A stainless steel flange joint that leaks on start-up is rarely a machining defect. In most cases the flange dimensions, the gasket and the fasteners are all within specification, and the joint still fails. The bolts never compressed the gasket evenly, or never compressed it enough. Torque is the number written on the work pack; preload is the number that actually seals the joint, and the two are not the same thing.
This guide covers what a bolted joint must achieve, why equal torque does not mean equal bolt load, how the tightening sequence changes the result, and what matters specifically with austenitic stainless flanges and bolts.
A joint must satisfy two conditions at once. During assembly, the bolts have to compress the gasket enough to close the leak paths in the gasket material — the seating stress. In operation, the remaining bolt load must stay high enough to hold that compression while internal pressure pushes the flanges apart and thermal movement changes the joint geometry.
Between those conditions sits a narrow window. Too little load and the gasket leaks. Too much load crushes the gasket, yields the bolts, or rotates a thin flange until the gasket is compressed at the outside diameter and barely touched inside. Every practical decision in bolt-up — bolt size, lubrication, method, sequence — is an attempt to land inside that window, repeatably.
The window also depends on the gasket: a spiral wound gasket needs a defined seating stress and then behaves elastically, a PTFE-based sheet gasket compresses easily but creeps, and a ring joint gasket seals by metal-to-metal wedging. A torque figure copied from another project with a different gasket is not a specification; it is a guess with a number attached.
Turning a nut does not directly stretch a bolt. It applies torque, and only part of that torque becomes tension. The remainder disappears into friction in the thread helix and under the nut face. The relationship is usually written as T = K × F × d, where T is torque, F is the target preload, d is nominal bolt diameter and K is the nut factor — an empirical coefficient bundling thread friction, under-head friction and thread geometry into one number.
Because K is empirical, it carries the uncertainty of everything it summarises. A dry, as-received fastener and the same fastener with a molybdenum disulphide lubricant differ enough that identical torque produces clearly different preload. Coating, thread finish, washer hardness, thread damage and whether the fastener has already been tightened once all move K. In practice only a limited share of applied torque becomes useful tension, and that share depends on conditions the torque wrench cannot see.
Published nut factors are averages, and real joints scatter around them. Torque-controlled tightening is therefore approximate by nature: fast and cheap, but a way of getting close to a target preload rather than guaranteeing one.
| Method | Practical assessment |
|---|---|
| Torque-controlled (calibrated wrench) | Low equipment cost, works on any bolt size. Accuracy limited by friction scatter, so preload can vary widely between bolts in one joint. Suitable for general service. |
| Turn-of-nut (angle control) | Snug first, then rotate each nut a defined angle. Friction is largely bypassed after snugging, improving repeatability, especially on long bolts. Needs a defined snug condition. |
| Hydraulic tensioning | Loads the stud directly and seats the nut with almost no torque, producing very even load. Needs radial clearance, longer studs and one tool per bolt or group. |
| Direct elongation measurement | Ultrasonic or mechanical bolt stretch converted to load — closest to true preload. Used on critical or large-diameter joints; needs prepared bolt ends and trained operators. |
A flange is an elastic ring, not a rigid plate. Tightening one bolt to full load pulls that side down and lifts the opposite side, so the gasket is compressed unevenly until every bolt is brought up. On thin or large flanges that rotation can open a leak path on the far side.
The countermeasure is a cross or star pattern combined with several passes at increasing load. Working sequentially around the circle — 1, 2, 3, 4 — drags the flange progressively in one direction and is the most common assembly error on large joints.
Because K depends on friction, lubrication is part of the torque specification, not a convenience. Two identical joints, one dry and one with a graphite or molybdenum-based anti-seize compound, reach different preload at the same wrench setting. A torque figure quoted without a lubricant has no defined meaning.
The rules are simple: specify the lubricant in the procedure, apply it to thread and bearing face, apply it to every bolt, and record it. Never mix products on one joint and never lubricate some bolts and leave others dry. If the lubricant changes, the torque target changes with it — otherwise a joint intended to be correctly loaded can end up over-stressed.
