Steel beams and columns typically experience a combination of axial and bending loads. Engineers and Steel design codes realize that there is a need to examine steel beams under combined axial and bending loading. Research in the field progressed the general understanding of how steel beams behave under single and combined loads.

In particular, since most steel sections are shaped as I members (flange and web), researchers progressively realized that many modes of flexural faiiure of the flange or the web could exist. As a result, codes have evolved to include flexural torsional buckling of the flange or localized buckling of the web. In regards to axial loads, the Euler equations have evolved and been modified to better match the observed behavior of steel columns. Many investigators in the AISC committee have proposed various equations that essentially yielded the same results (AISC: American Institute of Steel Construction).

Exhibit 1: Steel pipe corner braces for a diaphragm wall excavation in Boston.

Before the early 1990's all structural design was performed with the allowable stress method (ASD: Allowable Stress Design) where the material strength was reduced to an allowable value depending on the loading type (axial, bending, weak or strong axis etc), and service loads were applied. Equations tended to be relatively simple and straight forward and designs were generaly conservative. Equations for complicated modes of flexural torsional buckling were not provided but the user was encouraged to investigate according to published literature where such analysis was warranted.

Since the latest edition of ASD 9th edition, Steel-design philosophy has evolved towards Load Factored Resistance Design (LRFD). In this design philosophy, loads are mutliplied by appropriate factors depending on their nature (Dead load, live load, seismic, wind etc). The recommended factors are minimum recommended factors and the designer is permitted to use higher load multipliers when deemed appropriate. Once ultimate loads are obtained from structural analysis, the steel sections must be examined against their utlimate capacity in the respective mode of failure and in the combined stress condition. LRFD design equations are more explicit and sligthly more complicated for repetitive calculations than ASD equations. However, the latest editions of LRFD (2nd, and 3rd editions) provide detailed equations of flexural-torsional behavior of I-shaped sections. These manuals also recognize that there is no "totally" safe combined stress equation.

General experience with LRFD is that designs tend to be less conservative than the ASD approach. In regards to deep excavations, while some engineers and mostly some academic investigators feel that LRFD should be applied the majority of practicioners are hesitant to abandon the ASD approach. Indeed, LRFD may not be the best way to consider structural design of steel beams in deep excavations because it does not necessarily make sence to factor soil loads or hydrostatic loads uniformly as the resulting wall deflections from such a multiplication would probably ignore all the effects of construction staging. As an alternative, it is the feeling of the author that when LRFD equations loads should be obtained by multiplying the service reactions by an appropriate load factor in the order of 1.6 to 2.0 depending on the degree of conservatism and uncertainty in soil strengths and water pressures.

Steel-beam enables such evaluation of the combined stress condition of beams and columns very quickly with built in data bases of pipes and I-sections (US). Hence, designs can be optimized in less time and design costs on simple equation calculations can be minimized. Thus more of the design budget can be allowed for evaluating other important aspects of deep excavations.