# How do you calculate friction losses in post-tensioned concrete tendons?

aistructuralreview.com · August 25, 2026

> Friction loss calculation is one of the most consequential steps in the design and stressing of post-tensioned concrete structures, because it...

Friction loss calculation is one of the most consequential steps in the design and stressing of post-tensioned concrete structures, because it determines how much of the jacking force actually reaches the far end of each tendon. When a tendon is stressed from one end, two distinct mechanisms rob it of force along its length: curvature friction, caused by the tendon pressing against the duct as it changes direction, and wobble (or wave) friction, caused by unintended minor deviations of the duct from its theoretical profile during concreting. Getting these numbers wrong means either overstressing the anchorage end to compensate for phantom losses or delivering less prestress than the design assumed, which can show up years later as unexpected deflections, cracking, or insufficient load capacity in slabs and transfer girders.

## The Governing Equation: What You Are Actually Computing

**Also worth reading:** [How do you calculate development length for reinforced concrete bars with a worked example?](https://aistructuralreview.com/knowledge/how_do_you_calculate_development_length_for_reinforced_concrete_bars_with_a_worked_example.php) · [how to calculate concrete mix ratio for structural beams?](https://aistructuralreview.com/knowledge/how_to_calculate_concrete_mix_ratio_for_structural_beams.php) · [What are the best practices for post tension tendon layout design in concrete structures?](https://aistructuralreview.com/knowledge/what_are_the_best_practices_for_post_tension_tendon_layout_design_in_concrete_structures.php)

The standard formula used worldwide for post tension friction loss calculation comes from the classical theory of belt friction around a curved surface, adapted for prestressing tendons. The effective stress at any point along the tendon is expressed as:

Pe = Pj × e^-(μa + kx)

where Pe is the effective force at distance x from the jacking end, Pj is the jacking force, μ is the curvature friction coefficient, α is the accumulated angular change of the tendon profile in radians up to that point, k is the wobble coefficient (units of per unit length), and x is the distance along the tendon. The exponential form matters: losses are not linear with length. A tendon with a total angular change of 0.5 radians and μ = 0.20 loses roughly 10 percent to curvature alone before wobble is even counted.

Typical published values give you a starting point but never replace project-specific data. For seven-wire strand in galvanized metal sheathing, μ commonly ranges from about 0.15 to 0.25, while k sits near 0.0015 to 0.0020 per meter (roughly 0.0005 per foot). Plastic (HDPE) ducts generally reduce μ toward 0.12 to 0.18 because the surface is smoother and less prone to rust inside the duct. ACI 318 requires that friction losses be verified by measuring tendon elongation during stressing, and the Post-Tensioning Institute's recommendations treat the calculated elongation versus measured elongation comparison as the acceptance criterion — typically within ±7 percent for bonded systems before investigation is triggered.

## Curvature Friction Versus Wobble Friction: Two Different Animals

Curvature friction is deterministic and intentional. Every time the designer drapes a tendon to follow the bending moment diagram — up over supports, down at midspan — the tendon bears against the duct with a normal force proportional to the local tension and the local curvature. Multiply that normal force by μ and you get the resisting friction force. This is why heavily draped tendons in transfer beams, where profiles may change direction by 0.8 radians or more, routinely lose 25 to 40 percent of jacking force by the dead end, while nearly flat banded tendons in a two-way slab might lose only 4 to 8 percent.