Austenitic stainless fasteners behave differently from carbon steel bolting. The first issue is galling: the alloy work-hardens rapidly and its oxide layer does not regenerate quickly under local pressure, so thread surfaces in sliding contact tear and cold-weld. The result is a seized nut, damaged threads and a bolt that never reaches its intended load. Slow tightening, clean threads, matched grades and a suitable anti-seize compound are the standard defences.
The second issue is differential expansion. Stainless flanges and bolts expand more than carbon steel components, so at elevated temperature the flange, gasket and bolt stack move at different rates and joint load can fall or rise. Where temperature makes this significant, bolt grade, gasket type and any re-torque requirement should come from the project specification rather than a generic table.
Bolt load is not constant over the life of a joint: gaskets creep, bolts relax and thermal cycles ratchet the assembly. Graphite gaskets tolerate elevated temperature far better than PTFE-based materials, which creep appreciably and lose load even at moderate temperature.
Re-torquing after the first thermal cycle is specified on some systems and prohibited on others. Where required, it is a controlled operation with its own procedure, permit, thermal limits and protective equipment — not an informal second pass with a wrench on a hot line.
Bolt-up is a documented process, and the records are what allow a joint to be audited or reproduced. A useful joint record carries the flange size, class and facing; the flange and bolt grades with their certificates; the gasket type and batch; the lubricant; the tightening method and pass-by-pass targets; the wrench calibration reference; and the technician and inspector names.
Flange certificates are issued against specifications such as ASTM A182/A182M, covering forged or rolled alloy and stainless steel flanges for high-temperature service, and inspection documents are often requested in EN 10204 3.1 format. Where piping is designed to a code such as ASME B31.3, that code's assembly, examination and testing requirements take precedence over workshop habit.
A flange joint seals when the gasket is compressed evenly, to a defined load, and stays that way in service. Torque is only the input; friction decides how much becomes useful. Controlling method, sequence, lubrication and records turns bolt-up into a repeatable procedure — which is what a leak-free stainless steel joint depends on.
Only if flange size and class, gasket type and thickness, bolt grade and diameter, surface condition and lubricant all match, and the target preload is the same. Change any one and the figure must be re-derived.
Galling is adhesive wear between stainless surfaces under pressure. The alloy work-hardens at contact points and the passive oxide layer cannot reform fast enough, so threads tear and cold-weld. Prevention means clean threads, matched grades, a suitable anti-seize lubricant, slower tightening and no high-speed impact tools on dry threads.
No. Calibration controls the torque applied; friction controls how much becomes bolt load. A calibrated wrench with inconsistent lubrication, a single pass and a circular pattern will still produce an uneven joint. Calibration is one control, not a substitute for a procedure.
The two flanges form one joint, so they are tightened as one assembly. Align the faces, centre the gasket, close the gap evenly, then apply the sequence across all bolts. Fully tightening one flange first distorts the gasket and creates the uneven compression the sequence is meant to avoid.
Work across the diameter in several passes rather than around the circle: opposed pairs in turn — 1 and 5, 2 and 6, 3 and 7, 4 and 8 — at increasing fractions of target, then re-check from the first bolt.
Bolt-up only works when flange, gasket and fastener are specified as a system. When sending an inquiry, include the flange type, nominal size, pressure class, facing, material grade and the standard you are working to — for example ASME B16.5 or EN 1092-1 — with gasket type, bolt grade and quantity. That lets facing finish, bolt hole alignment and tolerances be checked against each other before manufacture.
You can review the range on our alloy and stainless steel flange page and send joint details through the contact page for a dimensional and material review.
This article provides general information only. Final selection and assembly practice must be based on the applicable standards, the project specification, and confirmation by qualified engineering personnel.
Pessoa de Contato: Kelly Yang
Telefone: 86-15301677597