Return Period on Thermal Loads

[dougantholz]

Does anyone in this forum know what the design return period is/should be for thermal induced loads, or deflections?
I have a problem that many of you may relate to. I am designing a large warehouse 300,000 sq. ft, that has an expansion joint in the middle. The research report that I am using, "Expansion Joints in Buildings - Technical Report No. 65" by the Standing Commitee on Structural Engineering (a subcomittee of the National Research Council), cites a design temperature with a 100 year return period for both cold and hot maximums.
If I am designing in BOCA '96, based on ASCE 7-95 I have to include thermal loads in the load combination D+S+T (seismic controls and deflections are small so D+S+W+T does not control). In ASCE 7-95 the design earthquake is 2/3 of 10% in 50 years, the design wind is 2% in 50 years; what thermal load should I use?
I say this because we recently had a building sustain damage due thermal expansion creating more eccentricity on a connection than it was designed for. The contractor substituted weld all around on a seat for just welds on the vertical sides. With the increased eccentricity of the thermal expansion the welds have started to un-zip. So any guidence would be greatly appreciated.
thanks

 

[ishvaaag]

The answer to your question lies in the statistical data and elaboration leading to the expansion and shortening thermal ranges mandated in your code for the application. I have given a look to a publication in Spain but also omits that. In the absence of data for any particular site, I would use the values of return for wind, since circulation of the atmosphere is in its way related to cycles of solar radiation, and these to temperatures.