P-Delta Plastic Analysis: Second-Order Frame Mechanics (2026)

1. Introduction to Geometric Nonlinearity in Plastic Frame Collapse

First-order plastic limit analysis assumes that structural equilibrium can be formulated on the undeformed geometry of the framework. While this assumption is valid for stiff braced systems, it significantly overestimates the collapse capacity of unbraced, sway-sensitive structural steel frames. In flexible frameworks, vertical gravity loads acting through lateral joint displacements produce secondary destabilizing overturning moments. Conducting a rigorous p-delta plastic analysis allows structural engineers to capture this interaction between material yielding and geometric nonlinearity.

                   FIRST-ORDER VS SECOND-ORDER PLASTIC COLLAPSE
     Load Factor λ
           ^
           |          First-Order Rigid-Plastic Collapse Plateau (λp)
       λp -+ - - - - - - - - - - - - - - - - - - - - - - - - - - -
           |                     * * * * *
           |               * *               * * (Negative Post-Yield Stiffness)
       λf -+ - - - - - - * <--- True Second-Order Failure Peak (λf < λp)
           |           *                         *
           |         *                             *
           |       *                                 *
           |     *   Elastic P-Δ Response              *
           |   *                                         *
       0.0 +----------------------------------------------------------> Sway Drift Δ

When evaluating p-delta plastic analysis, engineers observe that second order plastic effects erode the theoretical first-order collapse plateau $\lambda_p$, replacing it with a reduced peak load factor $\lambda_f$ followed by a descending post-yield response. To estimate this reduced capacity without non-linear finite element modeling, structural mechanics utilizes the classical merchant-rankine formula.

Understanding these second-order destabilizing mechanisms is essential to prevent sudden sway instability in high-rise frameworks and unbraced industrial portals.

2. Mechanics of P-Delta and P-delta Secondary Effects

2.1 Large Displacement Sway ($P$-$\Delta$) vs Member Curvature ($P$-$\delta$)

In structural stability theory, second-order effects are categorized into two distinct geometric phenomena:

1. Global Sway Effects ($P$-$\Delta$): The secondary overturning moment caused by total gravity load $P$ acting through the relative lateral inter-story sway displacement $\Delta$ between adjacent floor levels.
2. Local Member Curvature Effects ($P$-$\delta$): The secondary internal bending moment caused by axial compression $P$ acting through the lateral displacement $\delta(x)$ of a member relative to its chord centerline.

       GLOBAL P-Δ EFFECT (Sway)                   LOCAL P-δ EFFECT (Curvature)

       P (Gravity Load)                            P
       v                                           v
       +---------------+                           |
       |               |                           |    δ(x) (Chord Curvature)
       |               |                           (=====)
       |               |                           |
       |        Δ      |                           |
       +------->+------+                           v
                Secondary Moment = P * Δ           P    Secondary Moment = P * δ(x)

In unbraced framed structures subjected to combined lateral and gravity loads, the global $P$-$\Delta$ sway effect dominates structural behavior and governs the reduction in plastic collapse capacity.

2.2 Negative Post-Yield Frame Stiffness

Under first-order rigid-plastic theory, once a complete plastic hinge mechanism forms, the frame undergoes plastic deformation under a constant collapse load factor $\lambda_p$ (zero post-yield stiffness, $K_{post} = 0$).

However, when second-order geometry is included, each incremental increase in sway displacement $d\Delta$ generates an additional overturning moment $P \cdot d\Delta$. Because the plastic hinge capacities remain constant ($M_p$), the external horizontal load $H$ must decrease to preserve equilibrium:

$$H(\Delta) = H_{first-order} – \frac{\sum P}{h} \Delta$$

This generates a negative post-yield stiffness:

$$K_{post} = \frac{dH}{d\Delta} = -\frac{\sum P}{h} < 0$$

The structure transitions from a stable plateau to an unstable softening path, where dynamic collapse accelerates unless lateral displacements are strictly bounded.

3. The Merchant-Rankine Empirical Formulation for Imperfect Frames

3.1 Classical Merchant-Rankine Formula Derivation

In 1954, W. Merchant proposed an empirical interaction formula to predict the true failure load factor $\lambda_f$ of real framed structures by combining:
1. The theoretical first-order rigid-plastic collapse load factor $\lambda_p$, and
2. The elastic critical buckling load factor $\lambda_{cr}$ (Euler sway buckling load of the overall frame).

