"Searching for simplicity" not so simple

I and many others agree that wind load provisions have exceeded the bounds of practicality for many of the reasons given in the excellent article by S.K. Ghosh, Ph.D., titled, "Searching for simplicity" in the December 2006 issue of Structural Engineer.

As an active member of the ASCE 7 Wind Load Task Committee in past revision cycles, I think there is a constant tension between accuracy of the provisions (and the design result) and practicality of the procedures in arriving at reasonable, "error-free" solutions by practitioners. Many ideas have been considered and none are perfect.

However, there is one level of complexity in the wind provisions that appears to have been misrepresented in the article. The 10 psf minimum load criteria was studied and debated in the last revision cycle of ASCE 7 and approved for removal by consensus vote, only to be removed from consideration by a procedural decision. It was found to be outdated and it requires that a completely different wind load analysis (with different distribution of wind pressures on the building) be considered in any design—this is not adding to simplicity. If used, it also controls the design of many building structures in the lower wind hazard regions (less than 130 miles per hour (mph) depending on exposure) and there is no real technical basis for it—this does not add to efficiency of design solutions. Instead, the 10 psf relates to simple (but conservative) design practice pre-dating modern boundary-layer wind tunnels, which have lead to improved understanding of wind loads now represented in the ASCE 7 standard. Therefore, the 10 psf minimum load requirement should be removed and not enforced or practiced. Alternatively, a similar vertical projected area pressure method could be revamped and developed as a calibrated substitute analysis procedure for the more complex pressure distributions on a building that the standard now requires in Methods 1 and 2. This would result in reasonable accuracy, as well as improved simplicity in wind design practice.

In addition, the article seems to have mischaracterized the effect of the directionality factor, K_d, on ASD vs. LRFD design methods. The inclusion of the 0.85 K_d factor was predicated on also stopping the use of the 1.33 stress increase factor commonly used with ASD wind and seismic design. This was done and the net impact is a 13-percent increase in ASD wind loads, not a decrease! The article is correct in that the rounding up of the LRFD wind load factor did result in a smaller conservative increase to LRFD factored wind loads. Thus, this change resulted in a net increase and a departure from past accepted levels of reliability for wind design for both ASD and LRFD design methods, with a greater conservative impact to ASD.
Jay Crandell, P.E.
West River, Md.



Response from S.K. Ghosh, Ph.D.:

Mr. Crandell’s positive comments are much appreciated. However, he states that I have misrepresented the 10 psf minimum load criterion. Let me try to respond to the points he makes about this minimum requirement.

He states that the minimum requirement was approved for removal by a consensus vote, only to be removed from consideration by a procedural decision. The fact is that when a proposal to remove the minimum criterion was letter-balloted within the ASCE 7 Committee, it passed the two pass/fail criteria. However, there were seven negative votes also cast. Until such negative votes are resolved, it cannot be claimed that an item has been approved by a consensus vote. The consensus process includes resolution of the negative votes.

A memo dated April 20, 2004 from the Task Committee on Wind Loads Chair Larry Griffis describes what happened next. "I must comment on ballot item 5, the 10 psf Minimum Wind Load. While this ballot item technically passed, it received some very strong negatives from some of our most eminent wind engineers, namely Mehta and Isyumov. We clearly did not get our point across to them and some others on a proper justification for the removal. I elected, as is the chair’s prerogative, to withdraw the ballot item and replace it with a clarification clause and figure in the commentary to show the proper application of the load. I was very reluctant to do this, but I feel we did not make a strong enough case and demonstrate the point that the minimum load is not needed. I simply was not able to give a credible response to the negatives when they asked for documentation justifying removal. We had none!"

It is clear that the proposal was not withdrawn for procedural reasons, as claimed, but because the committee chair concluded that technical justification for the proposal was lacking.

Kishor Mehta’s comments on the proposal were particularly instructive. He pointed out that the most important reason for having a floor of minimum horizontal pressure is life safety. According to his ballot comments, one-story building frames with large, open areas can collapse, killing and injuring many occupants. This is particularly true of manufacturing facilities, warehouses, and hangar-type buildings. The 10 psf minimum value is likely to control primarily in one-story buildings. In 90 percent of the country, design (3-second gust) wind speed is 90 mph or less. For strength design, we have a load factor of 1.6. Thus, if everything is calculated, designed, and constructed optimally, the framing system is able to resist only SQRT(1.6) x 90 = 115 mph before collapse. Thunderstorms have turbulent winds of outflow, downburst, microburst, and occasional tornado. Ground-level gust speed can exceed 90 mph in a small percentage of thunderstorms and occasionally can even exceed 115 mph. Mehta concluded, "If buildings are designed/constructed efficiently, 20 years from now we will see building frames collapsing in severe thunderstorms with virtually no warning."

Mr. Crandell states, "the 10 psf relates to simple (but conservative) design practice pre-dating modern boundary-layer wind tunnels, which have led to improved understanding of wind loads now represented in the ASCE 7 standard." Mehta has pointed out that "not a single wind tunnel facility is able to simulate thunderstorms and downburst winds."

For the above reasons, I’d like to respectfully disagree with Mr. Crandell’s recommendation that "the 10 psf minimum load requirement should be not enforced or practiced." If, after due deliberation, ASCE 7 removes the limit in the future, that will make it a different matter. Until then, if the limit is part of legal codes, it needs to be enforced and practiced.

Mr. Crandell also states that it seems I have mischaracterized the effect of the directionality factor, K_d, on ASD vs. LRFD design methods. He believes that the inclusion of the K_d-factor (=0.85) was predicated on also stopping the use of the 1.33 stress increase commonly used with ASD wind and seismic design. I was surprised at this comment because I did not believe ASCE 7 ever allowed a one-third increase in allowable stresses for seismic or wind design using ASD.

In view of Mr. Crandell’s comment, I checked Section 2.4 Combining Nominal Loads Using Allowable Stress Design of ASCE 7-95, the last edition before the introduction of the directionality factor, with Section 2.4 of ASCE 7-98, the edition in which the directionality factor was introduced. Each has a Section 2.4.3 entitled Load Reduction. The ASCE 7-95 and the ASCE 7-98 sections are essentially identical. Neither allows an increase in allowable stresses, but both allow the same decrease in the effects of variable loads when two or more variable loads act in addition to the dead loads. Thus, I am unable to verify the claim made by Mr. Crandell.

I hope the readers of Structural Engineer will find this discussion useful.