Structural Bolting Tensioning Methods: Torque, Turn-of-Nut and Tensioning Compared
28 min·Yaxiio Engineering Alliance

Structural Bolting Tensioning Methods: Torque, Turn-of-Nut and Tensioning Compared

Compare torque, turn-of-nut, twist-off and hydraulic tensioning: where preload scatter comes from, what to inspect, and how to choose.

An M24 class 10.9 assembly needs about 247 kN of clamp force. A torque wrench set to 771 N·m delivers between 214 and 292 kN across the batch coefficient window of 0.11 to 0.15. Turn-of-nut reaches the same preload by turning the nut about 45° from snug. A twist-off bolt stops when its splined end shears. A tensioner pulls the bolt directly, then hands the load to a nut.

The method in the work instruction decides which of those numbers can be defended at handover. This entry compares the four methods on scatter, evidence and failure mode.

  • Routine steel frame, tested batch: torque method with batch K and a sampled check.
  • Bridge, tower or fatigue-critical joint: turn-of-nut or twist-off.
  • Equipment base or large stud: hydraulic tensioning with pressure and elongation.

The Four Methods in One View

All four finish with a nut seated against a closed joint. They differ in what is measured on the way there.

Method Preload set by Main scatter source Evidence Best fit
Torque Torque through K Batch K, finish, lubricant Coefficient certificate, calibrated wrench, sampled check Routine steel connections
Turn-of-nut Rotation from snug Snug point, joint stiffness Angle mark, sequence record, witness Bridges, towers, tight band
Twist-off Shear of splined end Batch calibration, groove tolerance Sheared end on every bolt Repetitive site work
Tensioning Direct axial load Load transfer, gauge, thread friction Pressure log, elongation after transfer Equipment bases, large studs

Torque and twist-off control a torque; turn-of-nut a rotation; tensioning a force. None measures retained preload at handover.

Torque Method: A Friction Estimate

Formula
T=K · F · d

T is torque (N·m), F preload (kN), d nominal diameter (mm). GB/T 1231-2024 §7.4.1 defines Ki = T/(F·d) from the coefficient test. For large hexagon assemblies, §6.3.1 requires batch mean K 0.110–0.150 and Vk ≤ 0.077. GB/T 32076.10-2018 Table 2 accepts 0.10 < k ≤ 0.23 with Vk ≤ 0.10 for HR and HV assemblies. A K value copied from one family does not transfer.

M24 class 10.9: As = 353 mm² (GB/T 1231-2024 Table 10), Rm = 1000 MPa (GB/T 3098.1-2010 §7.1), so Fp = 0.7·Rm·As = 247.1 kN. Table 14 gives 237–263 kN for 10.9S M24, and 247.1 sits inside. At K = 0.13 the torque is 771 N·m; at 0.11 and 0.15 it is 652 and 890 N·m, a 36% spread. At a fixed 771 N·m the delivered preload is 214 kN at K = 0.15 and 292 kN at K = 0.11.

At mean K 0.165 and Vk 0.10, one standard deviation is 0.1485 or 0.1815, and preload moves about +11%/−9% at fixed torque; two standard deviations put preload roughly +25%/−17% from the mean with a calibrated wrench.

Turn-of-Nut: Angle Sets the Preload

After the joint is snug, the nut is rotated through a fixed angle. The rotation advances the nut by p·θ/360, and that displacement is shared between bolt stretch and joint compression:

Formula
F=(θ/360°) · p · k_eq
1/k_eq=1/k_b+1/k_c

p is pitch, k_b = As·E/l_b,eff, and k_c is joint stiffness. GB/T 32076.2-2015 §4 defines l_b,eff as clamp length plus half the nut height.

M20 class 10.9: As = 245 mm², Fp = 171.5 kN. Clamp length 60 mm, nut height 20 mm, so l_b,eff = 70 mm and k_b = 245 × 210,000 / 70 = 735 kN/mm. Assume k_c = 1,470 kN/mm for steel plies with washer faces. Then k_eq = 490 kN/mm, δ = 171.5 / 490 = 0.350 mm, and θ = 0.350 / 2.5 × 360 = 50°. From a 20 kN snug point the angle is about 45°. The stiffness ratio and snug load are example assumptions.

The weak input is the snug point. At 40 kN instead of 20 kN, the same 45° ends 20 kN high, about 12% of target.

Twist-Off Bolts: The Splined End Is the Record

GB/T 3632-2008 defines a bolt with a splined end and a groove, a nut and a washer. The dedicated wrench grips the splined end and turns the nut. At the calibrated torque the splined end shears. The sheared end is the record.

GB/T 32076.10-2018 §8.3.3 sets initial and re-tightening torque at M1 = 0.065·F0·d. For M24 at F0 = 248 kN, M1 = 387 N·m. Final tightening continues until the splined end breaks, with all passes within 24 h. If access prevents the dedicated wrench, torque-method final tightening is allowed with k = 0.13, or 774 N·m for this M24. The remaining end is cut mechanically; oxygen cutting is prohibited.

