steel-structure-second-order-analysis
Steel Structure Second-Order Analysis: P-Δ, P-δ & the Direct Analysis Method
Engineering line drawing of a single-bay steel frame, dashed lines for the undeformed position and solid lines for the swayed deformed position under lateral load, column axial load P and story drift Δ labeled, clean white drafting style.
A frame that sways under load leans further—and the leaning columns push the frame to sway still further. That feedback loop is exactly what a first-order analysis ignores and what second order analysis in steel design is built to catch. Skip it and a model that "passes" can quietly under-design every sway column in the building.
Second-order analysis captures how axial load acting on the already-deformed frame creates extra bending moments—the P-Δ and P-δ effects. You can account for them either by multiplying first-order moments with amplification factors (B1 and B2), or by running the modern Direct Analysis Method (DAM), which builds stiffness reduction and imperfections straight into a true second-order run. Get the method right and every column and connection is sized for the moments that actually occur; get it wrong and you have a "passed" model that hides 10–30% of the real demand.
This article explains why first-order is not enough, the two distinct second-order effects, the moment amplification factor approach, the Direct Analysis Method, and how AISC 360 and GB 50017 differ. Member buckling ratios and slenderness checks are covered in our steel structure stability design guide. This one is about how the analysis itself accounts for geometry change under load.
Why First-Order Isn't Enough
A first-order (linear) analysis calculates member forces using the original, undeformed geometry. It asks: given the frame as drawn, what moments and shears do the loads produce? That is fine for stiff, low-rise, stocky frames. A second-order analysis recomputes the forces on the deformed geometry. Once a column drifts sideways or a beam bows, the axial force P acting on that displacement Δ creates an additional moment P×Δ that the linear model never sees. Iterate that feedback and the moments grow.
How much it matters depends on how flexible and heavily loaded the frame is. Tall and narrow frames (large height-to-width ratio), columns carrying high axial load, flexible bracing systems, and any sway-sensitive building all develop meaningful second-order moments. Even a simple portal frame under wind: the top of the column drifts, and that drift magnifies the base moment that the wind load alone produced. How far that drift may go—and whether the frame is braced or free to sway—sets the whole stability strategy upstream; see our overall frame stability guide on allowable story drift angles and the braced-versus-unbraced K-factor that precedes this second-order check.
The output of second-order analysis—the amplified member forces—is then fed into the member stability checks. In other words, second order analysis steel structure work does not replace stability design; it supplies the forces that stability design checks. For the member-level side of that handshake, see steel structure stability design. In a multi-story steel building, where drift accumulates over many floors, the effect compounds with every story.
| Frame Characteristic | First-Order Adequate? | Second-Order Priority |
|---|---|---|
| Low-rise, stiff braced warehouse | Usually yes | Low |
| Portal frame under moderate wind | Often borderline | Medium |
| Tall / narrow moment frame | No | High |
| Heavily loaded, slender columns | No | High |
| Flexible bracing / large drift | No | Very high |
Typical engineering judgment; the governing code sets the threshold. Confirm with your structural engineer.
Two Effects: P-Δ and P-δ
Second-order analysis actually names two separate effects, and confusing them is a classic mistake.
P-Δ (the frame-sway effect). When the whole frame drifts sideways by Δ at a floor level, every column's axial load P now acts at a lever arm Δ from its original line. That creates an extra overturning moment at the floor—P times Δ. P-Δ acts at the story level, amplifying total frame moments and increasing drift. It is the effect a tall, sway frame lives under.
P-δ (the member-curvature effect). Within a single beam or column, the member itself bends by δ along its length. The axial load P on that bowed member then acts at the local offset δ, adding extra moment at the member's own ends and midspan. P-δ acts inside one member, amplifying that member's moments even if the frame never sways.
The two are independent and both must be checked. P-Δ is driven by story drift; P-δ is driven by individual member deflection. Modern software can output them separately, but many engineers check one and forget the other. A column in a sway frame needs both: P-δ magnifies its end moments from within, and P-Δ magnifies them again from the whole story drifting. Checking both together is what separates a credible second-order analysis steel structure model from a guess.
| Effect | Cause | Acts On | Typical Size | Notes |
|---|---|---|---|---|
| P-Δ | Story drift Δ | Whole floor / frame | Grows with height and axial load | The "sway" effect |
| P-δ | Member bow δ | Single beam / column | Grows with member slenderness | The "curvature" effect |
| P-Δ + P-δ | Combined | Column in sway frame | Can add 10–30% to moments | Both must be included |
Illustrative relative effects; magnitudes depend on geometry, axial load, and stiffness. Never skip one while checking the other.
Moment Amplification Factor Method
The traditional way to get second-order forces without a full nonlinear model is the moment amplification factor method. You first run a first-order analysis to get the moments, then multiply them by factors that estimate the extra moment the deformed geometry would create.
The working relation is:
M (amplified) = B1 × M (non-sway) + B2 × M (sway)
- B1 amplifies the P-δ effect on a member that does not sway. It depends on the ratio of the member's axial load to its own Euler buckling load.
