Jump to content

Talk:Limit state design

Page contents not supported in other languages.
From Wikipedia, the free encyclopedia

I would like to know about the history of the limit state concept. More clearly I would want to know why people came to think of limit states and how they began to discard the previous concept - the permissible stress concept.--Arun T Jayapal 10:28, 21 January 2006 (UTC)[reply]

Can any body help me with the detailed philosophy of limit state Design —Preceding unsigned comment added by 77.220.15.133 (talk) 01:30, 11 July 2009 (UTC)[reply]

Working/allowable stress design generally treats all loads equally (except when combining loads) and factors strength down for safety. Limit states design factors up loads using various factors depending on their certainty and only factors strength down by a small amount. For example, structural self weight is more certain and can be factored up by a smaller amount than uncertain live loads like traffic. Safety is still maintained in limit states, but can allow for higher live load due to reduction in factors for dead loads.
There is a good treatment of ASD vs. LRFD in Coduto's Geotechnical Engineering: Principles and Practices. Introductory structural engineering textbooks are also likely to have useful information. — Preceding unsigned comment added by 199.168.243.252 (talk) 21:18, 14 July 2014 (UTC)[reply]

In Australia, AS 4100-1998 Amdt 1 covers limit states. Unfortunately not freely available so can't be used as a Wikipedia source. — Preceding unsigned comment added by 203.129.23.146 (talk) 07:00, 6 October 2013 (UTC)[reply]

Norwegian The reference to the norwegian (bokmål) Wikipedia should have been to "Grensetilstand" (=limit state) and not to "Bruddgrensetilstand" (=ultimate limit state). Regards Arne Kvitrud 31.10.2014. — Preceding unsigned comment added by Kvitrud (talkcontribs) 11:12, 31 October 2014 (UTC)[reply]

Limit state

[edit]

DennisK SE (talk) 16:26, 18 March 2021 (UTC)I'll throw this out for now and edit later, apologies.[reply]

LRFD is mentioned as another name for limit state design both here and in another source I checked. I'm not in agreement with that. To me, limit state is an analysis of a failure state---looking at normal and shear stresses on a Mohr's circle or similar. I believe such analysis has been extended to 3 dimensions but I'm decades away from academia. (BSCE 1981, MSCE 1990).

LRFD is Load factor Resistance Design. In lieu of sizing structural members on a safe stress level (in theory, the maximum material stress reduced by a factor of safety), LRFD uses different load factors applied to different categories of loads and compares the result to what (used to be called) the ultimate strength of the member. LRFD was initially called Ultimate Strength Design (USD) to differentiate it from Allowable Stress Design (ASD).

USD 1st found general acceptance in reinforced concrete design. Back in the 70s a concrete design method was developed that used a rectangular shaped block for the concrete stress. Consider simple bending of a singly reinforced (reinforcement in the tension zone only) beam. The strain in the member varies linearly from the centroid of the section to the outer edges. The stress is assumed to also increase linearly with the distance from the centroid. (Note that that isn't true. It arises from one of the 'simplifying assumptions' made to derive easy to solve equations from the differential equation for bending---plane sections remain plane after bending. The actual shape of the stress is curved, not straight, but the intent of engineering design is to produce safe structures that can be quickly designed using simplified equations.) In reinforced concrete bending, the exterior applied moment is balanced by a couple consisting of the compression in a rectangular compression block at the compression face and tension in the reinforcing steel. Along with the new method, we got load factors so that loads with different levels of uncertainty could be given different load factors. Dead loads, the weight of the structure and finishes, (seemingly easy to calculate) was given a load factor of 1.4. Live loads, which are sort of an educated guess, were given a load factor of 1.7. Since then, different load factors have continued to pop up (1.6 for wind, etc.) along with changes to the factor for dead and live loads (differentiating between floor live loads, various types of roof live loads, flood loads, seismic loads, etc.).

The load factors look like a different way of considering material strength such as we do for wood. Wood, however, is a totally different type of material in that it has different properties in different directions and it experiences significant creep. Creep is deflection caused by sustained loads. A structural member will deflect as it is initially loaded (elastic deformation) then most material will continue to deform under sustained loads (usually blamed on migration of moisture in the material). Single reinforced concrete will experience creep deformation, over time, that is approximately equal to the initial elastic deformation. In addition, all materials respond differently to rapidly loads than they do to slowly applied loads (static loads). So wood has always been given a load factor. recognizing that it will resist deformation from rapidly applied loads but experience more deformation to static loads. In past codes, steel was also given different allowable stresses for rapidly applied loads including wind and seismic, but most people, including plan checkers for large jurisdictions, were not aware of that.

The 1st problem with LRFD is that the easy to calculate loads aren't easy to calculate. Analysis of failures usually show that the design dead load was exceeded for the actual structure. Although live loads are usually well below design assumptions at failure (the Hyatt Hotel walkway being a notable exception).

The other problem with LRFD is that it seems so alluring to academics and code writers that they, in my opinion, have gotten way out of line. Example: the wind load provisions beginning with the 2010 edition of ASCE7. We were surprised to see that wind speeds had suddenly gotten faster in the 2010 edition of that standard. What ASCE did was increase wind speeds on maps so that the resulting pressure, using the same old equation, were pre-factored. This allowed for a load factor of 1.0 in load combinations including wind. Think about that for a minute. It is analogous to publishing a new density for concrete of 195 pcf so that one could use 1.0 times self weight in stead of 1.3 in one of 6 different load combinations. However, someone may want to know the actual force (not the factored force) on an object. I recently had to sort through the ASCE wind provisions to see whether a European window manufacturer was using a wind pressure, for a large window, that I was comfortable with. I also had to provide an uplift wind pressure to a joist manufacturer (supposedly) so they could specify the nailing schedule for roof sheathing (something I should have control over because it relates to the lateral capacity of the building). Also, the minimum nailing in the wood code for wind and seismic would easily resist uplift in any non-hurricane wind situation. It points out that nobody wants to use the wind speed to determine applied pressure anymore. (My drawings used the wind speed as given by current ATC maps via an on-line site that gives design parameters based on zip code). I gave them the number in a table in ASCE7-10 although I recognize that the actual force involved should have been that number divided by 1.6.

Enough whining about LRFD. This should be about limit state design except no body knows what that is. In Canada, they use a limit state design approach that results in the allowable loads for Simpson beam hangers to have half the value in Canada that they do in the US. This is because the Canadians are using a triangular pressure block under the bearing flange of the hanger instead of the US assumption of a uniform pressure. Although over conservative, the Canadians are correct in their assumption and that interpretation is closer to being a limit state than LRFD which isn't, actually, related to a limit state at all.