How does EN 1993-1-1 determine whether second-order effects must be considered?

EN 1993-1-1 Clause 5.2.1 uses the elastic critical load factor αcr. If αcr ≥ 10, the frame is non-sway and first-order analysis suffices. If 3 ≤ αcr < 10, the amplified sway moment method applies. If αcr < 3, a full second-order analysis is required. Most multi-storey steel buildings fall in the 3-10 range, triggering the amplified sway method.

What is the equivalent sway imperfection φ and how is it applied?

The equivalent sway imperfection φ = φ0 × αh × αm per EN 1993-1-1 Clause 5.3.2. The basic value φ0 = 1/200, reduced by height factor αh = 2/√h (0.67 ≤ αh ≤ 1.0) and column count factor αm = √(0.5(1 + 1/m)). Applied as equivalent horizontal forces H_i = φ × N_Ed at each storey, where N_Ed is the total vertical load above that storey.

How is the amplified sway moment method applied per EN 1993-1-1?

The amplified sway moment method (Clause 5.2.2(3)) amplifies first-order sway moments: M_II = M_I × k_amp where k_amp = 1/(1 - 1/αcr). First-order analysis runs twice: with lateral restraints for non-sway moments, without for sway moments. The sway component is amplified and added to the non-sway component. Valid only when αcr ≥ 3 and beams are laterally restrained at supports.

What is the difference between P-Δ and P-δ effects?

P-Δ effects (global) arise from gravity loads acting through inter-storey drifts. P-δ effects (local) arise from axial load acting through member curvature. EN 1993-1-1 addresses P-Δ through the αcr check and amplified sway method. P-δ is covered by member buckling checks in Clause 6.3. The two effects do not double-count — they address different physical mechanisms at different scales.

How is αcr calculated for a multi-storey frame in practice?

For each storey i: αcr = (H_Ed,i / V_Ed,i) × (h_i / δ_H,Ed,i). H_Ed is total horizontal load, V_Ed is total vertical load, h_i is storey height, and δ_H,Ed,i is relative horizontal displacement between top and bottom of storey i. Run first-order analysis with EHF from imperfections plus wind, measure inter-storey drifts, then compute αcr per storey. The lowest αcr across all storeys governs.

When can a frame be classified as non-sway and what are the implications?

A frame is non-sway when αcr ≥ 10 for ALL storeys per EN 1993-1-1 Clause 5.2.1(3). This typically requires a bracing system (X-bracing, shear walls, concrete cores) resisting at least 90% of horizontal forces. First-order analysis suffices, buckling lengths use non-sway values (L_cr ≤ L_system), and sway imperfections may be omitted. Bow imperfections for individual members still apply per Clause 5.3.2(6).

EN 1993-1-1 Frame Stability — αcr, Second-Order Effects & Sway Imperfections

Q: What is the equivalent sway imperfection φ and how is it applied?

The equivalent sway imperfection φ = φ0 × αh × αm per EN 1993-1-1 Clause 5.3.2. The basic value φ0 = 1/200, reduced by height factor αh = 2/√h (0.67 ≤ αh ≤ 1.0) and column count factor αm = √(0.5(1 + 1/m)). Applied as equivalent horizontal forces H_i = φ × N_Ed at each storey, where N_Ed is the total vertical load above that storey.

Q: How is the amplified sway moment method applied per EN 1993-1-1?

The amplified sway moment method (Clause 5.2.2(3)) amplifies first-order sway moments: M_II = M_I × k_amp where k_amp = 1/(1 - 1/αcr). First-order analysis runs twice: with lateral restraints for non-sway moments, without for sway moments. The sway component is amplified and added to the non-sway component. Valid only when αcr ≥ 3 and beams are laterally restrained at supports.

Q: What is the difference between P-Δ and P-δ effects?

P-Δ effects (global) arise from gravity loads acting through inter-storey drifts. P-δ effects (local) arise from axial load acting through member curvature. EN 1993-1-1 addresses P-Δ through the αcr check and amplified sway method. P-δ is covered by member buckling checks in Clause 6.3. The two effects do not double-count — they address different physical mechanisms at different scales.

Q: How is αcr calculated for a multi-storey frame in practice?

For each storey i: αcr = (H_Ed,i / V_Ed,i) × (h_i / δ_H,Ed,i). H_Ed is total horizontal load, V_Ed is total vertical load, h_i is storey height, and δ_H,Ed,i is relative horizontal displacement between top and bottom of storey i. Run first-order analysis with EHF from imperfections plus wind, measure inter-storey drifts, then compute αcr per storey. The lowest αcr across all storeys governs.