- B2 amplifies the P-Δ effect on a sway story. It depends on the ratio of the story's total axial load to the story's lateral buckling load.
Read the factors as a thermometer. A factor near 1.0 means second-order effects are small and safely ignored. A factor climbing toward 1.1 or above means the extra moments are real and must be carried into the member and connection design. The method is fast and clear for regular, orderly frames; for irregular geometry or large deformations, a direct second-order analysis is the better tool. Lateral stiffness from bracing is what keeps B2 small—see our steel building bracing system guide.
| Factor | Source Effect | Based On | When It Approaches 1.0 |
|---|---|---|---|
| B1 | P-δ (member curvature) | Member axial load / member Euler load | Short, stocky, lightly loaded members |
| B2 | P-Δ (frame sway) | Story total load / story lateral stiffness | Stiff, braced, low-rise stories |
| B1 and B2 > 1.1 | Combined | Combined ratios | Second-order effects significant |
Factor formulas are given in AISC 360 (Appendix 8) and GB 50017; exact values require the code tables.
Direct Analysis Method (DAM)
The Direct Analysis Method is the modern AISC 360 approach. Instead of first computing moments by linear analysis and then bolting on amplification factors, DAM runs a true second-order (P-Δ-δ) analysis that already accounts for imperfections and stiffness loss in one step. The payoff is directness: you no longer look up effective-length (K) factors from alignment charts—you model the imperfections and let the analysis find the buckling behavior.
DAM has three ingredients:
- Stiffness reduction. The member stiffness is reduced to reflect residual stresses and the onset of buckling, so the analysis does not overestimate how stiff the frame really is.
- Notional loads. Imaginary horizontal loads applied to the frame simulate the geometric out-of-plumb and initial imperfections a real frame always has—equivalent to starting the P-Δ feedback from day one.
- A real second-order analysis. The program then solves the equilibrium on the deformed geometry, capturing both P-Δ and P-δ in one run.
Because DAM does not split frames into "sway" and "non-sway" and does not rely on K ≥ 1.0 lookups, it suits complex, irregular, or sway-sensitive buildings, and it is built into mainstream packages (SAP, ETABS, MIDAS). A simple regular bent can still use the B1/B2 amplification method. That directness is why DAM has become the default route for any modern second-order analysis steel structure project.
| Aspect | Amplification Method (B1/B2) | Direct Analysis Method (DAM) |
|---|---|---|
| Core idea | First-order moments × factors | True second-order run |
| Effective-length (K) factor | Looked up from charts | Not required |
| Imperfections | Implied by K | Explicit notional loads |
| Best for | Regular, simple frames | Complex / irregular / sway-sensitive |
| Workflow | Two-step | One integrated step |
DAM requirements are in AISC 360 Chapter C; for the specification, see AISC 360.
Is Your First-Order Model Underestimating Moments?
In a tall or sway-sensitive frame, the P-Δ and P-δ effects can add 10–30% to moments a linear model never shows. Our engineers run second-order (DAM) analysis to AISC 360 / GB50017 and hand you the amplified member forces. Ask for a sample analysis report.
Codes: AISC 360 vs GB 50017
Both major codes require second-order effects where they matter, but they phrase the route differently.
AISC 360 (United States) offers two accepted paths: the moment amplification method (B1/B2) for simpler frames, or the Direct Analysis Method as the modern, recommended approach that drops the K-factor lookup. It is the standard most international and export projects are designed to.
GB 50017 (China) uses either a second-order elastic analysis or a first-order analysis with equivalent notional horizontal loads that simulate initial imperfections. The philosophy is close to DAM: inject the imperfection, then let the analysis capture the added moments. The Chinese steel standard explicitly requires second-order analysis for high-flexibility structures.
Which to follow is set by the project's governing code. International work commonly runs to AISC or Eurocode 3 (EN 1993-1-1); domestic export to a Chinese owner typically runs to GB 50017. Confirm the code of record with your local structural engineer before drawings are issued.
| Code | Accepted Method | Imperfection Handling | Notes |
|---|---|---|---|
| AISC 360 | B1/B2 amplification or DAM | Notional loads (DAM) | Modern preferred: DAM |
| GB 50017 | Second-order elastic, or first-order + notional loads | Notional horizontal loads | High-flexibility must use second-order |
| Eurocode 3 (EN 1993-1-1) | Second-order global analysis | Equivalent imperfections | International projects |
Method statements; exact clauses and factors per the current code edition your project cites.
When It Matters Most
Second-order analysis bites hardest on the structures people tend to design with a quick linear model. Tall, narrow frames with high height-to-width ratios, columns carrying large axial load, and flexible bracing all hide meaningful P-Δ. Cantilever and tower structures are the extreme case. Underestimating these moments does not usually cause a dramatic collapse—it quietly overstresses a column base plate or a bolted connection, which is exactly why a steel structure connection design review must use amplified forces, not first-order ones. Feeding those amplified forces downstream is the whole payoff of a rigorous second-order analysis steel structure run.
