Solid Rivet Installation Quality

Solid Rivet Installation Quality (manufacturing-quality/assembly/solid-rivet-installation-quality)

Use when the task is verifying the installation quality of solid rivets at assembly time. Solid rivets are deformation-driven permanent fasteners: the installer selects a rivet long enough to fill the stack and form a shop head, upsets the tail with a squeeze force against the material flow stress, and the driven head must land inside workmanship bands on diameter and height while the hole fill clearance stays within the installation limit. This leaf implements the four checks in pure Python, stdlib only: rivet length selection from the stack and the head-style allowance, the driven shop head geometry verdict, the squeeze force to upset, and the hole fill clearance check, plus a combined installation verdict. It is the deformation-fastener complement of manufacturing-quality/assembly/fastener-installation-quality, which owns threaded and lock-bolt fastener mechanics; hole drilling process control and rivet fatigue allowables are out of scope here.

Domain quick reference

• Rivet length selection: length = stack + allowance, where the allowance is a head-style factor times the shank diameter. Typical practice: protruding heads 1.5 d, countersunk heads 0.8 d (documented workmanship allowance, confirm against the governing process specification for the program).
• Driven shop head geometry: after upsetting, the driven head diameter should fall between 1.4 d and 1.5 d and the driven height between 0.4 d and 0.5 d, d being the nominal rivet shank diameter. A head below 1.4 d or 0.4 d is under-driven (insufficient fill), above 1.5 d or 0.5 d is over-formed; the band is symmetric around 1.45 d and 0.45 d.
• Squeeze force: F = factor * sigma_flow * (pi d^2 / 4), with sigma_flow the flow stress in MPa (N/mm2) and a typical allowance factor of 1.5 on the shank area upset force.
• Hole fill: clearance = hole diameter - rivet diameter; a driven rivet must fill its hole to within 0.1 mm of clearance (typical limit). Interference fits are out of scope for this check.
• Units: dimensions in mm, force in N, flow stress in MPa. The factors above are leaf-local typical values in the same epistemic class as the torque coefficient used by the threaded fastener sibling; they are workmanship practice, not standard data.
• AS9100 frames the process-control context (records, operator certification, traceability around the installation); the relations above are standard engineering methodology, summary-only.

Workflow

1. Fix the installation: stack thickness, rivet shank diameter, head style ("protruding" or "countersunk").
2. Select the rivet length with select_rivet_length: allowance plus stack, 1.5 d protruding or 0.8 d countersunk.
3. Measure the driven shop head after upsetting and judge it with shop_head_verdict against the 1.4 to 1.5 d diameter and 0.4 to 0.5 d height bands.
4. Compute the upset requirement with squeeze_force from the shank diameter, the material flow stress and the allowance factor.
5. Check the hole fill with hole_fill_check: the clearance between the drilled hole and the rivet must not exceed the 0.1 mm limit; a rivet larger than its hole (interference) is rejected as out of scope.
6. Combine the sub-checks with installation_verdict into one verdict with the overall ok flag when shop head and hole fill both pass.
7. Confirm the deterministic checks with the contract test scripts/test_solid_rivet_installation_quality.py.

Worked example

A 4.0 mm protruding rivet over a 6.0 mm stack, driven head 5.8 mm diameter by 1.8 mm height, flow stress 275 MPa, hole 4.08 mm.

• Length, protruding: allowance = 1.5 * 4.0 = 6.0 mm, so the rivet length is 6.0 + 6.0 = 12.0 mm (select_rivet_length real output).
• Length, countersunk: allowance = 0.8 * 4.0 = 3.2 mm, length 9.2 mm.
• Shop head: d_over_d = 5.8 / 4.0 = 1.45, h_over_d = 1.8 / 4.0 = 0.45, both mid-band, ok True.
• Under-driven head 5.0 x 1.2 mm: d_over_d 1.25, h_over_d 0.30, ok False (under the 1.4 d and 0.4 d edges).
• Squeeze force: area = pi * 4.0^2 / 4 = 12.566 mm2; F = 1.5 * 275 * 12.566 = 5183.6 N (real module output 5183.63 N).
• Hole fill: 4.08 - 4.0 = 0.08 mm clearance, within 0.1 mm, ok True; a 4.15 mm hole gives 0.15 mm, ok False.
• Combined: installation_verdict on the good head and 4.08 mm hole returns overall_ok True.