Q: When can a frame be classified as non-sway and what are the implications?

A frame is non-sway when αcr ≥ 10 for ALL storeys per EN 1993-1-1 Clause 5.2.1(3). This typically requires a bracing system (X-bracing, shear walls, concrete cores) resisting at least 90% of horizontal forces. First-order analysis suffices, buckling lengths use non-sway values (L_cr ≤ L_system), and sway imperfections may be omitted. Bow imperfections for individual members still apply per Clause 5.3.2(6).

Complete reference for EN 1993-1-1:2005 frame stability and second-order analysis. Covers the elastic critical load factor αcr (Clause 5.2.1), second-order P-Δ effects and the amplified sway moment method (Clause 5.2.2), equivalent sway imperfections φ and equivalent horizontal forces (Clause 5.3.2), buckling length determination in sway and non-sway frames (Clause 6.3 and Annex BB), and the storey-based approach for multi-storey frames. Includes a fully worked 6-storey steel frame example with αcr calculation, imperfection application, and amplified moment design.

Quick access: EN 1993 Beam Design | EN 1993 Column Buckling | EN 1993 Moment Connection | Beam Capacity Calculator

Frame Stability — The Fundamental Decision

Steel frames resist lateral loads through either a bracing system (non-sway frame) or frame action — the bending stiffness of beams and columns (sway frame). EN 1993-1-1 Clause 5.2 draws a bright line based on the elastic critical load factor αcr. This single parameter determines the entire analysis methodology.

αcr ≥ 10: Non-sway frame. First-order elastic analysis is sufficient. Bracing system provides at least 90% of lateral stiffness. Buckling lengths for columns may use the in-plane non-sway values (typically L_cr ≤ L_system). Equivalent sway imperfections may be omitted entirely.

3 ≤ αcr < 10: Sway frame, moderate sensitivity. Second-order effects cannot be neglected but may be approximated by the amplified sway moment method (Clause 5.2.2(3)). Buckling lengths must reflect the sway condition. This is the most common range for unbraced multi-storey steel frames.

αcr < 3: Sway frame, high sensitivity. A full second-order analysis (geometric non-linear, P-Δ-δ) is required. First-order or amplified methods are not permitted. Frame stiffness is insufficient; consider adding bracing or stiffening members.

In practice, EN 1993-1-1 drives designers toward braced frames (αcr ≥ 10) wherever possible because second-order effects substantially complicate both analysis and member design.

Elastic Critical Load Factor αcr — Clause 5.2.1

The αcr check answers one question: by what factor could we multiply the design loads before the frame reaches elastic buckling under sway? A higher αcr means a stiffer, more stable frame.

Storey-Based αcr Formula

For each storey i:

α_cr = (H_Ed,i × h_i) / (V_Ed,i × δ_H,Ed,i)

Where:

H_Ed,i = total horizontal design load at storey i (including EHF from imperfections + wind)
V_Ed,i = total vertical design load at storey i (dead + live above that level, in load combination)
h_i = storey height (floor-to-floor)
δ_H,Ed,i = relative horizontal displacement between top and bottom of storey i under H_Ed

The simplicity is deceptive: you need a full first-order analysis to get δ_H,Ed, and then compute αcr at every storey. The lowest αcr governs the frame.

Hand Estimate (Vianello's Method)

For a preliminary check during scheme design:

α_cr ≈ (π² × ΣEI_col) / (V_Ed × h_storey²)

The term π²ΣEI/h² is the sum of column Euler buckling loads divided by the storey shear. This is conservative for frames with stiff beams and fixed bases — use it to flag frames that will certainly fail the αcr check and require bracing.

Worked Example — αcr for a 3-Bay × 6-Storey Frame

Geometry: 6 storeys at 3.5 m = 21 m total height, 3 bays at 7.5 m = 22.5 m width. Columns: UKC 254×254×89 (S355). Beams: UKB 457×191×67 (S355). Simple construction with vertical X-bracing in end bays.

Loads per storey (ULS combination 6.10): G_k = 4.5 kN/m² × 7.5 m × 22.5 m = 759 kN dead load per floor. Q_k = 3.0 kN/m² × 3 bays × 7.5 m × 7.5 m = 506 kN imposed. Total V_Ed per floor = 1.35 × 759 + 1.5 × 506 = 1784 kN. Cumulative V_Ed at ground = 6 × 1784 = 10,704 kN.