The merchant-rankine formula is expressed as:

$$\frac{1}{\lambda_f} = \frac{1}{\lambda_p} + \frac{1}{\lambda_{cr}} \implies \lambda_f = \frac{\lambda_p}{1 + \frac{\lambda_p}{\lambda_{cr}}} = \frac{\lambda_p \lambda_{cr}}{\lambda_p + \lambda_{cr}}$$

3.2 Modified Merchant-Rankine Formula and Stability Limits

Extensive experimental testing on steel portal frames demonstrated that strain hardening often compensates for minor second-order effects in stiff frames. Consequently, Wood and Horne proposed the Modified Merchant-Rankine Formula:

• For $\frac{\lambda_{cr}}{\lambda_p} \ge 10$:
  $$\lambda_f = \lambda_p \quad (\text{Second-order effects negligible})$$

• For $4 \le \frac{\lambda_{cr}}{\lambda_p} < 10$:
  $$\lambda_f = \frac{\lambda_p}{0.9 + \frac{\lambda_p}{\lambda_{cr}}}$$

• For $\frac{\lambda_{cr}}{\lambda_p} < 4$:
  The frame is too flexible for simple empirical modification; rigorous geometrically nonlinear elastoplastic analysis is required.

4. Equilibrium Equations of Second-Order Plastic Mechanisms

For a single-story unbraced portal frame with column height $h$, total vertical gravity load $V = \sum P$, and lateral load $H$, the virtual work equation at finite sway displacement $\Delta$ is:

$$W_{ext} = H \cdot \Delta + V \cdot \Delta_{vert}$$

For a rigid-body sway mechanism, the downward vertical movement of the rafter is $\Delta_{vert} \approx \frac{\Delta^2}{2h}$, which is a higher-order term. However, the external work done by the vertical load $V$ acting through the lateral displacement $\Delta$ at joint rotations $\theta = \Delta / h$ introduces a first-order work increment $V \Delta \theta$:

$$(H \cdot h) \theta + V \Delta \theta = \sum_{j=1}^{N_h} M_{p,j} \theta_j$$

Dividing by $\theta$:

$$H \cdot h + V \cdot \Delta = \sum_{j=1}^{N_h} M_{p,j} \left( \frac{\theta_j}{\theta} ight)$$

The required lateral load capacity at finite sway $\Delta$ is:

$$H(\Delta) = \frac{\sum M_{p,j} (\theta_j / \theta)}{h} – \frac{V}{h} \Delta = H_p – \left( \frac{V}{h} ight) \Delta$$

where $H_p$ is the classical first-order plastic collapse load.

5. Step-by-Step Worked Calculation: Unbraced Portal Frame Second-Order Capacity

To illustrate the quantitative impact of p-delta plastic analysis, consider a single-story, single-bay fixed-base steel portal frame.

                 V = 400 kN (Total Vertical Gravity Load)
                 v
       B +---------------+ C
         |               |
   H = 50 kN             |   Frame Dimensions:
   --->  |               |   Span L = 8.0 m, Height h = 4.0 m
         |               |   Columns: HE 240 B (Mp_col = 320 kNm, Ix = 112.6 x 10^6 mm^4)
       A ///           D /// Rafter:   IPE 360  (Mp_beam = 410 kNm)
                             Steel Modulus E = 210,000 MPa (N/mm^2)

5.1 Frame Geometry, Loading, and First-Order Limit Load

• Column plastic moment capacity: $M_{pc} = 320\text{ kN}\cdot\text{ m}$.

• Beam plastic moment capacity: $M_{pb} = 410\text{ kN}\cdot\text{ m}$.

• Since $M_{pc} < M_{pb}$, plastic hinges form in the columns at $A$, $B$, $C$, and $D$ under a pure sway mechanism.

