation of Buildings) which refers to this article. It claims that a study in the Chicago area found that bowstring trusses constructed in the WWII era were underdesigned by today's standards.

I am working on an investigative report for a bowstring truss with a bottom chord failure. The failure initiated at a knot. The building is in Chicago. The truss actually failed during the summer when no snow loads were present. The only documented loads were moderate winds during the week that the truss failed.

If you find any good information on deterioration of protected wood over time, please let me know.

Thanks!

 

[JAE]

JWJ - I pulled out my copy of Vol. 1, No. 1 and there it was! Thanks for the reference. In fact, much of what is said in there we ended up following (great minds think alike??? )

We found that a number of the bottom chord splices were split in areas originating at the bolts (split ring connectors). Our calculations showed that the ring connector capacities were less than the chord forces that the original engineer showed on the plans.

The most unique issue, though, was that there were only two areas (of about 6 trusses each) that had significant problems. These were discrete areas along a very long building. We kept asking "why just in these areas - the conditions are the same throughout".....then we noted that these two banks of trusses were actually supported at one end by a steel support truss spanning over two side doors (each 100 feet wide). Apparently, the side (lintel) truss was more flexible than the other supports (concrete columns at each truss) and this may have been the straw that broke the camels back.

Our repair: we didn't use the post-tensioning rods shown in the article, but did do, basically, the same thing as we added steel brackets to the bottom chord along the length and attached steel tube sections, one on each side of the bottom chord, to ensure adequate strength should a splice fracture. We could live with a little extra deflection in this roof, but had to have a safe roof. Post-tensioning was determined to be too difficult as access was very limited at each end of the trusses.

thanks again for the reference.

 

[Aton]

Great to hear the follow-up story, JAE. This is how we all can learn.

Sustainable, Solar, Environmental, and Structural Engineering: Appropriate technologies for a planet in stress.

 

[oldrunner]

Within the last day, there was a bowstring truss failure in Sacramento, California. For more information, check the Sacramento Bee:

http://www.sacbee.com/

These kind of failures tend to have a ripple effect with one truss failing and then pulling down the adjacent trusses and so on.
About 10 years ago I investigated a similar truss failure in Richmond, California, where all of the trusses failed in a ripple effect. The trusses were approximately 50 years old and had been badly abused - ie: hanging stuff off of the bottom chords, non-engineered repairs to various members etc. I believe the final fix was steel tension rods. This particular design had been done by an architect and had lots of eccentric joints. He was still alive and verbally defending his design.

 

[scherry]

I repaired a lot of bowstring trusses in southern Californmia after earthquakes. We repaired everything from top chords to bottom chords to webs depending on the stresses and how the trusses broke. We also repaired large bowstring trusses after the third or fourth reroof in older buildings. in one building we had a part of the roofing removed and weighed. It wieghed over double what the original design allowed for. We used tension rods or steel cables for reinforcement of the lower chords, with specifically designed steel shoes that fit over the external part of the heel connection. then afterwards when everything was where we wanted it to be, we used clamps and epoxy on any splits. we replaced members that were broken or split through the bolt lines. We had one building where I did a third of the trusses after the whittier narrows, 1/3 after the northridge quake..and then I left CA. But we never had a repeat failure on any of the ones we repaired.