Four-Point Bend Testing of Dissimilar Metal Weldments

Four-point bend testing is the principal mechanical method for verifying the ductility, fusion integrity, and overall soundness of dissimilar metal weldments used in nuclear, oil and gas, petrochemical, and power generation equipment. By loading the specimen at two symmetric points, the test produces a uniform bending moment across the entire weld zone, making it far more revealing than three-point bending for joints between materials with mismatched strength and stiffness. This article explains the specimen types, test items, procedures, acceptance criteria, and typical applications of four-point bend testing of dissimilar metal welds, with reference to ISO 5173, ASTM E290, GB/T 2653, and related welding qualification standards.

Technical Overview

A dissimilar metal weld (DMW) joins two base metals with different chemical compositions and mechanical properties, such as carbon steel to austenitic stainless steel, low-alloy steel to nickel-based filler metal, or aluminum to steel. Because the weld metal, the two heat-affected zones, and the two fusion boundaries each respond differently to deformation, standard tensile testing alone cannot fully characterize the joint. Bend testing is therefore required to force plastic strain directly onto the weakest interface and reveal hidden discontinuities.

Four-point bend testing applies the load through two symmetric loading rollers while the specimen rests on two support rollers. This configuration creates a region of constant bending moment between the two loading points, so the entire weld cross section, including the weld metal, both fusion lines, and both heat-affected zones, is strained uniformly. This is a decisive advantage for dissimilar joints, where defects such as lack of side-wall fusion, hard martensitic layers, carbon migration zones, or interfacial delamination may be located anywhere across the weld width. The method is standardized in ISO 5173, ASTM E290, GB/T 2653, and AWS B4.0, and is routinely specified in welding procedure qualification frameworks such as ISO 15614-1 and ASME Section IX.

In a rectangular specimen of width b and thickness d under total load P, with the distance between each support and the adjacent loading point equal to a, the maximum fiber stress equals 3Pa/(bd²). The support span is typically set at twice the loading span, and the maximum strain develops on the two outer surfaces while the neutral axis passes through mid-thickness. For dissimilar joints, the elastic modulus mismatch shifts the true neutral axis of the composite section, which is one reason the test is a demanding screening tool for interface quality.

Testing Samples

Transverse Bend Specimens (Face and Root Bend)

Transverse specimens are cut across the weld so that the weld axis is perpendicular to the specimen length. In a face bend test, the weld cap side is placed in tension; in a root bend test, the root side is placed in tension. These specimens, typically about 10 mm thick with the weld centered in the gauge region, are used for thinner joints and assess the ductility of the weld metal, the fusion boundary, and the near-surface heat-affected zone on the tensioned side.

Side Bend Specimens

Side bend specimens expose the full cross section of the joint by placing the through-thickness direction in tension. They are mandated for thicker welds, typically above 10 to 12 mm, and for joints made with multiple passes or multiple processes. For dissimilar metal welds, side bend testing is especially valuable because it simultaneously strains the entire depth of both fusion boundaries and is highly sensitive to lack of fusion, slag, and interfacial cracking.

Longitudinal Bend Specimens

Longitudinal specimens are oriented with the weld axis parallel to the specimen length, so the base metals and the weld metal are bent together along the same fiber. ISO 5173 and related qualification standards recommend longitudinal bend testing when the two base metals differ significantly in tensile strength, because in transverse bending the weaker material would deform and fail before the stronger material and the weld interface are adequately strained. Longitudinal bending overcomes this shielding effect and directly interrogates the fusion boundaries.

Specimen Preparation Requirements

• Remove the weld reinforcement and root beads by machining unless the applicable specification permits them to remain, taking care not to alter the material through heating.
• Machine away all thermally cut or sheared edges and slightly round the specimen edges, with a corner radius not exceeding about 10 percent of the thickness, to prevent edge tearing artifacts.
• Clean the surfaces of oxide, scale, spatter, oil, and marking ink, and clearly scribe the weld centerline so the weld can be positioned accurately within the constant-moment region.
• Identify each specimen with material heat numbers, joint configuration, and specimen orientation to ensure full traceability within the test report.

Testing Items

Depending on the product specification and qualification code, a four-point bend test program for dissimilar metal weldments may include the following items:

• Face bend ductility of the weld and adjacent heat-affected zone.
• Root bend ductility, which is critical for single-sided joints with restricted access.
• Side bend soundness across the full joint thickness for thick or multi-pass welds.
• Longitudinal bend evaluation for joints between base metals with pronounced strength mismatch.
• Verification that the required bend angle, commonly 180 degrees, is achieved around a mandrel or roller diameter defined as a multiple of specimen thickness, typically 2t to 6t.
• Detection and sizing of surface-breaking defects on the tensioned face, including cracks, lack of fusion, incomplete penetration, inclusions, and interfacial delamination.
• Optional quantitative outputs such as maximum flexural load, apparent flexural stress, and the full load–deflection curve for research or fitness-for-purpose assessments.

