Broms method vs p-y method for laterally loaded piles

The Broms method (1964) is a hand calculation that gives the ultimate lateral capacity of a single pile in uniform clay or sand, treating the pile as rigid or yielding at a plastic hinge. The p-y method solves the full nonlinear soil–structure problem, producing deflection, moment, and shear at every depth in layered soils at any load level. Use Broms for preliminary sizing and sanity checks; use p-y for design.

What the Broms method actually is

Bengt Broms published the method in two companion papers — one for cohesive soils, one for cohesionless soils (Broms 1964a, b, ASCE Journal of the Soil Mechanics and Foundations Division). It is a limit-equilibrium calculation: assume the distribution of soil pressure against the pile at failure, then solve statics for the horizontal load that exhausts either the soil or the pile section. It does not model stiffness at all — only strength.

The two soil idealizations are deliberately simple:

• Purely cohesive soil (undrained clay). The ultimate soil resistance is taken as zero from the surface down to a depth of 1.5 pile diameters (the "dead zone," where the soil can heave upward), and a constant

  pᵤ = 9 · cᵤ · b

  below it, where cᵤ is the undrained shear strength and b the pile diameter. The factor 9 is the classical flow-around bearing factor — the same limit the p-y curve families approach at depth.
• Purely cohesionless soil (sand). The ultimate resistance grows linearly with depth,

  pᵤ = 3 · γ′ · z · Kₚ · b,   Kₚ = tan²(45° + φ/2)

  i.e. three times the Rankine passive pressure acting over the pile width, with γ′ the effective unit weight and φ the friction angle.

Broms worked the statics out for free-head and fixed-head (restrained) piles and condensed the results into dimensionless design charts: for short piles, ultimate load Hᵤ/(cᵤ·b²) or Hᵤ/(Kₚ·γ′·b³) versus embedment L/b; for long piles, ultimate load versus the dimensionless yield moment of the section. You read the chart, multiply back out, and apply a factor of safety — traditionally around 2.5 on the ultimate load.

Short (rigid) piles vs long (flexible) piles

The central idea in Broms is that a laterally loaded pile fails in one of two ways, and you must check which governs:

• Short (rigid) pile — the soil fails first. The pile rotates as a rigid body about a point near the toe (free head) or translates bodily (fixed head), and the capacity is set entirely by embedment and soil strength. For a fixed-head short pile in clay, for example, the statics reduce to Hᵤ = 9·cᵤ·b·(L − 1.5b); for a free-head short pile in sand, moment equilibrium about the toe gives Hᵤ = ½·γ′·b·L³·Kₚ/(e + L), with e the load eccentricity above grade.
• Long (flexible) pile — the pile fails first. The bending moment reaches the yield moment Mᵧ of the section and a plastic hinge forms (one hinge below grade for a free head, hinges at the head and below grade for a fixed head). Capacity is then controlled by Mᵧ, and making the pile longer adds nothing.

Whether a pile is "short" or "long" is judged with a dimensionless length — β·L in clay and η·L in sand, where β = (kₕ·b/4EI)^¼ and η = (nₕ/EI)^⅕ come from subgrade-reaction theory. Roughly, values below about 2 behave rigidly and values above about 4 behave flexibly, with an intermediate regime between. Most driven piles and drilled shafts at typical structural embedments land in the long-pile regime, where the section's yield moment governs.

What Broms gives you — and what it doesn't

Three outputs, all at the failure state or well below it:

• the ultimate horizontal load Hᵤ (or, inverted, the required embedment or Mᵧ);
• the maximum moment and its approximate depth, from the same statics;
• a rough head deflection at working load, from a separate set of charts based on linear subgrade-reaction theory (constant kₕ in clay, nₕ·z in sand). Broms himself framed these as estimates valid only at loads well below ultimate — the soil response is assumed linear, which real soil is not.

What it cannot give you: the load–deflection curve, the moment and shear diagrams along the pile, the response of a layered profile, cyclic degradation, or the effect of a cracked concrete section. Those are exactly what the p-y method was built for.