Wobble friction, by contrast, is unintentional and statistical. No crew places ductwork perfectly; supports sag, chairs shift under foot traffic, concrete pressure displaces the duct slightly, and the tendon inside never sits exactly on the duct centerline. These micro-deviations accumulate, so wobble loss grows linearly with length regardless of profile geometry. In short tendons under about 30 meters, wobble is often the dominant term. In long tendons exceeding 60 meters, both terms matter and the exponential compounding makes single-end stressing impractical, which is why codes and practice push designers toward two-end stressing beyond certain lengths.

| Feature | Curvature Friction | Wobble (Wave) Friction |
| --- | --- | --- |
| Physical cause | Intentional profile curvature bearing on duct | Unintended duct deviations and misalignment |
| Coefficient symbol | μ (dimensionless) | k (per unit length) |
| Typical value, strand in metal duct | 0.15–0.25 | 0.0015–0.0020 /m |
| Growth pattern | Exponential with accumulated angle | Linear with tendon length |
| Dominant in | Heavily draped beams, transfer girders | Short flat tendons, slabs |
| Designer control | High — adjust profile | Low — depends on workmanship |
| Verification lever | Elongation measurements | Duct placement tolerance inspection |

## Practical Step-by-Step Calculation Procedure
A defensible calculation follows a repeatable sequence. First, establish the jacking force, usually 75 to 80 percent of the specified ultimate tensile strength of the strand (ACI 318 caps jacking stress at 0.80 fpu, with typical practice at 0.75 fpu = approximately 1,395 MPa for Grade 1860 strand). Second, break the tendon into segments between points where the slope changes, and compute the angular change per segment from the drape geometry: for a parabolic segment, α ≈ 4e/L, where e is the mid-segment drape relative to the chord and L is the segment length. Third, apply the exponential formula cumulatively, segment by segment, carrying the reduced force forward as the input to the next segment rather than computing everything off the full jacking force — this is a common spreadsheet error that understates losses.

Fourth, compute predicted elongation using the average force method or, more accurately, by integrating numerically: Δ = ∫(P/AEs)dx over the free stressing length. Fifth, compare against field-measured elongation. Sixth, back-calculate implied coefficients if the discrepancy exceeds tolerance. Modern software — including AI-assisted structural platforms that auto-extract tendon geometry from BIM models — performs this segmentation automatically, but the engineer still owns the assumptions for μ and k, and those assumptions deserve scrutiny every time the duct supplier, duct material, or placing crew changes.

## Stressing End Selection and Two-End Stressing

Because friction always reduces force away from the jack, the choice of stressing end is itself a design decision. The convention is to stress from the end where the resulting friction-reduced force distribution best matches the demand diagram. For a simply supported beam with symmetric draping, either end works identically. For an asymmetric member, or one where the critical section sits closer to one support, stressing from the nearer end puts more force where it is needed. Many engineers run the calculation from both ends and pick the orientation that maximizes minimum effective prestress across all critical sections.

When computed losses exceed roughly 15 to 20 percent, or when tendon length exceeds about 45 to 60 meters depending on profile complexity, two-end stressing becomes the economical answer. Stressing simultaneously or sequentially from both ends roughly halves the maximum distance over which friction acts, cutting peak losses substantially — though it doubles jacking operations, adds anchorage hardware at the second end, and complicates elongation bookkeeping since each end records only part of the total extension. The crossover point is economic as much as technical: for a 40-meter tendon losing 12 percent, adding a second live anchorage may cost more than the extra strand needed to compensate for losses via higher jacking force within code limits.

## Common Mistakes That Corrupt the Numbers

The most frequent error is treating μ and k as universal constants copied from a textbook instead of values calibrated to the actual duct system and validated against first-stressing results. A second widespread mistake is ignoring the seating (anchorage) loss at the wedge, typically 6 mm of slip for standard wedges, which disproportionately affects short tendons — on a 10-meter tendon, 6 mm of slip can erase several percent of the intended elongation and must be added to the required elongation before stressing. Third, many calculations use the chord angle between endpoints rather than summing true segment angles on sharply curved profiles, understating α wherever the tendon curves through intermediate low points.