Testing Methods

Apparatus and Loading Configuration

The test is performed on a calibrated universal testing machine fitted with a four-point bend fixture consisting of two support rollers and two loading rollers. The rollers must be free to rotate and long enough to exceed the specimen width, and their diameter is selected so that the specimen contacts only the rollers and not the fixture shoulders. A support-to-loading span ratio of 2:1 is conventional, and the fixture alignment is checked so that the loading axis remains perpendicular to the specimen throughout the bend. Displacement is applied under controlled crosshead speed, and load and deflection are recorded continuously.

Test Procedure

• Center the weld, at the fusion line or full joint width as required, precisely in the constant-moment zone between the two loading rollers.
• Apply the load at a steady, strain-controlled rate, avoiding impact loading that could mask or artificially create discontinuities.
• Continue bending until the specified angle is reached, commonly 180 degrees, or until the specified outer-fiber strain condition defined by the standard is satisfied.
• Remove the load, extract the specimen, and examine the tensioned surface and both side faces visually and under magnification of at least 10x.
• Measure and record the length and location of any cracks or open defects relative to the weld centerline and fusion boundaries.

Evaluation and Acceptance Criteria

Acceptance is governed by the applicable fabrication code and qualification standard. A common criterion is that the specimen shall show no cracks or other open defects exceeding 3 mm, or approximately 1/8 inch, in any direction on the tensioned surface. Small cracks emanating from specimen corners, provided they are shown to be geometric rather than metallurgical in origin, are frequently judged separately from weld discontinuities. For dissimilar metal welds, particular attention is paid to cracking along the fusion boundary and to delamination between layers, since these failure modes indicate inadequate bonding across the dissimilar interface rather than simple ductility exhaustion. All results, including photographs of the tested surfaces, are documented in a traceable test report.

Applications

• Nuclear power generation: qualification and surveillance of safe-end dissimilar metal welds joining low-alloy steel reactor pressure vessel nozzles or piping to austenitic stainless steel piping using nickel-based filler metals.
• Oil and gas: procedure qualification for clad and lined pipes, flowlines, and pressure vessels that combine carbon steel strength with corrosion-resistant alloy layers.
• Petrochemical and chemical processing: quality control of stainless-clad plate welds and of transition joints between reactor vessels and stainless steel internals.
• Conventional and renewable power: verification of ferritic-to-austenitic tube joints in boilers, heat recovery steam generators, headers, and heat exchangers.
• Cryogenic and LNG systems: integrity assessment of aluminum-to-stainless steel transition pieces used to join low-temperature aluminum piping to steel structures.
• Shipbuilding, rail, and aerospace: testing of structural transition joints, including explosion-bonded aluminum–steel and aluminum–titanium laminates.
• Welding procedure and welder qualification programs executed per ISO 15614-1, ISO 9606-1, and ASME Section IX, as well as research programs in which four-point loaded weld specimens are used for environment-assisted cracking studies.

FAQ

Why is four-point bending preferred over three-point bending for weldments?

In three-point bending the maximum stress occurs only under the single loading nose, so a defect just millimeters away from that line may escape detection. Four-point bending generates a uniform bending moment between the two loading points, so the entire weld zone, including both fusion boundaries of a dissimilar joint, is strained equally and any discontinuity within the weld is exposed.

Which standards govern four-point bend testing of welds?

The principal standards are ISO 5173 for bend tests on welds in metallic materials, ASTM E290 for bend testing for ductility, GB/T 2653 as the corresponding Chinese national standard, and AWS B4.0 for mechanical testing of welds. Qualification standards such as ISO 15614-1 and ASME Section IX specify when each bend orientation must be performed and define the acceptance criteria.

Why are longitudinal bend specimens specified for dissimilar metal joints?

When the two base metals differ substantially in tensile strength, a transverse bend specimen will deform preferentially in the weaker metal, which shields the stronger metal and the weld interface from the required strain. A longitudinal specimen bends all constituents together along the same radius, ensuring that both fusion boundaries are genuinely tested.

What is the typical acceptance criterion after bending?

The most widely used criterion is that no crack or other open defect on the tensioned surface exceeds 3 mm, approximately 1/8 inch, in any direction. Specific product standards may impose stricter limits, and for dissimilar joints any cracking along the fusion boundary is normally treated as a cause for rejection regardless of length.

Can the test produce quantitative data as well as pass/fail results?

Yes. While the primary purpose is a ductility and soundness assessment, instrumented four-point bend testing yields maximum flexural load, apparent flexural stress, and complete load–deflection curves, which are valuable for comparing filler metals, welding procedures, and post-weld heat treatments in dissimilar joint development.

Where should such testing be performed?

Testing should be carried out by accredited third-party laboratories or in-house facilities with calibrated universal testing machines, qualified four-point bend fixtures, and inspection personnel capable of evaluating defects at the required magnification, ensuring that results are traceable and acceptable to certification bodies and end users.