Predicting post-fire mechanical properties of grade 8.8 and 10.9 steel bolts: Fid-Bau Portal

Predicting post-fire mechanical properties of grade 8.8 and 10.9 steel bolts

Ketabdari, Hesamoddin / Saedi Daryan, Amir / Hassani, Nemat

Abstract Structural fire safety is one of the primary considerations in the design of high-rise buildings where steel is often a popular material for structural members selection. Therefore, predicting post-fire mechanical properties of steel bolts as a crucial element in steel structures is highly valuable. In this paper, the behavior of the High-Strength Steel Bolts (HSSB), after exposing to fire, is investigated and practical equations for the mechanical properties, including the ultimate strength, the yield strength, and the modulus of elasticity, are proposed as well. Accordingly, Grade 8.8 and Grade 10.9 steel bolts are employed in a variety of sizes, from M6 to M24, experiencing six different target temperatures. After natural cooling, a tensile test is applied to all the bolts, the corresponding stress-strain curves are derived, and finally all required data for each specimen are obtained by means of these curves. Results from these curves indicate that at 400 °C or fewer temperatures, more than 80% of the mechanical properties are recovered. Between 400 °C and 500 °C, the features began to reduce, however, above 500 °C, a sudden drop was noticeable. Besides, by using both the obtained data and the Gene Expression Programming (GEP) as a branch of Genetic Algorithm, equations for mechanical properties of HSSB are derived. The validation results indicate that the relative error of GEP-based models is less than 10%. All in all, the minimum error in the GEP-based models demonstrates favorable equations for post-fire mechanical properties of HSSB and also better predictions than the traditional models.

Highlights • The behavior of the High Strength Steel Bolts (HSSB) is investigated after exposure to fire. • The functional equations for the mechanical properties of HSSB are proposed by Gene Expression Programming. • The proposed GEP equations are reliable and applicable due to their minimum errors.