Beam Splice Welding Guide: CJP Flange Welds, Web Connections, and Weld Sequencing

A beam splice connects two beam sections end-to-end to create a continuous member longer than a single mill length. The standard detail uses CJP groove welds on both flanges and a bolted web splice plate. The flanges carry the bending moment, so they get full-penetration welds. The web handles shear and gets a bolted connection with splice plates.

Beam Splice Anatomy

A typical beam splice has four main components:

Flange CJP groove welds. Both the top and bottom flanges are joined with complete joint penetration groove welds. These welds must develop the full strength of the flange material because they’re transferring bending forces across the splice.

Web splice plates. Steel plates bolted to both sides of the web transfer shear forces across the splice. The plates are typically sized by the engineer based on the shear demand at the splice location.

Weld access holes. Shaped cutouts in the web at the web-flange junction provide clearance for welding the flange CJP joints. Without access holes, the web blocks the welding electrode from reaching the flange joint at the web-flange intersection.

Backing bars. Steel flat bars placed behind the flange groove joint support the root pass and prevent blow-through. Standard backing is 1/4 x 1 inch steel bar, tack-welded in place before the groove weld is deposited.

Flange CJP Groove Welds

The flange welds are the critical element of a beam splice. They must achieve complete joint penetration through the full thickness of the flange. Incomplete fusion or lack of penetration at the root compromises the splice’s ability to transfer bending forces.

Joint Preparation

Flange preparation for a beam splice typically uses a single-bevel or single-V groove configuration:

The bevel is typically cut in the shop using a beveling machine, plate processor, or thermal cutting (plasma or oxy-fuel) followed by grinding to remove the heat-affected material. Field beveling is done with portable bevelers or grinders when shop beveling isn’t practical.

Backing Bar Installation

The backing bar sits behind the root of the groove joint, bridging the gap between the two flange ends. It’s tack-welded to the back side of the bottom flange (for overhead welding) or held in position by fit-up for the top flange (which is welded in the flat or horizontal position).

Backing bar fit-up must be tight to the flange surface. Gaps between the backing bar and the flange create a pathway for porosity and incomplete fusion at the root. The tack welds holding the backing bar must not interfere with the groove weld passes.

For demand-critical welds under AWS D1.8 (seismic applications), the backing bar is removed after welding, the root is back-gouged to sound metal with an air-arc or grinder, and a reinforcing back weld is deposited. This eliminates the notch effect at the backing bar interface.

Welding Sequence for Flange CJP

The flange groove weld is deposited in multiple passes, starting from the root against the backing bar and building up to the cap:

1. Root pass: Deposited against the backing bar. Must achieve fusion to the backing bar and both groove faces. This pass sets the foundation for the entire weld.
2. Fill passes: Build up the groove cross-section. Each pass must fuse to the previous pass and both sidewalls. Bead placement progresses from one side of the groove to the other, stacking passes to fill the joint evenly.
3. Cap pass: Final surface layer. Must provide a smooth, uniform crown with complete fusion to both bevel edges. Reinforcement height is typically limited to 1/8 inch maximum.

For overhead flange welds (the bottom flange, welded from below), the technique demands precise heat control. The puddle wants to drip, so shorter arc length, faster travel speed, and lower amperage per pass are typical compared to flat-position work.

Weld Access Holes

Weld access holes (sometimes called “cope holes” or “rat holes”) are cutouts in the beam web at the flange splice location. They serve two purposes: provide physical clearance for the welding process and reduce stress concentration at the web-flange junction.

AISC and AWS Requirements

Both AISC 360 and AWS D1.1 specify requirements for weld access hole geometry:

The access hole geometry matters because a poorly shaped access hole creates a stress concentration that can initiate fatigue cracks or brittle fractures. Sharp re-entrant corners are particularly dangerous. All thermally cut edges must be ground to a smooth profile.

