FATIGUE STRENGTH AND STIFFNESS OF REINFORCED CONCRETE BRIDGE DECKS. FINAL REPORT

Tests were conducted under static, stationary pulsating and moving constant wheel-load on 1/3- and 1/6.6-scale models of an 8.5-in. thick non-composite concrete-on-steel girder deck for a simply supported 50-ft long bridge. Three deck designs are studied: (a) AASHTO design; (b) Ontario design (OHBDC); and (c) isotropic. The measured static deck ultimate strength (P sub u) is up to 5 times the Johansen load (safety factor = 8 divided by 21). The ACI and CEB Codes considerably underestimate the deck ultimate capacity, especially for the central deck region. Most of the decks eventually punched through at a maximum deck deflection less that 50% of the deck thickness. Punching shear deck failure appears to be related with the snap-through instability of an arching mechanism in the deck. Under moving wheel-load, approaching deck failure the initial two-way deck slab action changes to a one-way slab action. Under stationary pulsating load, the two-way deck slab action is maintained up to deck failure. Decks subjected to stationary pulsating load exhibited flexural radial cracking, while under moving constant wheel-load a grid-like cracking pattern formed similar to the bottom flexural steel mesh. Based on exponential curve fitting of the deck fatigue data, while the predicted 2.5 x 10 to the 6th power load cycle deck fatigue strength under stationary pulsating load is 0.47 P sub u divided by 0.54 P sub u (safety factor = 6 divided by 12), that under moving load (2.5 x 10 to the 6th power wheel-load passages) is considerably lower at 0.21 P sub u divided by 0.28 P sub u (safety factor = 2 divided by 4). The 2.5 x 10 to the 6th power wheel-load passage deck fatigue strength (moving load) is comparable to the deck's cracking load level, while the 2.5 x 10 to the 6th power load cycle deck fatigue strength (pulsating load) is similar to the yielding load level in the bottom steel. If the effectiveness of the deck design is determined by the number of wheel-load passages the deck can carry at a given load ratio P/P sub u (P = moving load) without deck failure, the OHBDC deck design appears to be more effective than the AASHTO design.