IS 800:2007 vs AISC 360 for PEB frame design

Why the comparison matters for PEB work

Pre-engineered buildings in India sit at an interesting crossroads. The industry grew out of US metal building practice, so many PEB suppliers learned their trade on AISC and MBMA. At the same time, government and many private projects in India require IS 800:2007. It is not unusual for one design office to run an AISC job in the morning and an IS 800 job in the afternoon, sometimes for buildings that look almost identical.

Because the codes reach their answers differently, the same frame under the same loads can end up with different flange and web plates. If you are pricing a project, or reviewing a design done under the other code, you need to know where those differences come from.

Design philosophy and safety factors

IS 800:2007 is a limit state code. You factor the loads using the partial safety factors in Table 4 and compare the effects with design strengths that are themselves reduced by material factors: γ_m0 = 1.10 for resistance governed by yielding and buckling, and γ_m1 = 1.25 for resistance governed by ultimate stress (net-section fracture, bolts and welds have their own values).

AISC 360 gives you a choice of two formats on the same nominal strength R_n:

• LRFD: φR_n ≥ R_u, with factored load combinations and resistance factors such as φ_b = 0.90 for flexure and φ_c = 0.90 for compression.
• ASD: R_n/Ω ≥ R_a, with service-level load combinations and safety factors such as Ω_b = Ω_c = 1.67.

We go deeper into the two AISC formats in LRFD vs ASD in steel design. For this comparison the key point is that IS 800’s method is structurally closest to AISC LRFD: factor the loads up, factor the resistance down.

Loads and load combinations

The codes draw their loads from different places. IS 800 works with loads from IS 875 (Parts 1 to 3 for dead, imposed and wind, with IS 875 Part 3:2015 for wind) and IS 1893 for earthquake. AISC 360 relies on ASCE 7 or the governing building code.

The rows simplify both codes for comparison; your project’s full set of combinations, including collateral, crane and seismic cases, comes from the governing standards.

For light PEB roofs, the uplift case with dead load reduced to 0.9 often governs the rafters and the outer column flange. Because roof dead load is small, the net effect is sensitive to the wind pressures, which is one reason the two codes can disagree: the wind loads themselves (IS 875 Part 3 vs ASCE 7) are different before any factors are applied.

Section classification

Both codes classify the section by the width-to-thickness ratios of its plates, and both treat welded built-up sections more strictly than rolled ones.

• IS 800 Table 2 sorts sections into plastic, compact, semi-compact and slender, with limits expressed in ε = √(250/f_y).
• AISC Table B4.1 uses compact, noncompact and slender for flexure (with λ_p and λ_r), and nonslender or slender for compression.

The categories line up loosely (plastic and compact in IS 800 both map to AISC compact for flexure), but the boundaries are not identical. A flange that is semi-compact in IS 800 might be compact or noncompact under AISC depending on grade and web slenderness. For a tapered member the web class can change along the length, so classification has to be checked at several stations. Our guide to tapered built-up I-sections gives the actual limits.

Lateral-torsional buckling

Lateral-torsional buckling often decides the rafters near the haunch and the columns under uplift, so the method matters.

IS 800 clause 8.2.2 computes the elastic critical moment M_cr, a non-dimensional slenderness λ_LT, and a reduction factor χ_LT from a buckling curve. The imperfection factor α_LT is 0.21 for rolled sections and 0.49 for welded sections, which is a significant penalty for built-up members.

AISC 360 Chapter F uses the unbraced length against limits L_p and L_r, with a moment gradient factor C_b that rewards non-uniform moment. For web-tapered members, AISC Design Guide 25 extends the approach with stress-ratio methods.

In both codes the biggest lever is the same: bracing. Flange braces from purlins and girts shorten the unbraced length of the compression flange. If the bracing in your model does not match what will be built, neither code will save you.

Combined axial force and bending

Columns and rafters in a portal frame carry axial force and bending together, so the interaction check is usually the governing one.

• IS 800 clause 9.3 has a cross-section check (9.3.1) and member buckling checks (9.3.2.2) with interaction factors k_y, k_z and k_LT that depend on the axial load ratio and the equivalent uniform moment factors.
• AISC 360 Chapter H uses the familiar bilinear equations: for P_r/P_c ≥ 0.2, P_r/P_c + 8/9 (M_rx/M_cx + M_ry/M_cy) ≤ 1.0, with a different form below 0.2. Second-order effects come from the stability analysis required by Chapter C (the direct analysis method or an alternative).

Because the shapes of these interaction curves differ, a member with a modest axial force and a large moment can come out lighter under one code and a member with high axial force lighter under the other.

So which code gives the lighter frame?

Honestly, it depends. We have seen frames come out lighter under each code. The things that swing it are:

• the wind code and site wind speed, which change the governing combination before any steel design starts;
• the ratio of dead to live and wind load;
• the steel grade (E250 or E350 to IS 2062, or A572 Grade 50 and similar under ASTM);
• how well the frame is braced;
• and the local buckling and LTB treatment of welded sections.

If you are bidding a project where the client allows either code, the sound approach is to optimise the same frame both ways and compare. With the PEB Optimization Tool that means switching the design code in your STAAD.Pro model (IS 800:2007, AISC 360 LRFD or ASD) and running the optimisation again; it reads the code from the model and searches flange, web and taper against that code’s clauses. You can preview the idea in the browser bench.

Whichever code you use, the lightest passing section is almost never the first one that passes, which is the whole argument for PEB optimisation in the first place.