The Riveting Process, From Rivet Selection to an Accepted Joint

The riveting process is a permanent mechanical joining method that joins two or more parts by placing a rivet through or into the material stack and deforming part of its shank. Once installed, the rivet creates a second head or mechanical interlock that clamps the stack. Joint quality depends on the rivet, hole, access, material pairing, tooling, and verified process window.

This guide starts with five control stages used to organize the analysis, then moves into the choices that matter on a production line: rivet family, forming motion, hole geometry, material pairing, defect diagnosis, monitoring limits, qualification, and the information an equipment supplier needs.

Updated September 2026 · By XCX

What the Riveting Process Actually Is

The riveting process forms a permanent mechanical joint by deforming a rivet after it reaches the intended position in a material stack. The factory head forms the head on one end, while the tail end of the rivet becomes a shop head, flare, bulb, or interlock on the other side.

In practical terms, riveting is a mechanical assembly process; the research institute TWI describes it as generally permanent even though a rivet can be drilled out. That distinction matters: removal is possible, but removal usually destroys the fastener and may alter the hole. A riveted joint is therefore not a reusable bolted joint, and the exact mechanism changes with the rivet family and rivet material.

1. Define the joint — record the materials, stack thickness, access, load direction, environment, appearance, rate, and inspection requirement.
2. Prepare the parts — make and verify any required holes, remove harmful burrs, align the stack, and close unintended gaps.
3. Insert the rivet — confirm family, material, diameter, length or grip range, head style, orientation, and die match.
4. Form the tail — apply the specified axial force, impact, orbital motion, mandrel pull, or self-piercing stroke while supporting the joint.
5. Accept the joint — inspect formed geometry and process evidence, then perform the joint tests required by the release plan.

How does riveting work?

During setting, the riveting tool pushes, squeezes, hammers, rotates, or pulls the rivet so material flows into a retained shape. Inserting a rivet is only the start: to install the rivet, place the rivet through the aligned hole or at the defined self-piercing point, then deform the rivet without damaging the rivet shank.

With a blind fastener, the rivet body—the deformable body of the rivet—forms a rear bulb as the mandrel is pulled; solid rivets form a shop head, while self-piercing rivets flare into the lower sheet against a die. TWI reports tail expansion of about 1.5 times the original stem diameter as a general description, but that is not a universal acceptance number; the drawing, fastener specification, and joint qualification control the finished geometry. A useful definition distinguishes the head of the rivet from the tail of the rivet, and a release requirement must name the applicable limits before a station can install rivets repeatedly.

Rivet Types: What Each Family Needs From the Joint

Rivet selection starts with classification axes, not a memorized family count. Ask whether the riveting method requires access to both sides, whether a prepared hole is allowed, how the tail can form, what head can remain, and which materials must touch. Different types of rivets can belong to several descriptive groups at once, so no fixed count is canonical.

TWI lists eight widely used types—blind, drive, flush, friction-lock, Oscar, self-piercing, solid, and structural steel rivets—as its practical taxonomy. These main types are one cited teaching list, not every rivet form in service; a pop rivet is a common blind-rivet name. The reference table below stays within that taxonomy and its cited formation descriptions; it is not a cross-process selection framework.

One TWI self-piercing-riveting study shows how quickly a family label becomes a specific process: its example fixes aluminium 6016, a 5 mm rivet diameter, 5 mm rivet length, a DZ 9-0.25 die, and 5 mm edge distance. Those values describe that study matrix, not a universal self-piercing-riveting recipe.

Rivet nuts sit at the boundary of this taxonomy. They install from one side but create a reusable internal thread for a separate bolt or screw, so the final joint is not completed by the rivet nut alone. Treat one as a threaded insert when defining assembly sequence and service access.

Self-piercing riveting deserves a similar scope boundary. It is a cold-forming route with a semi-tubular rivet, punch, and supporting die. Readers comparing that equipment family can continue to self-piercing riveting machines; this article keeps its focus on process control.

Hot Riveting, Cold Riveting, and the Tool Motion That Forms the Tail

Hot versus cold riveting is a material-formability decision, while squeezing, impact, orbital motion, and mandrel pulling describe how the forming load reaches the rivet. These axes are related but not interchangeable. A viable route must form the head without cracking the rivet, damaging the stack, or exceeding the required cycle.

EAA calculations give about 345 lbf to set a 3/32-inch solid rivet and 615 lbf, about 2.7 kN, for a 1/8-inch rivet. Its description separates continuous axial force from a squeezer and repeated axial impacts from a rivet gun. Either route fails when the load is not aligned with the shank.

