Qualifying Leak Tight Joints on Fuel Tanks and Pressure Vessels
A continuous seam on a fuel tank or air receiver has a different success criterion from a structural weld. Nobody asks whether it looks strong; they ask whether it holds pressure after painting, assembly, thermal cycling and years of vibration. Resistance seam welding is well suited to these parts because it creates a chain of overlapping nuggets that is both fused and forged under pressure, without adding filler material or a separate sealant. But the difference between a seam that passes a bench test and one that stays tight in service lies in qualification: a documented chain linking material condition, machine settings, section evidence and a leak test that reflects real duty. That chain is what a properly commissioned Seam Welding Machine is built to support.
Define the Joint and the Test Before Setting the Machine
Qualification starts with a drawing that states the overlap, the allowable mismatch, the seam path, the required leak rate and the test method. On thin low-carbon steel enclosures, a lap of a few millimetres is typical, and the machine range of 2+2 mm through 3.5+3.5 mm defines the thickness envelope within which a single recipe can be expected to work. The test should be appropriate to the product: a low-pressure air-under-water test for a small tank, a calibrated pressure-decay test for a production line, or a helium or tracer test where very small leaks matter. Deciding this after the first welds are made invites the common mistake of tuning the machine until a convenient test passes, rather than until the joint is genuinely sound.
Control the Inputs That Are Not on the Controller
A weld recipe records current, speed, force and wheel geometry, but several of the strongest influences on leak tightness sit outside the controller. Surface condition is first among them: drawing compound, oil, oxide scale, moisture and zinc coating all change contact resistance at the faying surfaces. Material lot changes, even within the same specification, can shift the resistance enough to move a marginal process into expulsion. Overlap consistency is second: a joint that varies from 3 mm to 6 mm along its length changes both the current path and the shunting behaviour. Fit-up and clamping are third, because a gap between sheets lets the wheels press the metal apart instead of forging it together. A qualification programme should specify cleaning, acceptable overlap tolerance and fixture clamping force with the same precision as the electrical parameters.
Establish the Process Window With Sections
Before any finished part is tested, build a heat input ladder on production-representative coupons. Hold speed in the middle of the 0.8 to 3.2 metres per minute range and force at a measured, repeatable value, then increase current in small increments. Section and etch each coupon transversely and longitudinally. The longitudinal section is the critical one for leak tightness: it shows whether successive nuggets overlap by enough margin to leave no unfused path between them. A transverse section alone can look perfect while a longitudinal gap permits a slow leak. Record nugget width, penetration and overlap percentage against each setting, and identify the centre of the window rather than the lowest current that happens to produce a bond.
Dual closed-loop feedback control, which holds motor speed and therefore welding speed constant as load varies, is especially important on circumferential seams. A cylindrical tank presents changing inertia and friction through a revolution; without speed regulation, the drive slows at the heavy part of the rotation and heat input rises locally. Qualification should therefore include sections taken at several angular positions around a real part, not only from a flat coupon. If nugget overlap varies with position, the problem is the drive or the fixture, not the current setting.
Force, Cooling and the Long-Seam Drift
Leak failures that appear only on long seams usually come from drift rather than a bad initial setting. Electrode force that decays as a pneumatic regulator warms, or coolant flow that falls below the specified roughly 120 litres per hour, changes contact resistance progressively along the weld. Wheels that heat up expand, change their contact patch and wear faster. During qualification, produce at least one seam of full production length and section its beginning, middle and end separately. Compare nugget geometry across the three locations. A stable process will show nearly identical overlap throughout; a graded result tells you to investigate force stability, cooling and wheel dressing before adjusting current.
Document the wheel diameter and edge radius at the time of qualification and re-enter the diameter into the controller after every dressing. Because travel speed is derived from wheel geometry, a dressed-down wheel with an unchanged parameter quietly changes heat input per unit length. Similarly, verify that the roughly 20 mm electrode travel is free and that the upper arm is not deflecting when the cylinder applies force on a formed edge or a flange.
Leak Testing That Reflects Service
A bubble test on a bench is a useful screening tool but a weak qualification test. Real service includes pressure cycles, temperature changes, vibration and occasional mechanical impact on the seam area. Where the product standard permits, add a pressure-cycling test on a small sample of qualification parts, followed by a final leak measurement. Include at least one part welded at the low end and one at the high end of the allowed parameter window, because a process that is only sound at its nominal setting is fragile in production. Retain the test records with the recipe version, material lot, wheel state and coolant readings so that any later field failure can be traced to a known configuration.
Reject classification should also be defined. A leak at a seam start or stop, at a cross seam, or at a formed corner has a different cause from an intermittent leak distributed along the whole seam. Start and stop defects point to current ramping and overlap at the run-on tab; corner defects point to fit-up and speed changes; distributed leaks point to nugget overlap or force drift. Sorting rejects by location multiplies the diagnostic value of every failed part and shortens the path to a corrective action.
Production Control and Re-Qualification
Once qualified, the process needs a control plan: a daily coupon or a defined first-off check, a periodic section audit, continuous monitoring of coolant flow and air pressure, and a scheduled wheel dressing interval based on metres of seam rather than elapsed time. Every change that touches the physics of the joint should trigger some level of re-qualification — a new material supplier, a change in coating, a replacement cylinder, a controller update, a wheel alloy change, or a fixture modification. The depth of the re-test can be proportional to the change, but the decision should be recorded rather than left to informal judgement.
What a Good Qualification Package Contains
The final package is not a single certificate. It is a folder containing the joint drawing and leak criterion, the cleaning and fit-up instruction, the measured force calibration, speed verification at the recipe setpoint, the heat input ladder with etched transverse and longitudinal photographs, sections from multiple positions on a production-length part, pressure or leak test results including window-extreme samples, the wheel and coolant baseline, the stored controller recipe, and the re-qualification triggers. With that package, a manufacturer can demonstrate not merely that one tank passed a test, but that the process is understood, bounded and repeatable — which is what customers and auditors actually mean when they ask whether a welded seam is qualified.