*CISC provides this column as part of its commitment to the education of those interested in the use of steel in construction. Neither the CISC nor the author assumes responsibility for errors or oversights resulting from the use of the information contained herein. Suggested solutions may not necessarily apply to a particular structure or application and are not intended to replace the expertise of a licensed professional engineer or architect.*

QUESTION 1: Connection design forces in seismic applications sometimes depend on whether or not the governing failure mode is ductile. Which failure modes are considered ductile?

ANSWER: Capacity-protected elements and connections in seismic force-resisting systems are required to resist the gravity loads together with the lateral loads corresponding to the probable resistance of the yielding elements (CSA S16:19 Clause 27.1.2.2). However, the connection design forces need not exceed those based on Rd Ro = 1.3 when the governing failure mode is ductile, or Rd Ro = 1.0 otherwise.               For Conventional Construction (Clause 27.12.2, Rd Ro = 1.5 x 1.3), and depending on the site-specific seismic parameters, connections must be proportioned so that the expected failure mode is ductile, or
else designed to resist gravity loads combined with the seismic load multiplied by Rd. The CISC Commentary on CSA S16:19 in Part 2 of the Handbook of Steel Construction (Clauses 27.1.2 and 27.12) lists the following ductile failure modes for connections of primary framing members such as bracing connections, yielding in tension, bolt bearing and gusset plates detailed for ductility. For moment frames, extended end-plate and flangeplate moment connections are considered ductile if appropriately proportioned. Welded connections consisting of fillet welds may not possess sufficient ductility and should be designed for amplified loads.

 

QUESTION 2: When installing pretensioned bolts on a project with a heavy coating system, it was found that the bolts seem to relax after installation as the coating compresses in the joint. Are there any recommendations to ensure that the bolts retain their pretension?

ANSWER: Bolt relaxation in joints with thick coatings is a known issue. According to the RCSC Specification for Structural Joints Using High-Strength Bolts (2020, Commentary on Clause 8.2), joints with
galvanized coatings are particularly susceptible to bolt relaxation, especially those consisting of many plies of thickly coated material. In such cases, repretensioning of bolted assemblies may be required after a settling-in period following initial pretensioning. The relaxation characteristics of assemblies with galvanized plates and bolts are about twice as great as those with plain bolts and connected material, according to
the Guide to Design Criteria for Bolted and Riveted Joints (1987, Section 4.4). Bolt force relaxation can be expected to increase as the grip length decreases. And for a given grip length, increasing the number of plies will also lead to an increase in bolt relaxation.

 

“WHEN SINGLE ANGLES ARE DESIGNED ACCORDING TO CSA S16:19 CLAUSE 13.3.2.2 (INDIVIDUAL MEMBERS AND PLANAR TRUSSES; SEE FIGURE 1), THE VALUE OF KL/R IS NEVER LESS THAN 72, EVEN WHEN L/RX = 0.”

 

QUESTION 3: In Part 4 of the Handbook ofSteel Construction, the tables for single angle struts connected through one leg indicate a factored compressive resistance of Cr = 362 for a L102x102x9.5 member of
zero length. But the tables for double-angle struts indicate Cr = 1,170 kN for the same member size. Shouldn’t the single-angle resistance be half the double-angle value when KL = 0?

ANSWER: When single angles are designed according to CSA S16:19 Clause 13.3.2.2 (individual members and planar trusses; see Figure 1), the value of KL/r is never less than 72, even when L/rx = 0. The factored “yield load,” Cro = φ A Fy , would be the expected resistance when the length is zero, at least in theory. In practice, however, the handbook tables are used to interpolate the resistances for member lengths between 0 and 500 mm. The interpolation is only accurate when Cr is tabulated consistently using the formula in S16 for all member lengths. For double angles, there is no minimum value of KL/r, and these members can reach Cro when KL = 0

 

 

Questions on various aspects of design and construction of steel buildings and bridges are welcome. They may be submitted via email to info@cisc-icca.ca. CISC receives and attends to a large volume of inquiries; only a selected few are published in this column.