Horizontal load (EHF from imperfections, no wind for this check): φ = 1/200 × (2/√21) × √(0.5(1+1/4)) = 1/252. H_EHF per floor = φ × V_Ed = 1784/252 = 7.1 kN. Cumulative H_Ed at ground = 42.5 kN.

First-order sway: Under the 7.1 kN per floor, the top-of-frame sway from a quick analysis is approximately δ_top = 18 mm. Inter-storey drift at storey 1 (ground to first): δ_H,1 = 3.8 mm = 0.0038 m.

αcr for storey 1: α_cr = (42.5 × 3.5) / (10,704 × 0.0038) = 148.75 / 40.68 = 3.66.

Result: αcr = 3.66 falls in the 3-10 range — sway-sensitive. Second-order effects must be considered using the amplified sway moment method. Full second-order analysis is not required (αcr > 3).

Design decision: For αcr = 3.66, the sway moment amplifier k_amp = 1/(1 - 1/3.66) = 1.38. All first-order sway moments are multiplied by 1.38. This is a manageable amplification. If k_amp had exceeded 1.5 (αcr < 3), the frame would be reworked — likely by adding bracing or increasing column sections.

Amplified Sway Moment Method — Clause 5.2.2

When 3 ≤ αcr < 10, EN 1993-1-1 permits the amplified sway moment method in lieu of a full second-order analysis. The procedure:

Run first-order analysis WITH lateral restraints at each floor to obtain non-sway moments M_I,ns and axial forces N_Ed.
Run first-order analysis WITHOUT lateral restraints to obtain total moments M_I.
Extract sway moments: M_I,sway = M_I - M_I,ns.
Compute amplifier: k_amp = 1/(1 - 1/αcr).
Design moment: M_II,Ed = M_I,ns + k_amp × M_I,sway.

The non-sway moments (from pattern loading, beam end fixity) do not need amplification because P-δ effects are separately accounted for in the member buckling check using the appropriate buckling length and buckling curve.

Storey-Based Amplification — Clause 5.2.2(6)

When αcr varies between storeys, apply storey-specific amplifiers:

k_amp,i = 1 / (1 - 1/α_cr,i)

Where α_cr,i is computed for each storey i. The amplification is applied to the sway moment diagram piecewise — each storey's sway component is amplified by its local k_amp,i before combining with non-sway moments.

Equivalent Sway Imperfections — Clause 5.3.2

Real frames are never perfectly plumb. EN 1993-1-1 accounts for erection tolerances and residual stresses through the equivalent sway imperfection φ.

Imperfection Components

φ = φ_0 × α_h × α_m

Parameter	Symbol	Description	Value / Range
Basic value	φ_0	Basic sway imperfection	1/200
Height	α_h	Reduction for column height h (metres)	2/√h, 0.67-1.0
Columns	α_m	Reduction for number of columns m in a row	√(0.5(1+1/m))

The α_h factor recognises that tall columns are built more accurately because the same absolute tolerance produces a smaller out-of-plumb angle. The α_m factor accounts for statistical reduction: in a row of m columns, the worst one does not govern because adjacent columns restrain it.

Application as Equivalent Horizontal Forces (EHF)

At each storey level i, apply a notional horizontal force:

H_EHF,i = φ × (N_Ed,i - N_Ed,i+1)

Where (N_Ed,i - N_Ed,i+1) is the total vertical load applied at that storey. For the roof level, N_Ed,roof+1 = 0. The EHF may be applied in any horizontal direction but must act simultaneously with gravity loads and wind in the load combination.

Full Frame Stability Worked Example

6-storey unbraced steel office building, 21 m tall, S355 steel.

Step 1: Equivalent Sway Imperfection. h = 21 m, α_h = 2/√21 = 0.436 → capped at 0.67 (minimum). m = 4 columns per row, α_m = √(0.5(1+1/4)) = √0.625 = 0.791. φ = (1/200) × 0.67 × 0.791 = 1/378.

Step 2: Equivalent Horizontal Forces. Total vertical load at ground: V_Ed = 1.35 × (6 × 759) + 1.5 × (6 × 0.7 × 506) = 6148 + 3188 = 9336 kN. EHF at each floor = φ × floor load = 1784/378 = 4.7 kN per floor. Total EHF = 28.3 kN.