Cutting Access Holes

Access holes are typically cut in the shop before the beam is erected. The cutting method depends on the shop’s equipment:

• Plasma cutting followed by grinding gives the best results for most shops
• Oxy-fuel cutting works but requires more grinding to achieve a smooth surface
• Mechanical cutting (hole saw, milling) produces the cleanest finish but is slower

Regardless of the cutting method, the finished access hole surface must be ground smooth and free of notches. The inspector will check the surface condition as part of the fabrication inspection.

Web Connection Options

Bolted Web with Welded Flanges (Most Common)

The standard beam splice uses CJP welded flanges with a bolted web splice plate. This is the most common configuration because:

• The flanges carry bending moment and need full-strength CJP connections
• The web primarily carries shear, which bolted connections handle efficiently
• Bolted web connections are faster to install in the field than welded web connections
• The combination avoids the complexity of welding in the tight space between flanges

The web splice plate is designed by the engineer for the shear demand at the splice. Typical details include a single plate on each side of the web, with high-strength bolts (A325 or A490) in bearing or slip-critical connections depending on the design requirements.

Fully Welded Splices

Some beam splices require CJP welds on both the flanges and the web. This is less common but may be specified when:

• The splice is located in a zone of high combined bending and shear
• The engineer requires a fully rigid splice
• The beam is a built-up section with thick web plates

Welding the web splice requires careful sequencing to control distortion, because the web weld runs perpendicular to the flange welds and introduces competing shrinkage forces.

All-Bolted Splices

All-bolted beam splices use splice plates on both flanges and the web, connected entirely with high-strength bolts. These are faster to erect than welded splices and don’t require welders in the field. The tradeoff is bulk: the flange splice plates and bolts add significant weight and visual mass to the connection.

Weld Sequence to Minimize Distortion

Weld sequencing on beam splices controls distortion and residual stress. An improper sequence can pull the beam out of alignment, create locked-in stresses, or cause lamellar tearing in thick flanges.

Recommended Sequence

1. Complete the bottom flange weld first (if welding from below) or the flange that will be welded in the less favorable position.
2. Allow the first flange to cool to below the maximum interpass temperature before starting the second flange.
3. Weld the second (top) flange. The shrinkage of the second flange weld counteracts some of the distortion from the first flange weld.
4. Complete the web splice (if welded) after both flange welds are done.

Some engineers specify alternating passes between flanges: a pass on the bottom flange, then a pass on the top flange, back and forth until both are complete. This balanced approach minimizes the differential shrinkage that causes angular distortion.

Distortion Control Tips

• Clamp or brace the beam before welding to resist movement
• Use the minimum number of passes consistent with sound welding practice
• Avoid excessive weaving, which increases heat input and distortion
• Monitor the beam straightness as welding progresses and adjust the sequence if distortion develops
• Backstep technique (welding short segments in the direction opposite to overall progression) reduces longitudinal shrinkage

Inspection Requirements

Beam splice welds on structural steel projects are inspected per the project quality assurance plan, which references AWS D1.1 acceptance criteria.

Visual Inspection

Every weld receives visual inspection. The CWI checks profile, size, appearance, and visible discontinuities against D1.1 Table 8.9 criteria. The backing bar interface, access hole surfaces, and weld termination points all receive close attention.

Ultrasonic Testing (UT)

CJP groove welds on beam flanges typically require volumetric inspection in addition to visual. Ultrasonic testing is the standard method for field welds because it provides immediate results and can be performed on the erected structure.

UT examines the weld for internal discontinuities: lack of fusion, slag inclusions, porosity, and cracks. The UT technician scans the weld from the flange surface and reports any indications that exceed the D1.1 acceptance criteria.

Radiographic Testing (RT)

Radiography is sometimes specified for shop-welded splices where access allows film or digital detector placement. RT provides a permanent record of the weld’s internal condition but requires radiation safety controls that make it impractical for most field applications.

The engineer’s inspection plan specifies which NDT methods are required and the percentage of welds to be tested. Critical splices in high-stress locations may receive 100% UT, while routine splices might be inspected on a sampling basis.

Back to structural steel welding for more structural topics. See also structural welding procedures for the WPS and qualification details behind beam splice welding.