The Pacific Northwest National Laboratory publication on AZ31 magnesium riveting shows why temperature cannot be reduced to “large equals hot.” Its abstract describes magnesium as 33% lighter than aluminium and 75% lighter than steel, then reports brittle head fracture at room temperature, extra cycle time from preheating, and a rotating-hammer process that formed a head in 0.23 seconds while refining grains from about 15 to 2.6 µm.

Production equipment operates on a different scale from hand setting. The Simitch servo-riveting page states an 80 kN capacity, 200 mm effective stroke, 0–333 mm/s speed range, ±2% pressure accuracy, and 0.01 mm repeat positioning for its named system. These are first-party specifications for that system, not a joint recipe or proof that maximum force is preferred.

Once trials show that controlled position and force are required, evaluate the joint, feed, fixture, die, inspection, and traceability interfaces of servo-controlled riveting assembly systems. Equipment selection begins after the stack and acceptance plan are defined.

Hole, Grip, Edge Distance, and Pitch

Hole and grip geometry control how the rivet fills, clamps, and transfers load. Specify the target drill separately from the maximum accepted hole, then match rivet diameter and grip to the actual stack. Edge distance and pitch must follow the governing fastener and joint design, because one universal multiple cannot represent every material or rivet family.

Published supplier guidance shows why a drawing note must name its source and process. Atlas Copco’s solid-rivet sizing guide describes a drill about 0.003 inch larger than the rivet. Orbitform’s solid-rivet design guide describes a rivet diameter 0.010 to 0.015 inch smaller than the hole and ties protrusion to forming motion. These are source-specific design guides, not a universal tolerance stack or a substitute for the current fastener specification.

Numbers That Do Not Decide Joint Quality Alone

Separate supplier guides give a roughly 0.003-inch drill allowance and a 0.010–0.015-inch rivet-to-hole difference; neither is a universal acceptance clearance for a 1/8-inch rivet. Blind-rivet holes may differ again because the body and mandrel system behave differently. Reject a datasheet comparison that merges source-specific guidance into one tolerance stack.

Grip range is the fastener’s permitted stack-thickness interval, not simply body length. Measure the compressed stack, including coatings and gaps that will remain after clamping. Edge distance is measured from hole center to the sheet edge, pitch is center-to-center spacing, and a countersink needs its own angle and depth; all interact with material, thickness, load direction, and failure mode.

Material Pairing and Galvanic Risk

Material pairing must consider electrochemical potential, exposed area, crevice geometry, coating damage, moisture, and the joint’s stress state. Once hole and grip geometry are controlled, a galvanic series can show which metal tends to act as anode or cathode in a stated environment, but potential difference alone does not rank the damage that a real rivet-and-sheet geometry will produce.

University of Delaware corrosion notes place aluminium near −0.8 V and 300-series stainless steel near −0.07 V, a gap of about 0.73 V. Their rivet example changes the outcome without changing that gap: a small aluminium rivet in a large stainless plate is a small anode coupled to a large cathode and can be attacked severely.

Area ratio is still not the whole model. University of Kentucky’s record for Matzdorf and co-authors reports worse nearby coating attack around 316 stainless fasteners than around more noble titanium fasteners in the tested system. Available cathodic current, coatings, electrolyte, and geometry can reverse a simple nobility ranking.

The multi-institutional 2021 study with General Motors co-authors reports less coupled-region corrosion in self-piercing riveted joints than in spot-welded joints because the smaller crevice sustained a higher local pH. Galvanic corrosion still occurred in both. Riveting changed the crevice; it did not remove the mechanism.

Material pairing is one part of the broader sheet-metal fastening decision. For a focused treatment of mixed-metal stacks, see the aluminium joining guide.

What Goes Wrong: Defect Modes in a Riveted Joint

Riveting failures belong to two different lifecycle stages. After material compatibility is bounded, formation defects can still occur during installation: wrong hole, grip, alignment, gap, die, or upset. Load-induced failures appear after a sound joint carries service loads. Mixing the stages misdirects corrective action.

Use the 2-Stage Formation–Service Failure Split

The 2-Stage Formation–Service Failure Split is an editorial diagnostic, not a standard: first decide whether the defect was created during setting or emerged under service load, then choose the inspection and corrective action. A “bad rivet” result is only a symptom label, not a diagnosis. This keeps a malformed head from being confused with a qualified joint that later failed by bearing, shear-out, or fatigue.

Formation defects

For solid riveting, the EAA guide uses 1.5D protrusion before setting and a finished shop head around 1.5D across and at least 0.5D deep as its shop guidance. It also contrasts 36,000 psi AD rivets with softer 16,000 psi A-category rivets and describes a 2 sec to 3 sec driving window for AD material. Those figures belong to that aircraft solid-rivet context and do not transfer to every blind or self-piercing fastener.

