Lap splices are used in reinforced concrete structures to transfer tension forces across discontinuous reinforcing bars to allow for continuity of load path in structural elements. Lap splices of straight reinforcing bars present a number of disadvantages when used in connections of large precast concrete elements typical of bridge substructure. Most importantly, lap splices of large (e.g., No. 11) straight bars are substantially long. Since the closure joint connecting two precast elements must be at least long enough to fit the lap splice, traditional lap splices result in impractically large closure joints, offsetting the benefits of using precast concrete elements. To address this problem, bridge designers are using hooked bars in noncontact lap splices to connect precast elements, presuming that hooked bars will allow for shorter required splice lengths. However, there exists neither substantial design guidance nor studies of the behavior of hooked bar lap splices in large precast elements justifying this design philosophy. To develop design guidance permitting the use of noncontact hooked bar lap splices and address the knowledge gap regarding the behavior of such splices, an extensive experimental and computational research program was conducted which is described in this dissertation. Fifty-eight large-scale beam-splice specimens containing hooked bar lap splices were tested to physically study the behavior of hooked bar lap splices and develop a dataset to justify design guidance permitting the use of such splices in practice. Bond variables were parametrically varied among the test specimens to produce guidance applicable over the wide range of geometric configurations and material properties expected in bridge design. The specimens were subjected to monotonic, four-point loading and were designed to fail in a mode related to anchorage to study splice behavior. Nonlinear finite element analyses were conducted to examine the mechanism of force transfer in hooked bar lap splices and numerically assess splice configurations not experimentally studied. A simple approach to modelling hooked reinforcing bars in solid concrete elements which accounts for conditions of imperfect bond was developed and validated using the experimental results. Test results from the 58 specimens were used to assess the appropriateness of using existing guidance for hooked bar anchorages to design hooked bar lap splices. Because the existing guidance was found to be deficient for this application, descriptive and design equations characterizing hooked bar lap splices were developed using power regression analyses. The results demonstrated that all else being equal, a bottom-cast hooked bar lap splice can develop approximately 40% greater stress than contact lap splices of straight bars. Accordingly, hooked bars can be used to splice bars over a substantially shorter length than straight bars. Noncontact hooked bar lap splices without secondary reinforcement (e.g., ties) can fail due to a mode termed "hook side bulging", resulting from eccentricity between the lapped bars. Splices with secondary reinforcement typically fail due to more typical modes observed in the literature, such as side-face blowout and concrete crushing. Unlike as suggested by code authorities and some researchers for noncontact lap splices of straight bars, noncontact hooked bar lap splices were found to exhibit weaker splice strengths than contact splices as the splice spacing increased. The use of steel fibers and increases in lap length, concrete compressive strength, cover depth, amount of secondary reinforcement, or the number of lap splices allowed for greater stress on average to be developed in spliced bars. All else being equal, an increase in either bar size or the number of spliced reinforcement layers decreased the stress that could be developed in the spliced bars. A descriptive equation characterizing splice strength with an average test-to-calculation ratio and coefficient of variation of 6% was developed. The descriptive equation was adapted to develop a design equation for the minimum required lap length of hooked bars which uniformly characterizes the influence of the bond variables over the ranges explored in this study. Design examples and code language facilitating technology transfer of the design equation into immediate practice were developed. / Doctor of Philosophy / Precast concrete is widely used in highway bridges to enable more rapid and economical construction than could be achieved using cast-in-place concrete. However, the connection of two or more precast, prefabricated bridge elements introduces several difficulties which may inhibit construction, thereby reducing overall economy. One of the most significant difficulties is that connections of substructure elements supporting the superstructures are impractically long using a common, code-approved detail―lap splices of straight reinforcing bars. Such splices are quite long (e.g., 5 ft in length) since large bars are typically used in substructure elements, requiring long splice lengths to transfer the large forces in each bar across the connection. Details which would shorten the required splice length would consequently reduce the required connection length, thereby reducing the amount of cast-in-place concrete construction required in the field. Consequently, the speed of construction, economy, and worker safety would increase. This dissertation thus summarizes an extensive experimental and numerical study aimed at validating the use of noncontact hooked bar lap splices to shorten the required splice length of large precast elements. In support of this objective, the anchorage behavior of noncontact hooked bar lap splices was studied through static load testing of 58 large-scale beam-splice specimens and nonlinear finite element models accounting for bond-slip behavior. These efforts revealed that hooked bar lap splices can develop on average approximately 40% more stress over the same lap length than contact splices of straight bars. Existing design provisions which might presently be used to design hooked bar lap splices were evaluated against the experimental results and were found to be deficient in characterizing splice strength. Thus, a design equation was developed for the splice length of hooked bars which accurately characterizes anchorage behavior and allows for significantly shorter splices lengths than what could be achieved with straight bars.

The tensile resistance of No. 15 lap spliced reinforcing bars with varying transverse spacing and lap splice length was evaluated in full-scale concrete block wall splice specimens. The range of the transverse spacing between bars was limited to that which allowed the bars to remain within the same cell, and included the evaluation of tied spliced bars in contact. Two-and-a-half block wide by three course tall double pullout specimens reinforced with contact lap splices were initially used to determine the range of lap splice length values to be tested in the wall splice specimens such that bond failure of the reinforcement occurred. The double pullout specimens were tested in direct tension with six replicates per arrangement. Three values of lap splice length: 150, 200, and 250 mm, were selected from the testing of the double pullout specimens and tested in the wall splice specimens in combination with three values of transverse spacing: 0, 25, and 50 mm, with three replicates per configuration. A total of twenty-seven two-and-a-half block wide by thirteen course tall wall splice specimens reinforced with two lap splices were tested in four-point loading. Both the double pullout and the wall splice specimens were constructed in running bond with all cells fully grouted. The tensile resistance of the lap spliced bars in the double pullout specimens was measured directly. The contact lap splices with a 150, 200, and 250 mm lap splice length developed approximately 38, 35 and 29% of the theoretical yield load of the reinforcement, respectively. The difference between the mean tensile resistances of the three reinforcement configurations tested in the double pullout specimens was found to be statistically significant at the 95% confidence level. Different than expected, the tensile resistance of the lap spliced reinforcing bars in the double pullout specimens was inversely proportional to the lap splice length provided. For the short lap splice lengths used in this investigation, the linear but not proportional relationship between bond force and lap splice length known from reinforced concrete is believed to have caused this phenomenon. An iterative sectional analysis using moment-curvature response was used to calculate the tensile resistance of the lap spliced reinforcement in the wall splice specimens. The calculated mean tensile resistance of the reinforcement increased with increasing lap splice length, and was greater when the bars were in contact. Securing the bars in contact may have influenced the tensile capacity of the contact lap splices as higher stresses are likely to develop as a result of the bar ribs riding over each other with increasing slip. Results of the data analysis suggest that the tensile resistance of non-contact lap splices within the same cell is generally independent of the spacing between the bars. A comparison of the experimental results for the wall splice specimens with the development and splice length provisions in CSA S304.1-04 and TMS 402-11 indicate that both the Canadian and U.S. design standards are appropriate for both contact and non-contact lap splices located within the same cell given the limited test database included in this investigation.