Critical Temperature Relationships for the Performance-Based Design of Realistically-Restrained Steel Composite Floor Assemblies

This dissertation presents an investigation into the structural-fire response of one-way spanning, realistically-restrained composite steel-concrete floor beams. This includes the experimental testing of two large-scale structural-furnace specimens with realistic shear tab framing connections to a support frame with finite stiffness as well as with terminal slab edge rotational restraint to replicate the continuous slab placement condition seen in a full building system. The tests were loaded to 65% of the specimen's factored design flexural capacity, representative of a full-service gravity load. The steel beams were protected with spray-applied fire resistive material (SFRM). Extensive instrumentation captured the specimen's thermal and structural response. The specimens were heated to both a standard fire curve as well as a natural fire curve which included a fire decay phase to burnout. The standard fire resistance was determined, and the specimen was observed to maintain structural integrity through burnout. The experimental research conducted in this study as well as results obtained from tests previously performed by others is used to validate simplified thermal and structural modeling techniques. A simplified, three-lumped mass heat transfer model was shown to provide good agreement with the measured steel temperatures. A basic, one-dimensional heat transfer fiber model was shown to provide an accurate slab temperature prediction capturing the effects of deck rib shielding. Temperature-dependent thermal properties (thermal conductivity and specific heat) were obtained experimentally for the lightweight concrete and SFRM material through the transient plane source technique using a HotDisk TPS2500s Thermal Constants Analyzer. Both complex and simplified structural modeling strategies were implemented and validated against the test findings. The use of the complex model compared against test data allowed for confidence in the simplified modeling approach proposed in the work. In addition to the realistic testing, a computational investigation into 16 standard fire tests of varying composite action, end restraint, and loading utilization is presented. The standard test results are compared to the results of 29 additional fire experiments in the literature for a wide range of test conditions. Steel temperatures at structural failure are noted as "critical temperatures", and are shown to be a clear function of flexural load utilization regardless of end condition, specimen length, beam depth, slab profile, fire protection, and ambient composite action. This crucial finding leads to the development of simplified performance-based design methods framed around the concept of maximum permissible steel temperature for a calculated level of loading. A final parametric study extends the validated models to long span cases and provides recommended performance-based critical temperature design solutions. SFRM thickness and thermal properties are also highlighted stochastically to demonstrate the variability associated with the application of minimum amounts of fire protection, most often leading to higher degrees of conservatism than accounted for in design.