The recurring mistakes are running a straight linear analysis straight to drawings, forgetting P-δ while handling P-Δ, and skipping the stiffness reduction that DAM requires. A disciplined steel structure drawing review catches the "analysis type" note on the drawings—if it says linear first-order on a tall frame, stop and ask.
Conclusion
Second order analysis steel structure work is the discipline of capturing the extra moments that geometry change under load creates: P-Δ from whole-frame sway and P-δ from individual member bow. You capture them either by multiplying first-order moments with B1/B2 amplification factors or by running the Direct Analysis Method, which folds stiffness reduction and notional loads into one true second-order run. A first-order model that "passes" is not evidence the real moments are small—on a tall, narrow, or heavily loaded frame, the second-order run can raise column moments by 10–30%. Design to the amplified forces under AISC 360 or GB 50017, and hand those forces to the connection and member checks.
Know the Real Moments Before You Fabricate.
We run second-order (P-Δ/P-δ, Direct Analysis Method) structural analysis to AISC 360 / GB50017 so every column and connection is sized for the amplified forces a linear model hides. Ask for a sample second-order analysis report for your frame.
🏭 Explore our products: Steel Factory · Steel Workshop · Steel Warehouse
Case Example
A six-story, 18,000 m² (≈194,000 sq ft) office in a moderate-seismic zone was first modeled as a first-order sway frame. Every column passed at unity ratio, but the reviewers flagged that P-Δ amplification had been estimated with a conservative B2 factor that risked both over-design and, if wrong, under-design.
Key challenges: rigid moment frames in both directions, a weakly restrained roof story, and a client unwilling to over-specify steel tonnage.
Solution: the team re-ran the model using the AISC 360 Direct Analysis Method—introducing stiffness reduction τ_b, notional lateral loads, and a true second-order run—instead of multiplying first-order moments. Effective length factors were no longer guessed: K moved from the assumed 1.0 on several top-story columns to about 1.6 once real joint restraint was modeled.
Results: twelve columns were upsized, the rest stayed, and total steel tonnage rose only about 6% versus the first-order "pass." Compared with a blanket B2 conservative run, the DAM approach cut steel weight by roughly 18% and gave the connection designer moments that actually match the deformed geometry. See overall stability design and member local stability for the companion member checks.
Reference Links
- AISC 360 Specification for Structural Steel Buildings
- ASCE 7 Minimum Design Loads and Associated Criteria for Buildings and Other Structures
- ISO 12944 Corrosion protection of steel structures by protective paint systems
- GB 50017 Standard for Design of Steel Structures (China)
- Eurocode 3 (EN 1993) Design of Steel Structures
About the Author
Senior Structural Engineer
With over 20 years of hands-on experience in steel structure design and prefabricated building engineering, our in-house senior structural engineer has personally contributed to more than 500 steel building projects—including warehouses, industrial factories, aircraft hangars, agricultural buildings, and commercial structures. The focus is on translating design codes such as AISC 360, ASCE 7, and Eurocode 3 into buildable, cost-effective steel solutions that balance structural performance, fabrication efficiency, and total project cost.
Learn more about our engineering team
Frequently Asked Questions
What is second-order analysis in steel design?
Unlike first-order (linear) analysis, which calculates forces on the undeformed frame, second-order analysis accounts for the deformed geometry. Once the frame sways or a member bends, axial load acting on that displacement creates extra bending moment—the feedback that linear design ignores. It is required where frames are tall, slender, or heavily loaded.
What is the difference between P-Δ and P-δ?
P-Δ is the frame-sway effect: total column axial load acting on the story drift Δ adds overturning moment at the floor level. P-δ is the member-curvature effect: axial load on a single beam or column's own deflection δ adds moment within that member. They are separate effects and both must be checked.
How does the moment amplification method work?
You first run a first-order analysis to get moments, then multiply them by factors—B1 for P-δ (member curvature) and B2 for P-Δ (frame sway), each derived from the ratio of axial load to the relevant buckling load. A factor near 1.0 means second-order effects are small; above about 1.1 they cannot be ignored.
What is the Direct Analysis Method (DAM)?
The Direct Analysis Method (AISC 360 Chapter C) runs a true second-order analysis that already reduces member stiffness and applies notional (imperfection) loads, so it directly outputs the amplified member forces—removing the need to look up effective-length (K) factors. It is the modern preferred method for irregular or sway-sensitive frames.
Do AISC 360 and GB 50017 handle second-order analysis differently?
Both require second-order effects where they matter. AISC 360 offers either the B1/B2 amplification method or the Direct Analysis Method; GB 50017 uses second-order elastic analysis or a first-order analysis with equivalent notional horizontal loads, philosophically similar to DAM. Confirm the governing code with your local structural engineer.
steel-structure-fire-resistance-design
steel-structure-deflection-control