First-Order Plastic Collapse Load ($H_p$):
Using virtual work for sway mechanism with hinges at 4 column ends:

$$H_p \cdot h \cdot \theta = 4 M_{pc} \cdot \theta \implies H_p = \frac{4 M_{pc}}{h} = \frac{4 \times 320}{4.0} = 320.0\text{ kN}$$

The first-order plastic collapse load factor under reference lateral load $H_0 = 50\text{ kN}$ is:

$$\lambda_p = \frac{H_p}{H_0} = \frac{320.0}{50.0} = 6.40$$

5.2 Elastic Critical Buckling Load Factor Calculation

For an unbraced rectangular portal frame with stiff girder ($I_{beam} \gg I_{col}$), the elastic critical sway buckling load $V_{cr}$ is governed by column sway stiffness:

$$V_{cr} = 2 \times \frac{\pi^2 E I_x}{(K h)^2}$$

For fixed-base columns with rigid top joints in sway, the effective length factor is $K \approx 1.2$:

$$K h = 1.2 \times 4.0\text{ m} = 4.80\text{ m} = 4800\text{ mm}$$

$$P_{cr,col} = \frac{\pi^2 \times (210,000\text{ N/mm}^2) \times (112.6 \times 10^6\text{ mm}^4)}{(4800\text{ mm})^2} = \frac{2.3338 \times 10^{14}}{23.04 \times 10^6} = 10,129,340\text{ N} \approx 10,129\text{ kN}$$

Total elastic critical frame gravity capacity:

$$V_{cr} = 2 \times 10,129\text{ kN} = 20,258\text{ kN}$$

The elastic critical load factor under design vertical load $V_0 = 400\text{ kN}$ is:

$$\lambda_{cr} = \frac{V_{cr}}{V_0} = \frac{20,258}{400.0} = 50.645$$

5.3 Merchant-Rankine Second-Order Collapse Factor Evaluation

1. Check Stability Ratio:
   $$\frac{\lambda_{cr}}{\lambda_p} = \frac{50.645}{6.40} = 7.913$$
   Since $4 \le \frac{\lambda_{cr}}{\lambda_p} = 7.913 < 10$, the frame falls into the intermediate stability interaction zone where second order plastic effects must be included.

2. Calculate Failure Load Factor via Modified Merchant-Rankine:
   $$\lambda_f = \frac{\lambda_p}{0.9 + \frac{\lambda_p}{\lambda_{cr}}} = \frac{6.40}{0.9 + \frac{6.40}{50.645}} = \frac{6.40}{0.9 + 0.12637} = \frac{6.40}{1.02637} = 6.235$$

3. Comparison with Classical Merchant-Rankine:
   $$\lambda_{f,classical} = \frac{\lambda_p}{1 + \frac{\lambda_p}{\lambda_{cr}}} = \frac{6.40}{1 + 0.12637} = \frac{6.40}{1.12637} = 5.682$$

The true failure load is reduced from the first-order capacity $\lambda_p = 6.40$ to $\lambda_f = 6.235$ (a $2.6\%$ reduction using Modified Merchant-Rankine, or $11.2\%$ using Classical Merchant-Rankine).

5.4 Post-Yield Deflection and Stability Limit State Verification

The post-yield softening slope is:

$$K_{post} = -\frac{V}{h} = -\frac{400\text{ kN}}{4.0\text{ m}} = -100.0\text{ kN/m} = -0.10\text{ kN/mm}$$

If lateral sway exceeds $\Delta_{limit} = 50\text{ mm}$, the lateral load resistance drops by:

$$\Delta H = K_{post} \cdot \Delta = (-0.10\text{ kN/mm}) \times 50\text{ mm} = -5.0\text{ kN}$$

This drop demonstrates why drift limits must be enforced during p-delta plastic analysis to avoid accelerated collapse.

6. Engineering Failure Modes and Forensic Stability Investigations

Forensic investigations of historical steel frame collapses during extreme wind and seismic events highlight critical vulnerabilities related to second-order effects:

1. Progressive Plastic Sway Softening: In multistory buildings with unbraced lower levels (soft stories), sequential hinge formation in columns triggers rapid $P$-$\Delta$ overturning that outpaces material ductility.
2. Amplified Connection Rotations: Secondary $P$-$\Delta$ drifts demand extreme rotational capacity from beam-to-column moment connections, frequently exceeding bolt or weld shear limits.
3. P-delta Local Flange Plastic Buckling: Compressive stresses amplified by member curvature induce localized plastic flange buckling, precipitating hinge softening.

