Portal Frame Optimization Guide — Section Selection, Haunch Length, and Deflection Limits

A practical guide to optimizing portal frame designs for minimum weight and cost. Covers section selection trade-offs between column and rafter stiffness, haunch length optimization, the interaction between deflection limits and frame weight, and proven rules of thumb for cost-effective portal frame design.

PRELIMINARY — NOT FOR CONSTRUCTION. All results are for educational and reference use only. Must be independently verified by a licensed Professional Engineer (PE) or Structural Engineer (SE) before use in any project.

The Optimization Problem

Portal frame design has never been a purely strength-governed exercise. The governing constraint is almost always serviceability — specifically, lateral sway at the eave under wind load and vertical deflection at the rafter mid-span — rather than ultimate strength. A frame that satisfies all strength checks may still be 20–30% heavier than necessary because the designer has not tuned the stiffness distribution between columns and rafters.

The portal frame optimization problem has three interacting variables:

  1. Haunch length — Longer haunches reduce rafter moments but increase fabrication cost
  2. Column-to-rafter stiffness ratio (I_c / I_r) — Affects moment distribution and sway
  3. Base fixity — Fixed bases reduce sway but increase foundation cost

The objective is to find the combination that minimises total frame weight (columns + rafters + haunches) while satisfying strength, stability, and deflection requirements.

Section Selection: Balancing Column and Rafter Stiffness

The moment distribution in a portal frame is controlled by the relative stiffness of the column and rafter. For a pinned-base frame under uniform gravity load, the eave moment is:

M_eave = (w × L² / 8) × [ (h/L) × (I_c / I_r) ] / [ 2 + 3 × (h/L) × (I_c / I_r) ]

This relationship reveals a fundamental trade-off: increasing column stiffness (raising I_c / I_r) increases the eave moment and reduces the mid-span rafter moment, while decreasing column stiffness does the opposite.

The Optimum I_c / I_r Ratio: For typical portal frames with h/L between 0.2 and 0.4, the optimum I_c / I_r ratio (the ratio that minimises total frame weight) falls between 1.20 and 1.60. This means the column section should be 20–60% stiffer than the rafter section.

A ratio below 1.0 forces the rafter to carry more moment, requiring a heavier rafter. A ratio above 2.0 drives excessive moment into the eave, requiring a heavier haunch and heavier column, with diminishing returns on rafter weight reduction.

Practical Selection Strategy:

  1. Select an initial rafter section based on span/depth ratio (depth = span/40 to span/55). For a 20 m span, start with a 450–500 mm deep section.
  2. Select a column section with I_c / I_r ≈ 1.3. For the rafter above (I_r ≈ 147 × 10⁶ mm⁴), choose a column with I_c ≈ 190 × 10⁶ mm⁴.
  3. Analyse the frame. If the eave moment is less than 60% of the column moment capacity and the mid-span moment is greater than 80% of the rafter capacity, the rafter is too light — increase rafter stiffness.
  4. If the eave moment exceeds 80% of the column capacity, increase either column stiffness or haunch length.
  5. If sway exceeds the h/300 limit, increase column stiffness or specify a fixed/nominally fixed base.

A common optimisation mistake is selecting identical sections for columns and rafters (I_c / I_r = 1.0). While this simplifies procurement, it increases total frame weight by 10–15% compared to the optimum because the columns are under-utilised while the rafter governs.

Haunch Length Optimization

The haunch serves two purposes: it increases the moment capacity at the eave (the most highly stressed region), and it stiffens the frame, reducing lateral sway. The haunch length is the single most powerful optimisation variable because it affects both strength and stiffness.

Effect on Rafter Moments: Increasing the haunch length from 10% to 15% of the span reduces the mid-span rafter moment by 8–12% because the increased stiffness at the eave attracts more moment away from the rafter. However, beyond 15% of span, the benefit plateaus — the additional steel in the haunch exceeds the saving in the rafter, and total frame weight increases.

Effect on Frame Sway: The haunch stiffens the eave joint, reducing frame sway by 15–20% for a haunch length increase from 10% to 15% of span. This is often the cheapest way to solve a drift problem — adding 1 meter to the haunch costs less than upsizing both columns by one section.

Economic Optimum: For hot-rolled sections, the optimum haunch length (minimising total frame weight) is 12–14% of the span. For spans under 20 m, the optimum shifts to the lower end (10–12%) because the haunch becomes a larger fraction of the rafter length. For spans over 40 m with plate girder haunches, the optimum can extend to 16–18% because the haunch plate cost is low relative to the rolled section cost.

Fabrication Cost Consideration: Longer haunches require longer stiffened web panels, more flange-to-web welding, and larger end plates at the eave. Fabrication cost increases roughly linearly with haunch length. A haunch of 2.5 m may add 15% to the fabrication cost compared to 1.5 m, even if the material cost is only 5% higher. In regions with high labour costs (Western Europe, North America, Australia), the optimum haunch length shifts toward the shorter end of the range.

The Deflection Limit Trade-Off

Deflection limits are the hidden driver of portal frame weight. A frame that satisfies all strength checks at 100% utilisation may still fail the L/180 (rafters) or h/300 (sway) deflection limits. Addressing deflection without over-design requires understanding which variable controls each deflection mode.

Rafter Vertical Deflection: Governed by rafter stiffness (Ir) and the eave moment (which is controlled by I_c / I_r). The most efficient way to reduce rafter deflection is to increase I_r (upsize the rafter). Increasing haunch length helps moderately (5–8% reduction). Increasing column stiffness actually _increases rafter deflection because it drives more moment into the rafter by stiffening the eave restraint — counter-intuitive but important.

