Structural Connections: Bolted Joints and Welded Details in Steel Design

A steel structure is assembled from individual members fabricated in a shop and erected in the field. At every point where two members meet, a connection must transfer forces from one to the other without allowing excessive slip, rotation, or premature failure. Connections are consistently where structural failures initiate when they occur, and they are where the greatest percentage of engineering and fabrication cost concentrates in most steel buildings. Understanding both bolted and welded connections, their failure modes, and their appropriate application domains is essential for anyone working with steel structures.

Bolted Connections: Shear and Bearing

A bolt in a joint transfers load between connected plates through one of two mechanisms depending on how it is installed and how the joint is detailed. In a bearing-type connection, the bolt is installed snug-tight (the full effort of a worker using an ordinary wrench), and load is transferred by the bolt shank bearing against the walls of the holes in the connected plates. The bolt is in shear, and the plate is in bearing. In a slip-critical connection, high-strength bolts are tightened to a prescribed pretension (typically 70 percent of the bolt's tensile strength), clamping the connected plates together tightly enough that friction between the faying surfaces, rather than shear and bearing, transfers the applied force.

AISC J3 governs bolted connection design. Two grades of high-strength bolts dominate structural practice. ASTM A325 bolts (now superseded by ASTM F3125 Grade A325) have a minimum tensile strength of 120 ksi for diameters up to 1 inch and 105 ksi for larger diameters. ASTM A490 bolts (F3125 Grade A490) have a minimum tensile strength of 150 ksi. The nominal shear strength of a bolt depends on the ratio of bolt area in shear to tensile area, and whether threads are included or excluded from the shear plane. With threads in the shear plane (the more common field condition), the nominal shear stress is F_nv = 54 ksi for A325 bolts and 68 ksi for A490 bolts.

Four failure modes must be checked for any bolted connection. Bolt shear failure is typically the first check: the bolt shank fails in shear across the plane between the connected elements. Bearing failure of the plate occurs when the bolt crushes or tears through the plate hole; the nominal bearing strength is 2.4F_udt per bolt, where F_u is the plate ultimate strength, d is the bolt diameter, and t is the plate thickness. Tearout occurs when the bolt is too close to the edge of the plate and shears out a block of material; the nominal tearout strength is 1.2L_ctF_u, where L_c is the clear distance from the hole edge to the plate edge or adjacent hole. Net section fracture of the plate must also be checked, since holes reduce the gross section area available to resist tension.

Slip-Critical Connections: When Movement Cannot Be Tolerated

Certain connections must remain slip-free under service loads even though bearing-type behavior would be adequate for strength. These include connections subject to load reversal (where slip in one direction would be followed by slip in the other, causing fatigue damage), connections to oversized or slotted holes, and connections in structures where slip would be functionally unacceptable. For these, slip-critical design is required.

The design slip resistance per bolt is R_n = μ D_u h_f T_b n_s, where μ is the mean slip coefficient for the faying surface condition (0.35 for Class A surfaces such as unpainted clean mill scale, 0.50 for Class B surfaces such as blast-cleaned steel), D_u is the ratio of mean installed bolt pretension to specified pretension, h_f is a factor for fillers, T_b is the minimum bolt pretension, and n_s is the number of slip planes. Class B surfaces with blast-cleaned or roughened finishes transfer significantly more load by friction than Class A surfaces, reducing the number of bolts required or allowing smaller bolt diameters.

Installation of slip-critical connections requires controlled bolt tensioning. Turn-of-nut tensioning rotates the nut a specified fraction of a turn from the snug-tight condition (typically 1/3 to 2/3 turn depending on bolt length and diameter) to achieve the required pretension. Calibrated wrench tensioning uses a torque wrench calibrated daily against a tension-measuring device. Direct tension indicators (DTIs) are washers with surface protrusions that flatten when the required tension is reached, providing a visual verification without requiring torque measurement. Each method has specific requirements for inspection and verification of the pretension achieved.

Welded Connections: Fusion Bonding

Welds join steel by melting the base metal and filler metal together into a single continuous piece. Properly made, a weld is stronger than the base metal it joins. The challenge is ensuring that welding is done correctly, because weld defects such as porosity, undercut, incomplete fusion, or cracks are invisible after the weld is finished and can drastically reduce capacity.

Fillet welds are the most common weld type in structural connections. They are triangular in cross-section, placed in the corner between two plates meeting at an angle. The effective throat of a fillet weld is 0.707 times the leg size (for equal-leg welds loaded in shear), and the design shear strength per unit length is 0.6F_EXX × 0.707w, where F_EXX is the electrode classification strength (typically 70 ksi for E70XX electrodes commonly used with A36 and A992 steel) and w is the weld leg size. A 5/16-inch fillet weld using E70XX electrode has a design shear strength of approximately 5.57 kips per inch of weld length.

Complete joint penetration (CJP) groove welds fuse through the full thickness of the connected element. They develop the full strength of the base metal in tension, compression, and shear. CJP welds require backing bars, root passes, or back-gouging to ensure complete fusion through the full thickness, making them more expensive to produce than fillet welds. They are required at moment connections where tension must be transmitted through the full thickness of a flange or web.

Partial joint penetration (PJP) groove welds penetrate part of the base metal thickness. They are sized like fillet welds with a reduced effective throat and have lower tension capacity than CJP welds. PJP welds in tension applications are checked against the effective throat area, with tension perpendicular to the weld axis limited to 0.9 times the base metal yield in some loading directions where the unfused root creates a stress concentration.

Moment Connections vs. Shear Connections

The structural function of a connection determines which forces it must transfer. A simple shear connection, also called a pinned connection, is designed to transfer only vertical shear from the supported beam to the supporting column or girder. It allows the beam end to rotate freely under load, producing the simply supported beam behavior assumed in gravity-load analysis. Simple connections include shear tabs (single-plate connections), double angles, end plates without moment-resisting bolts, and seated connections. The connection must have enough ductility to accommodate the end rotation that develops under the design loads without fracturing.

A moment connection transfers shear, axial force, and bending moment between the connected members. It creates a rigid or semi-rigid joint that influences how lateral loads are distributed in the frame. The extended end plate moment connection and the welded unreinforced flange moment connection (WUF-W) are common configurations. In moment frames designed for seismic resistance, special moment frame (SMF) connections must be prequalified through testing to demonstrate that they can sustain large inelastic rotations without brittle failure. The flange connection details, column continuity plates, and panel zone reinforcement all influence the behavior under cyclic loading and are governed by AISC 341 for seismic applications.