7. Structural Code Standards: AISC 360 and Eurocode 3 Provisions

Modern structural codes establish clear criteria for assessing when p-delta plastic analysis is mandatory:

7.1 AISC 360 Direct Analysis Method (DAM)

AISC 360 (Chapter C) requires that second-order effects ($P$-$\Delta$ and $P$-$\delta$) be included in all frame designs. The Direct Analysis Method incorporates:
1. Geometric Imperfections: Direct modeling of initial out-of-plumbness ($\Delta_0 = L / 500$) using explicit geometry or equivalent lateral notional loads $Y_i = 0.002 \sum P_i$.
2. Stiffness Reductions: Application of reduced flexural stiffness $E I^* = 0.8 \tau_b E I$ and axial stiffness $E A^* = 0.8 E A$ to account for residual stresses and early plasticity.

7.2 Eurocode 3 (EN 1993-1-1) Stability Criteria

Eurocode 3 uses the critical load ratio $\alpha_{cr} = F_{cr} / F_{Ed}$ to determine whether second-order effects can be neglected:

• For elastic analysis: Second-order analysis required if $\alpha_{cr} < 10$.

• For plastic analysis: Second-order analysis required if $\alpha_{cr} < 15$.

When $\alpha_{cr} \ge 15$, first-order plastic limit analysis is legally sufficient.

8. Synthesis and Engineering Wrap-Up

Mastering p-delta plastic analysis bridges the gap between theoretical first-order limit capacity and real-world frame stability. By accounting for the destabilizing overturning moments induced by vertical gravity acting through lateral sway displacements, structural engineers design frames that avoid dangerous post-yield softening.

Balancing first-order strength with second-order stiffness guarantees robust, collapse-resistant frameworks.

References & Standards Cited

1. American Institute of Steel Construction (AISC). (2022). Specification for Structural Steel Buildings (AISC 360-22). Chicago, IL: AISC.
2. European Committee for Standardization (CEN). (2005). Eurocode 3: Design of steel structures – Part 1-1: General rules and rules for buildings (EN 1993-1-1). Brussels: CEN.
3. Merchant, W. (1954). “The Failure Load of Rigidly Jointed Frameworks as Influenced by Stability.” The Structural Engineer, 32(7), 185–190.
4. Horne, M. R., & Merchant, W. (1965). The Stability of Frames. Oxford: Pergamon Press.
5. Wood, R. H. (1958). “The Stability of Tall Buildings.” Proceedings of the Institution of Civil Engineers, 11(1), 69–102.

Frequently Asked Questions (FAQ)

What is the fundamental difference between P-Delta and P-delta effects?

$P$-$Delta$ (large Delta) refers to the global overturning effect caused by gravity loads acting through inter-story lateral sway displacements. $P$-$delta$ (small delta) refers to the localized secondary bending caused by axial loads acting along the curved chord deflection of an individual member.

When does Eurocode 3 require second-order plastic analysis?

Eurocode 3 requires second-order plastic analysis whenever the elastic critical load ratio $alpha_{cr} = F_{cr}/F_{Ed} < 15$ for plastic limit states (or $alpha_{cr} < 10$ for elastic limit states).

How does the Merchant-Rankine formula approximate second-order plastic capacity?

The **merchant-rankine formula** combines the first-order rigid-plastic collapse factor $lambda_p$ with the elastic sway buckling factor $lambda_{cr}$ via the harmonic relation $1/lambda_f = 1/lambda_p + 1/lambda_{cr}$, capturing the interaction between yielding and instability.

Why does an unbraced frame develop negative post-yield stiffness?

After plastic hinges form to create a mechanism, hinge capacities remain fixed at $M_p$. As lateral sway $Delta$ increases, the secondary moment $P Delta$ increases, forcing the external horizontal resistance $H$ to decrease to maintain equilibrium.

How do AISC notional loads account for geometric imperfections?

AISC notional loads ($N_i = 0.002 Y_i$) apply a fictitious horizontal load equal to $0.2%$ of the factored gravity load at each floor level, simulating a code-mandated initial out-of-plumbness of $Delta_0 / L = 1/500$.