Load-induced failure modes

Bearing, net tension, shear-out, tear-out, cleavage, rivet shear, and fatigue cracking describe how a loaded joint may fail. ASTM D5961’s B, C, L, S, and T codes are specifically bearing-response codes for fastened polymer-matrix composite laminate coupons. They are useful vocabulary inside that scope, not a generic metal-rivet defect checklist.

One 2024 carbon/epoxy bolted-coupon study illustrates the geometry interaction: changing edge-distance-to-diameter ratio within the same material system shifted failure from shear-out toward progressive bearing. The lesson is bounded but practical—geometry can change the mode even when material stays fixed; it does not make material or load direction irrelevant.

What In-Process Monitoring Can and Cannot Tell You

A force-displacement curve is a process indicator, not direct proof of joint load capacity. It can reveal changes in force, position, stroke sequence, stack response, or tool behavior, but interior interlock, remaining sheet thickness, cracks, and destructive strength need other evidence. Acceptance should combine signals instead of asking one trace to certify every property.

ISO 12996:2013 specifies tensile-shear testing for single mechanical joints in one-layer and multilayer specimens up to a single sheet thickness of 4.5 mm. Its purpose is to determine mechanical characteristics and failure modes. The standard’s destructive route makes the boundary clear: an in-process trace and a joint-strength result are different evidence.

The 2021 self-piercing-riveting review reports Mucha’s 2014 steel-joint study, in which raising force from 14 to 32 kN increased strength by 36%. The same review warns that higher force can raise residual stress and die wear. Force is a real lever with a real cost, and it is the wrong lever when the fault is a wrong die or bad stack.

Ford Global Technologies LLC patent US20200094310A1 illustrates a pre-stroke die-height check: the system uses a low-force contact position to infer installed die height before authorizing the operation. The example shows that the pre-stroke check and the final setting curve establish different facts. The client guide on force-displacement curve validation explores the adjacent monitoring layer.

When Riveting Test Results Point to Another Joining Route

A riveting trial should trigger a broader route review when the chosen rivet family cannot meet access, material compatibility, surface, load, environment, rate, or repair requirements inside a defensible process window. That is a stop condition for this process, not a verdict that one alternative is universally superior.

Is riveting stronger than welding?

Not as a universal rule. Published monotonic comparisons disagree because material, sheet thickness, geometry, load direction, and process setup change the result. The introduction to a study of aluminium 6111 self-piercing-riveted joints, summarising earlier work rather than reporting its own measurement, states that fatigue strength at 10⁶ cycles was about twice that of spot-welded joints. Treat that figure as directional evidence from a secondary summary, not a universal design value.

Mechanism matters more than a winner label. Interfacial slip can relax stress concentration around a weld nugget, and cold forming can add local work hardening where welding adds heat. Corrosion behavior adds a third axis. This article stops at the conditions that invalidate the selected riveting process; use the metal joining methods framework for actual cross-process selection.

Qualifying the Process: Standards Scope and Guarding

Qualification begins by matching each standard to the joint and test it actually covers. Those process limits then determine which dimensional, blind-rivet, destructive mechanical-joint, installation, and machine-guarding documents apply. A familiar document number provides no protection when its specimen, process, material, or jurisdiction differs from the production case.

ISO 13469:2014 covers test procedures for interference-fit and form-fit blind-rivet joints and lock-bolt joints; it does not cover self-piercing riveting. ISO 12996 covers destructive tensile-shear testing for mechanical joints. The ISO catalogue for rivets records dimensional standards from 1999–2002 at stage 90.93, meaning reviewed and confirmed—not proof of why they remain unchanged.

United States scope has a second trap. 29 CFR 1910.217(a)(5) excludes riveting machines from that section’s mechanical-power-press requirements. It does not exclude them from 29 CFR 1910.212, which requires point-of-operation guarding where machine operation exposes an employee to injury.

“The point of operation of machines whose operation exposes an employee to injury, shall be guarded.”

— eCFR, 29 CFR 1910.212(a)(3)(ii)

Section 1910.212 does not name riveting machines in its illustrative list, so the accurate statement is based on the general duty, not a fabricated named-machine clause. Rivet making is different again: 1910.218(i)(2) points rivet manufacturing to forging-machine provisions. A concise machine-guarding review should follow the application risk assessment.

The Rivet Process Release Record: What to Freeze Before Production

A Rivet Process Release Record binds one selected joint stack, rivet, hole or die, setting motion, controlled variables, inspection method, and abnormal response to one revision. It is a release document for a qualified riveting route, not a supplier-selection scorecard or a comparison among joining technologies.

Complete One Record for One Qualified Route

This record is the process-data attachment that should travel with a tooling or equipment discussion. Machine buying questions remain outside this article; the riveting equipment guide and riveting equipment category own that selection intent.

Frequently Asked Questions