Comparison standards API-RP-2A, ISO 19902(2007) and AISC 360-10/16

Here some notable issues are discussed that were encountered when comparing the three standards.

The AISC code is one of the most flexible standards compared to the others. API and ISO consider the beam connection as rigid (fully restrained). In AISC more options are available to also partially restrain the moment connection between brace and chord and also allow rotation between the members. This is mentioned in AISC as “simple connections”.

However, the consideration how the connection is treated should be consistent throughout the whole structural calculation (see AISC 360-10 B3.6b). All three standards treat torsion and axial stresses separately, which make these standards less suitable for supported monopile structures like e.g. windmills.

For joints all standards use the same equation to combine axial load and moments:

But only in AISC it is stated that this equation is only valid for TY and X joints.  For K joints AISC states: "K-connections with moment loading on the branches are not covered by this specification", see Commentary K3; HSS to HSS moment connections.

API and ISO are for (offshore) tubular structures only. Use API-RP-2A when WSD method is required and ISO-19902 for LRFD method. AISC seems somewhat simpler with less complicated formulas, but this leads also to more conservative answers.

Furthermore, the standards according to API and ISO allow the first order analysis as acceptable, while using AISC, the engineer needs to (at least) consider a second order analysis including P-Δ and P-δ effects and other effects of initial imperfections and stiffness reductions due to inelasticity (see AISC 360-10 C2).

Try the differences between the codes using the hand calculation sheets of FEMDS.COM:

ISO 19902 edition 2007 vs 2020 and the relation with API

API-RP-2A LRFD 1st edition was released in 1993 but never commonly used and therefore retracted in 2012. On the ASCE site is written that:

"The API RP 2A-LRFD format was reintroduced in 2014 as a draft 2nd edition. This study was carried out in the context of the plan to prepare the API RP 2A-LRFD 2nd edition under the effort to align and merge the API and ISO standards as a single global standard."

Now the following can't be a coincidence:

The joint check code in ISO 19902(2007) contained strange behavior related to the Q_f parameter that could become negative, leading to unrealistic results. It is therefore notable that ISO19902 2020 edition has integrated most of the code from API-RP-2A. ISO itself announces it as:

tubular joint strength formulae nearly all changed through adoption of the API RP 2A-WSD 21st Edition Supplement 2 (October 2005) tubular joint formulae supplemented by some limited nonlinear FEA;

It will be clear that both standards are closely interrelated

Note that the one-third stress increase is removed from AISC 360-10/16.

ASCE/SEI standards no longer permit the familiar one-third stress increase in allowable stress design. This is in contrast to the API standard. When is this factor actually applicable?

AISC is originally a code for buildings that were exposed to load combinations of dead weight, live load and wind etc. The one-third stress increase has a long history starting around the year 1900. This factor was introduced at the time for load combinations of which the maximum load would never occur in practice. From AISC 360-10 it follows a new approach.  The one third stress increase is disappeared and the definitions for loadcases for buildings is defined in standard ASCE 7, see also "The Origin of the One-Third Stress Increase." For offshore and maritime industry that uses the AISC for certification, unity factors should be chosen in accordance with the requirements of the applicable certification authority. For clearness the one-third increase corresponds to a unity factor of 0.8.

In the ASCE 7-02 (issued in 2002), standard reduction factors are included in the previously described load combinations to be more realistic, because it is extremely rare that all variable loads reach their maximum at the same time. In section C2.4 some more clearance is given for combining loads in allowable stress design for buildings where the reduction factor is already incorporated such that the one-third stress increase on the material side is not permitted anymore:

"However, when more than one variable load is considered, it is extremely unlikely that they will all attain their maximum value at the same time. Accordingly, some reduction in the total of combined load effects is appropriate. This reduction is accomplished through the 0.75 load combination factor. The 0.75 factor applies only to the variable loads, not to the dead load."

Bear in mind that combinations of loadcases could lead to load levels that will never occur during life of a construction, the ASCE-7 gives some direction for reduction, but the one-third stress increase must in that case be skipped. Interesting is what is stated in this article: "The One-Third Stress Increase. Where is it now?". In the first paragraph of the conclusions an interesting statement is made:

"Code-specified ASD load combinations today provide for the proper amplification of loads, with one transient load at its maximum lifetime value and other transient load(s) reduced to their arbitrary point-in-time value(s). This accomplishes what the one-third stress increase used to accomplish on the material-strength side of the equation. Consequently, the one-third stress increase usually is inappropriate for use with current ASD load combinations. Engineers that “double-dip,” i.e., use both the one-third stress increase on the material side and reduced ASD load combinations, are violating the code in an unconservative way."

We read this as follows: If the calculation is fair so that the maximum load can actually occur during design life, then you can't apply the one-third stress increase!