GB/T 3632-2008 Table 12 accepts a batch mean clamping force of 248 kN for M24, min 225 kN, max 272 kN, standard deviation ≤ 22.5 kN. Eight assemblies are sampled per batch (§7.2). The band is −9%/+10%, and the permitted standard deviation is 9% of the mean.

Twist-off removes the operator’s torque decision, not K: the shear torque is set at production for the batch, and residual scatter is batch calibration plus groove tolerance.

Hydraulic Tensioning: Load Applied, Load Retained

A tensioner pulls the bolt shank, runs the nut down against the joint, then releases pressure and transfers load to the nut:

Formula
F_applied=p · A
F_retained=F_applied − load transfer loss

A 2,000 mm² piston at 123.5 MPa applies 247 kN. The gauge reads applied load. What stays is lower because threads settle and the bolt relaxes as the load path changes. An 8% transfer loss leaves 227 kN, below the 237 kN lower bound of the GB/T 1231-2024 Table 14 window for M24 10.9S. The 8% is an illustrative field value, not a standard limit; measure it on the first bolts.

Elongation is the check that survives transfer. For M24 with l_b,eff = 72 mm, δ = F·l/(As·E) = 247,000 × 72 / (353 × 210,000) = 0.240 mm. A reading of 0.22 mm after transfer means about 226 kN retained.

Boundary: GB/T 32076.10-2018 §7.1 lists torque, torque-angle, twist-off and direct tension indicator methods. Hydraulic tensioning is not on that list. Where used, the procedure comes from the project specification, often a flange-joint guideline; the GB installation standard does not cover it.

Snug Tight Is the Zero Point

Snug tight means every ply is in firm contact, hole clearance is taken up, and the nut has begun to carry load. The joint is closed but not at design preload.

Turn-of-nut counts angle from this point; twist-off uses initial torque to reach it; the torque method assumes final torque is applied to a closed joint. If the joint is still opening, torque closes it instead of stretching the bolt.

Align the holes, remove burrs and paint from faying surfaces, bring plies together with drift pins, and replace each pin with a bolt before final tightening. GB/T 32076.10-2018 §8.1.4 requires aligned holes, clean hole walls and faying surfaces free of dirt, paint and weld spatter; §8.1.5 allows torque only on the nut. Tighten until the plies stop moving, mark nut and bolt, and check for a gap with a feeler gauge.

A gap produces systematic low preload, not random scatter. A 0.2 mm gap consumes 29° of a 45° turn, and the remaining angle develops about one third of target. The bolts pass the angle check and the joint is loose.

Sequence and Elastic Interaction

Bolts are springs in parallel with a joint that is also a spring. Tightening one bolt compresses the plies around it, moving the head and nut of an already tightened bolt closer together, reducing its stretch and clamp load.

Work from the stiff part toward the free edge, or from the center outward. Tighten primary members before secondary. Start near locating pins; if one side is restrained by a weld, start there. GB/T 32076.10-2018 §8.3.4 states these rules.

GB/T 32076.10-2018 §8.3.2 sets initial and re-tightening torque at 50% of final and asks for three passes within 24 h. Many procedures use 30%/60%/100%; the percentages are project practice, the principle is the same.

Field experience across steel joints puts the loss on the first bolts of a single-pass group at 10–20%. A second pass after all bolts are tight restores most of it, which is why the check follows the final pass.

Turn-of-nut and twist-off do not escape interaction. The angle or sheared end certifies that bolt, not its neighbors.

Inspection, Sampling and Records

GB/T 32076.10-2018 §9.2.1 checks large hexagon assemblies 1–48 h after final tightening: 10% of nodes, at least 10 (all if fewer), then 10% of assemblies in each node, at least 2. The inspector marks a line across opposite nut flats onto bolt end, washer and plate, loosens the nut 60°, and retightens to the line with a wrench reading within ±2%. Torque must fall in 0.9–1.1 Mu. A failure doubles the sample; a second failure re-tightens every assembly at that node.

Twist-off inspection is visual: §9.2.2 accepts a detached splined end. Tensioning needs pressure at hold and elongation after transfer; pressure alone records only applied force.

§8.2.1 retests each batch of 3,000 assemblies on 5 samples. §8.2.3 calibrates the wrench each shift with torque error within ±4%. §6.4 limits storage to 6 months, after which coefficient or clamping force is retested.