Lateral Sway at Eave: Governed by column stiffness (I_c) and base fixity. Increasing column stiffness is the direct — and most expensive — solution. Two cheaper alternatives:

  1. Haunch stiffening: Lengthening the haunch reduces sway by 15–20% at minimal material cost
  2. Nominally fixed base: A base detail with four anchor bolts arranged outside the column flanges can achieve 20–40% of the full fixity moment, reducing sway by 15–30% without the foundation cost of a fully fixed base

Pre-Camber: Where deflection is visible but non-structural (e.g., no brittle finishes), pre-cambering the rafter to offset dead load deflection can satisfy the visual requirement without additional steel. AISC Code of Standard Practice Section 7.13 recommends camber equal to 75% of the dead load deflection.

Weight Minimization: A Systematic Approach

A systematic approach to minimising portal frame weight:

  1. Start with a pinned base. This is the default. Move to a nominally fixed base only if sway exceeds h/300.

  2. Optimise I_c / I_r to 1.2–1.6. Select column and rafter sections from the same section family (e.g., both W460 or both 410 UB) to simplify procurement.

  3. Set haunch length to 12% of span. Analyse the frame. If all strength checks pass with utilisation > 70%, the frame is competitive.

  4. Check deflection. If rafter deflection governs, upsize the rafter by one section (add 25–50 mm depth). If sway governs, try lengthening the haunch by 2% of span before upsizing the column.

  5. Check the uplift (wind) case. In light roof construction with high wind, the uplift combination may govern — the bottom chord (in compression under uplift) and the end diagonals (also in compression) may require heavier sections than gravity alone.

  6. Iterate once. After the first adjustment, re-analyse. Portal frame optimisation converges quickly — two iterations are usually sufficient.

Worked Example — Optimisation:

Starting frame: 25 m span, 7 m eave height, 8° pitch, 6 m spacing. Rafter W460 × 52 (I = 147 × 10⁶ mm⁴), column W460 × 52 (I = 147 × 10⁶ mm⁴), haunch = 2.5 m (10% span). I_c / I_r = 1.0.

Optimisation: Increase I_c / I_r to 1.3 by upsizing column to W460 × 68 (I = 193 × 10⁶ mm⁴). Lengthen haunch to 3.0 m (12% span).

The frame required only a column upsize and haunch lengthening — the rafter, which was adequate for strength, was left unchanged. A less targeted approach (upsizing both column and rafter) would have added 350+ kg.

Flange and Web Proportioning for Plate Haunches

For plate girder haunches (fabricated from plate rather than rolled sections), the flange and web proportions can be optimised independently:

The transition from the haunch flange to the rafter flange requires a butt weld with backing or a bolted splice. The haunch web must be continuously welded to the rafter web for shear transfer.

Optimization for Different Steel Grades

Using higher-strength steel (A572 Gr 50, S355, AS/NZS 3678 Grade 350) instead of A36/S235 can reduce frame weight by 20–25%. However, the benefit is partially offset by:

  1. Deflection control: Stiffness (E × I) is independent of steel grade. Higher-strength steel does not reduce deflections. The governing case may shift from strength to serviceability, negating the weight saving.

  2. Local buckling limits: Higher-strength steel has stricter width-to-thickness limits for compact sections. A section that works in A36 may be classified as noncompact in A572 Gr 50, reducing the available capacity.

  3. Cost per tonne: High-strength steel costs 10–20% more per tonne. The economic benefit depends on whether the weight saving (proportional to section area) exceeds the unit cost premium.

For portal frames where deflection governs (the common case), the weight saving from high-strength steel is typically 5–10%, not the 20–25% suggested by yield strength alone. The decision to use high-strength steel should be based on total project economics, not just frame weight.

Engineering Best Practices

FAQ

Q: What is the most cost-effective way to reduce portal frame sway? A: In order of cost-effectiveness: (1) lengthen the haunch by 2–3% of span — essentially zero additional fabrication cost beyond the extra steel, reduces sway 15–20%; (2) detail the base with four anchor bolts outside the column flanges to achieve partial fixity — reduces sway 15–30% at the cost of slightly larger anchor bolts and base plate; (3) upsize the column — the most expensive option but guaranteed to work. Avoid upsizing both column and rafter simultaneously.

Q: How much does pre-camber save in frame weight? A: Pre-cambering the rafter to offset dead load deflection (typically 30–50 mm for a 20 m span) can save 5–10% of the rafter weight because it addresses the visual deflection requirement without additional stiffness. However, pre-camber does not help with live load deflection or sway, so the saving is limited to cases where dead load deflection alone governs.

Q: Should I always use compact sections for portal frames? A: For frames designed plastically (Class 1/compact), yes — plastic hinges require adequate rotation capacity. For elastically designed frames, noncompact sections are permitted, but the reduced moment capacity (M_n instead of M_p) and increased deflection sensitivity usually make them uneconomical. Compact sections represent the sweet spot for portal frame economics.

Q: What is the optimum frame spacing for an industrial building? A: The optimum frame spacing balances the cost of primary steel (frames, which get heavier with wider spacing) against the cost of secondary steel (purlins and girts, which get heavier with wider spacing) and foundations. For typical industrial buildings with metal cladding, the optimum spacing is 6–8 m. For buildings with precast concrete wall panels (heavier than metal cladding), the optimum shifts to 5–6 m.

References

Try Our Free Calculator

Try our free steel design calculator to optimise portal frames, check member capacities, and evaluate deflection and sway per AISC 360, EN 1993, and AS 4100. No signup required. Get instant capacity checks and step-by-step calculation reports.

Try it now: Optimise your portal frame design with our free calculator →