An analytical approach to compute restrained shrinkage effects on composite concrete deck – steel truss members is presented and validated using strains and deflections obtained by Brattland and Kennedy, who in 1986 tested two full-scale (11.8 m × 2.35 m in plan) composite trusses and companion drying specimens for approximately 90 d. The test-to-predicted ratios for these quantities are excellent, with a mean value of 0.89 and a coefficient of variation of 0.11. It is also demonstrated that (1) single test specimens with a single drying face exhibit a large gradient of free shrinkage strain across the depth of the specimen that must be properly considered in the analysis; (2) shrinkage parameters estimated using generic rules based on the concrete mixture composition are not realistic for the particular mix used by Brattland and Kennedy; and (3) history-independent analysis methods, as described in Annex H of CSA S16-09, do not accurately predict the slab strains and truss deflections observed by Brattland and Kennedy unless the variation of the free shrinkage strain with slab depth, tensile creep, and concrete modulus of elasticity are properly accounted for.
Read full abstract