Item Requirement Source Record
Batch coefficient, large hexagon K 0.110–0.150, Vk ≤ 0.077 GB/T 1231-2024 §6.3.1 Certificate
Batch coefficient, HR/HV 0.10 < k ≤ 0.23, Vk ≤ 0.10 GB/T 32076.10-2018 Table 2 Retest report
Initial and re-tightening 50% of final, within 24 h GB/T 32076.10-2018 §8.3.2 Pass record
Torque check 1–48 h, 10% of nodes, 0.9–1.1 Mu GB/T 32076.10-2018 §9.2.1 Check sheet
Twist-off check Splined end detached on every bolt GB/T 32076.10-2018 §9.2.2 Visual record

Temperature, Lubrication and Winter Torque

K is a friction coefficient in disguise. Lubricant, coating, washer hardness and roughness change it. Temperature changes the lubricant film: grease that is fluid at 20 °C is stiffer at −5 °C, more torque goes into friction, and K rises.

When K rises from 0.13 to 0.15, the same torque delivers 13% less preload, because 0.13/0.15 = 0.867. The winter torque to hold preload is T_winter = 771 × 0.15/0.13 = 890 N·m.

The compensation is valid only if K_winter was measured. A blanket winter percentage is a guess. Measure batch K at the lowest expected temperature, or switch to turn-of-nut, twist-off or tensioning.

GB/T 32076.10-2018 §5.2.1 permits untreated, zinc flake per GB/T 5267.2, hot-dip galvanized per GB/T 5267.3, or another system by agreement. Each has its own K range and lubrication requirement.

Record ambient temperature and lubricant. GB/T 1231-2024 §7.4.6 holds fixture, instruments and assembly at the recorded temperature for 2 h before a coefficient test; GB/T 3632-2008 §6.1 tests at 10–35 °C with arbitration at 20 ± 2 °C. A K value from those conditions is not a K value for a winter site.

Choosing a Method for the Joint

Four questions: what the joint carries, what access the wrench has, what the acceptance clause asks for, and whether batch K is available.

1What does the joint carry?
Static steel frame
✅
torque method with batch K and a sampled check
Bridge, tower or fatigue-critical connection
⚠️
turn-of-nut or twist-off
Equipment base or large stud
🔧
hydraulic tensioning
↓
2What access do you have?
Nut access only
✅
torque, turn-of-nut or twist-off
Bolt end exposed, no nut access
⚠️
tensioner
↓
3What evidence does the clause require?
Batch K certificate and sampled check
✅
torque method
Sheared splined end on every bolt
⚠️
twist-off
Pressure and elongation per bolt
🔧
tensioning

The tree is not exhaustive: a bridge specification may require turn-of-nut and permit twist-off only with engineer approval; one plant may use both torque and tensioning. The higher-cost method is justified where the preload band is tight or the joint is fatigue-critical.

For strength values, see the property class guide. The torque-to-preload relation is worked through in the K factor entry.

The Decision in Three Lines

Torque is a friction estimate: use it with batch K, a calibrated wrench and a sampled check, or do not claim the preload.

Turn-of-nut and twist-off replace the friction estimate with geometry or a shear fuse. Both still depend on a correct snug point, and neither removes elastic interaction.

Tensioning applies load directly, but the gauge reads what was applied. Elongation after transfer is the number that stayed.

Takeaway Checklist
Copy this checklist into your tender documents
Selected conditions
Tap to check, this is your procurement checklist

The preload numbers behind each method are listed for common metric sizes in the bolt torque chart, and the K-factor explainer covers where the scatter between methods comes from.

References

Key Standards

Standards and clauses referenced by this article.

  1. High-strength structural bolting assemblies for preloading — Part 10: Assembly specifications§5.2.1, §6.4, §7.1, §8.1.4, §8.1.5, §8.2.1, §8.2.3, §8.3.2, §8.3.3, §8.3.4, §9.1.1, §9.2.1, §9.2.2, Table 2

    Installation methods, torque-coefficient acceptance, batch retest, wrench calibration, pass sequence, torque check sampling and the 0.9-1.1 Mu window. Values are quoted from the standard; the full text is not reproduced.

  2. High-strength large hexagon head bolted assemblies for steel structures§6.3.1, §7.4.1, §7.4.6, Table 10, Table 14

    Batch torque coefficient 0.110-0.150 with Vk no higher than 0.077; Ki = T/(F*d); M24 stress area 353 mm2 and 10.9S preload window 237-263 kN.

  3. Twist-off high-strength bolted assemblies for steel structures§5.3, §6.1, §7.2, Table 12

    Batch clamping force for M24: mean 248 kN, min 225 kN, max 272 kN, standard deviation no greater than 22.5 kN; 8 assemblies sampled per batch; test at 10-35 C with arbitration at 20 +/- 2 C.

  4. High-strength structural bolting assemblies for preloading — Part 2: Suitability for preloading§4

    Specified preload Fp = 0.7 Rm As and the effective bolt length l_b,eff used in the turn-of-nut calculation.

  5. Mechanical properties of fasteners — Bolts, screws and studs§7.1, Table 3

    Nominal tensile strength for property class 10.9 is 1000 MPa.

Deep Reading

More systematic selection, procurement, or inspection guides.

This page is one entry in the fastener engineering reference. The library covers standards, dimensions, materials, and failure analysis; every figure is traced back to its source.

Continue reading in the library →