Verification

• Confirm select_rivet_length(6.0, 4.0, "protruding") returns allowance 6.0 mm and length 12.0 mm, and the countersunk case 3.2 and 9.2 mm; doubling the stack adds exactly the stack increment.
• Confirm shop_head_verdict(5.8, 1.8, 4.0) returns 1.45 / 0.45 ok True and (5.0, 1.2, 4.0) returns 1.25 / 0.30 ok False; band edges 1.4 d / 0.4 d and 1.5 d / 0.5 d pass inclusively, and 1.6 d or 0.55 h fail.
• Confirm squeeze_force(4.0, 275) returns 5183.6 N within 0.1 N, and that the force scales with the diameter squared, linearly with flow stress and linearly with the factor.
• Confirm hole_fill_check(4.08, 4.0) passes at 0.08 mm, (4.15, 4.0) fails at 0.15 mm, and the exact 0.1 mm boundary passes (clearance limit is <=).
• Confirm every non-positive stack, diameter, dimension, flow stress or factor, every unknown head style, every hole at or below the rivet diameter (interference), and every negative max clearance raises ValueError.
• Run the contract test offline: python3 scripts/test_solid_rivet_installation_quality.py (30 tests, deterministic).

Related leaves

• manufacturing-quality/assembly/fastener-installation-quality: the threaded and lock-bolt sibling (grip, thread protrusion and clamp load checks); boundary is torque mechanics versus deformation fasteners, this leaf never applies torque-tension relations.
• manufacturing-quality/assembly/ewis-installation-quality: the wiring installation sibling in the same assembly pack.

Pitfalls

• Sizing with the wrong head-style allowance: protruding heads take 1.5 d and countersunk heads 0.8 d, so a countersunk rivet sized at 1.5 d is 0.7 d too long and its shop head will not form to band.
• Judging a driven head on diameter alone: the shop head verdict needs both the diameter (1.4 to 1.5 d) and the height (0.4 to 0.5 d) in band - an over-formed diameter with an under-driven height is still a fail, and a head below either low edge is under-driven rather than acceptable.
• Treating the workmanship values as standard data: the 1.5 d / 0.8 d allowances, 1.4-1.5 d and 0.4-0.5 d bands, 1.5 flow-stress factor and 0.1 mm hole fill limit are documented typical practice in the same epistemic class as the threaded sibling's torque coefficient - confirm them against the governing process specification.
• Feeding an interference fit: a hole at or below the rivet diameter raises ValueError - interference fits are out of scope for this check, not something to score with a negative clearance.
• Applying torque-tension mechanics here: this leaf is the deformation fastener complement of fastener-installation-quality and never applies the F = T / (k * D) relation used for threaded and lock-bolt fasteners.
• Passing on one sub-check: the combined installation verdict requires the shop head and the hole fill to pass together, so a mid-band head with a 0.15 mm hole clearance is still an overall fail.

Behavior contract (gate 3)

Run the deterministic contract test (stdlib unittest, offline):

python3 scripts/test_solid_rivet_installation_quality.py

The test covers the worked-example anchors (12.0 mm protruding and 9.2 mm countersunk lengths, 1.45/0.45 ok and 1.25/0.30 fail shop head verdicts, 5183.6 N squeeze force to 0.1 N, 0.08 mm pass and 0.15 mm fail hole fill), the inclusive band edges and band symmetry identity, the d^2, flow stress and factor scaling laws, the exact 0.1 mm hole fill boundary, the combined installation verdict and its keys, run to run determinism, and ValueError rejection of every non-physical input listed in Verification.

Compliance

• Standards referenced, not reproduced: AS9100 is cited as the process control frame (reference-only per standards-map.yaml); the workmanship factors and relations above are typical engineering practice, summary-only.
• compliance: STANDARDS-REF, gated: false.