Field-side errors compound the analytical ones. Elongation should be measured after the initial take-up seating that removes slack, not from zero ram travel. Ram calibration drift skews the force-elongation relationship silently. And crews sometimes chase matching elongation by pumping past the code stress limit, which risks strand rupture at the anchorage and violates the 0.80 fpu ceiling — if elongation will not come in range, the correct response is to stop, investigate duct obstruction or wrong coefficients, and re-run the numbers, not to keep pulling.

## Verification, Tolerances, and When to Recalculate

Acceptance hinges on the elongation check. Most specifications require measured elongation within ±7 percent of calculated for each tendon, with some owners tightening to ±5 percent on critical members. If a tendon falls outside tolerance, the standard protocol is to measure additional tendons in the same group to distinguish a systematic problem (wrong coefficients, blocked duct) from an isolated measurement issue. Persistent short elongation suggests internal obstruction or higher-than-assumed friction; persistent long elongation suggests lower friction than assumed, which sounds benign but means the real force distribution differs from design and may need documenting.

Recalculation is warranted whenever conditions change materially: switching from rigid metal to semi-rigid plastic duct, changing the placing contractor, revising the tendon profile during coordination with openings and embeds, or extending tendons through revised member depths. On projects using AI-driven structural review tools, automated cross-checks now flag tendons whose predicted loss exceeds thresholds or whose field elongation trends deviate across a pour, catching systematic errors that used to surface only during special inspection disputes. Whatever the tooling, the underlying physics has not changed since the belt-friction derivation, and no software output should be accepted without a hand check on at least one representative tendon.

## Cost and Consequence Perspective

Direct costs of doing friction analysis properly are trivial: engineering hours plus modest software licensing, often under a few thousand dollars even on large jobs. The consequences of skipping or botching it scale far higher. Understated losses mean undersized prestress, which may force remedial external post-tensioning or strengthening costing tens of thousands per member, plus schedule disruption. Overstated losses lead to oversized tendons and wasted strand, a smaller but real penalty repeated across hundreds of tendons in a typical parking structure. Wedge failure from overstressing carries safety consequences that dwarf any accounting. Treat the friction calculation, the elongation verification, and the coefficient validation loop as inseparable parts of one quality process rather than paperwork tasks.

## Quick answers

### What are typical friction coefficient values for post-tensioning tendons?

For seven-wire strand in galvanized metal ducting, μ typically ranges from 0.15 to 0.25 and k from 0.0015 to 0.0020 per meter. Smooth HDPE plastic ducts usually reduce μ to roughly 0.12–0.18. Always verify against the duct supplier's tested values and calibrate with early elongation measurements.

### At what tendon length does two-end stressing become necessary?

Practice generally moves to two-end stressing when losses exceed about 15–20 percent or tendon length exceeds roughly 45–60 meters, depending on profile curvature. Heavily draped transfer beams may justify two-end stressing at shorter lengths than flat slab tendons.

### Why is my measured elongation shorter than calculated?

Short elongation usually indicates higher-than-assumed friction, a partially blocked duct, unaccounted wedge seating slip, or incorrect segment angles in the calculation. First confirm measurement procedure and ram calibration, then check neighboring tendons to determine whether the issue is systematic before considering re-stressing or duct flushing.

### How much force do post-tensioned tendons typically lose to friction?

Flat banded slab tendons often lose only 4–8 percent of jacking force, while heavily draped beam tendons can lose 25–40 percent by the dead end. Losses grow exponentially with accumulated angular change, so profile geometry dominates the outcome.

### Does wedge seating loss count as friction loss?

No, it is accounted separately. Standard wedges slip about 6 mm when the jack releases, and this seating loss must be added to the required elongation before stressing. Its effect is proportionally largest on short tendons, where 6 mm can represent a meaningful percentage of total extension.

Canonical: https://aistructuralreview.com/knowledge/how_do_you_calculate_friction_losses_in_post-tensioned_concrete_tendons.php
Markdown: https://aistructuralreview.com/knowledge/how_do_you_calculate_friction_losses_in_post-tensioned_concrete_tendons.php/index.md