What the p-y method gives you

The p-y method models the pile as a beam-column on nonlinear springs and solves the governing equation

EI·y⁗ + Pₓ·y″ + Eₚᵧ·y − W = 0

by finite differences, iterating on the secant modulus of a nonlinear p-y curve at every node — the COM624P/LPILE formulation (Wang & Reese 1993, FHWA-SA-91-048). Each depth gets its own curve from a family calibrated against full-scale load tests: Matlock (1970) soft clay, Welch & Reese (1972) and Reese-Cox-Koop (1975) stiff clay, Reese (1974) and API/O'Neill-Murchison (1983) sand, Reese (1997) weak rock. The output is the complete response — deflection, rotation, moment, shear, and soil reaction at every node, at every load step — for arbitrarily layered profiles, static or cyclic loading, free, fixed, or elastically restrained heads, and moment–curvature-based nonlinear EI.

Note the family resemblance: the ultimate values that p-y curves flatten toward are Broms-type limits (Matlock's deep-flow factor reaches 9·cᵤ·b; the sand curves use wedge and flow failure modes closely related to Broms' passive-pressure argument). The p-y method keeps Broms' strength physics and adds the stiffness path to get there.

Broms vs p-y at a glance

When Broms is still the right tool

• Preliminary sizing. With one cᵤ or one φ you can bracket the required diameter, embedment, and section before any detailed analysis exists.
• Hand-checking software. A Broms ultimate load is an independent calculation from first principles. If a p-y run reports a working capacity above the Broms ultimate, something is wrong with the model — soil parameters, units, or head fixity.
• Sparse data. Early-phase sites often have nothing more than an SPT log and an index strength. Broms asks for exactly that much.
• Ultimate-strength questions. If the design question is purely "does the pile break or the soil fail first," Broms answers it directly.

Where Broms is inadequate

• Layered profiles. Soft clay over sand, fill over till — the single-soil assumption breaks down exactly where real sites live. Picking one "equivalent" soil is guesswork.
• Serviceability governs. Most lateral designs are controlled by a deflection limit at the pile head or a rotation limit at a bridge seat — not by ultimate capacity. Broms' linear-subgrade deflection estimate is not defensible for that check; the nonlinear load–deflection curve from a p-y analysis is.
• Cyclic loading. Wave, wind, seismic, and traffic loading degrade soil resistance near the surface. The cyclic p-y curves model this; Broms does not address it.
• Nonlinear EI. A reinforced-concrete shaft cracks well before it yields, which redistributes moment and increases deflection. Only an analysis that couples moment–curvature to the solution captures it.

Use both: Broms as the sanity check on a p-y run

The two methods are complementary, not competing. A sensible workflow, expanded in our step-by-step lateral design guide, is: size the pile with Broms, then analyze it with p-y. Afterward, close the loop — push the p-y analysis to large loads and confirm the head load at which the solution stops converging or deflections run away is commensurate with the Broms ultimate; check that the p-y maximum moment at factored load stays below Mᵧ, mirroring Broms' long-pile check. If the two disagree wildly, find out why before stamping anything. That habit — an independent hand check on every machine result — is also why we publish PileCalc's validation cases instead of asking you to trust the engine.

Common questions

Is the Broms method still used?

Yes, but almost exclusively for preliminary sizing and for hand-checking software results. Broms (1964) remains a fast, defensible way to estimate ultimate lateral capacity from a single soil strength. Final design of laterally loaded piles is now done almost universally with the p-y method, which handles layered soils, cyclic loading, and deflection limits that Broms cannot.

What is the main difference between Broms and the p-y method?

Broms is a limit-equilibrium method: it gives the ultimate lateral load at failure for a uniform clay or sand, plus a rough linear deflection estimate. The p-y method is a numerical soil–structure interaction analysis: it computes nonlinear deflection, moment, and shear at every depth and every load level, in layered profiles, for static or cyclic loading.

Can the Broms method handle layered soil profiles?

No. Broms assumes one uniform soil that is either purely cohesive or purely cohesionless, and its charts are derived for that idealization. Real profiles — soft clay over dense sand, fill over till — must be forced into a single equivalent soil, which is judgment, not analysis. Layered profiles are the standard use case for the p-y method.

Does the Broms method give pile deflection?

Only approximately. Broms provided separate charts for head deflection at working load based on linear subgrade-reaction theory — constant kₕ in clay, linearly increasing nₕ·z in sand. These are order-of-magnitude estimates valid well below ultimate load. Where a serviceability deflection limit governs the design, a nonlinear p-y analysis is required.

Want to see the difference on your own pile? Run a free lateral pile deflection calculation in the browser — full p-y analysis with the deflected shape and moment diagram — and compare the result against your Broms hand check.