Step 3: First-Order Analysis. Run global elastic analysis with EHF applied at each floor. Extract inter-storey drifts:

Storey	h_i (m)	δ_H (mm)	V_Ed (kN, cumul.)	H_EHF (kN, cumul.)	α_cr
1	3.5	4.2	9336	28.3	2.53
2	3.5	3.8	7552	23.6	2.88
3	3.5	3.3	5768	18.9	3.42
4	3.5	2.7	3984	14.2	4.8
5	3.5	1.9	2200	9.5	6.9
6	3.5	1.0	892	4.7	14.1

Storey 1 fails — αcr = 2.53 < 3. The frame is too flexible. Full second-order analysis would be required, but this indicates the frame is under-designed.

Design decision: Add vertical X-bracing (SHS 150×150×10, S355) in the two end bays. With bracing, αcr jumps to > 20 (non-sway). The amplified sway method is no longer needed. Column buckling lengths revert to non-sway values. This is the recommended design strategy for UK multi-storey commercial buildings.

Second-Order P-Δ and P-δ Effects

Understanding the distinction avoids double-counting:

Effect	Characteristic	EN 1993 Treatment
P-Δ	Global — gravity × storey drift, affects all members	αcr check + amplified sway method
P-δ	Local — axial load × member curvature, affects member	Member buckling check (Clause 6.3)

The αcr check and amplified sway method handle P-Δ. P-δ is addressed through the member stability check using appropriate buckling lengths. Doing both the amplified sway analysis AND a member buckling check with effective lengths appropriate for the sway condition does NOT double-count — they address different physical mechanisms.

A common error is applying sway amplification to a frame that has already been analysed with geometric non-linearity turned on in the software. The amplification is only applied to first-order sway moments; if analysis is already second-order, do not amplify further.

Practical Design Rules for Frame Stability

Frame Type	Typical αcr	Design Approach
Portal frame, haunched, 15-30 m span	6-12	Check αcr; often needs bracing over 20 m
4-storey unbraced, stiff beams	4-8	Amplified sway method, monitor column sizes
4-storey braced (X-bracing)	20-50	First-order, non-sway lengths
8-storey unbraced	2-5	Almost always needs bracing or core
8-storey core-braced	25-60	First-order by default

Six Rules for Stability Design

Check αcr first, not last. It is the first analysis output to review after the initial run. If αcr < 3, do not proceed to member design — stiffen the frame first.
Bracing is cheaper than column steel. The tonnage of additional column steel to raise αcr from 3 to 10 is typically 2-3× the tonnage of a simple X-brace system.
Apply imperfections in BOTH directions. EN 1993-1-1 Clause 5.3.2(7) requires EHF to be applied in any direction that produces the worst effect. For a rectangular building, apply in the two orthogonal directions separately.
Account for cracked concrete in composite frames. If the frame relies on a concrete core for stability, reduce the EI of the core by 50% (cracked section) per EN 1992-1-1 Clause 5.4.
The αcr for wind combination is different from the αcr for gravity-only. EN 1993-1-1 requires checking αcr under all relevant load combinations because the vertical load V_Ed varies.
Software non-linear analysis: run first-order first. Most structural analysis packages offer P-Δ analysis as a checkbox. Run the model linearly first to compute αcr. If αcr ≥ 10, no need for non-linear. If 3 ≤ αcr < 10, decide between amplified sway and P-Δ analysis.

EN 1993 vs Other Codes

Criterion	EN 1993-1-1	AISC 360	AS 4100
Sway classification	αcr ≥ 10 = non-sway	B2 ≤ 1.1	λ_ms ≤ 0.2 × storey
Second-order method	Amplified sway moments	B1-B2 amplification	Moment magnifier
Imperfection φ	φ = φ_0 × α_h × α_m	Δ_o = L/500	λ_c dependent
Direct analysis	Not explicitly in EN 1993	DAM (Chapter C)	Advanced analysis (Cl 4)

The philosophy differs: AISC 360 provides the Direct Analysis Method (DAM) as the primary route, embedding imperfections and stiffness reductions directly into the analysis model. EN 1993-1-1 retains the traditional αcr-based classification with imperfection application as separate steps. Both achieve the same goal — second-order effects properly accounted for — but the steps differ.

Educational reference only. Verify all design values against the current EN 1993-1-1 and the applicable National Annex for your jurisdiction. Frame stability is project-specific — αcr depends on the complete structural system and cannot be determined in isolation. Always verify with a full elastic critical load analysis. Results are PRELIMINARY — NOT FOR CONSTRUCTION without independent verification by a qualified structural engineer.