Abstract
Based on the previous experimental results by authors, the empirical equations for splitting bond strength had been developed in the Part 1. In this paper, these equations were extended to a more generalized form and verified from the applications to many test data of other investigators. For the comparative study, the proposals developed by Orangun, Jirsa and Breen and by Jimenez, White and Gergely were introduced and also applied to the same test data. The available test data for development length include 354 specimens in total, which failed in splitting of surrounding concrete before bar yielding. From this comparative study, the authors' proposal seems to give the best prediction for bond strength. The important characteristics of splitting bond failure which were taken into account in the authors' equation are summarized as follows : 1) The bond stress at which the surrounding concrete is split to failure depends sensitively on the geometrical bar location in a cross section. The authors' proposal accounts for this fact through classifying the failure modes into three typical splitting patterns; side-to side splitting, corner splitting and V-notch splitting. 2) The efficiency of stirrups in improving bond strength depends also on the geometrical arrangement of longitudinal bars and stirrups. It should be emphasized that yielding strength of stirrups higher than 3000 kgf/cm^2 gives no influences on the bond strength. The authors' equation reflects these experimental evidences reasonably by introducing the coefficient k which depends on the splitting patterns. The proposed equation was applied also to previously reported lap splice test data including 151 specimens, and the good correlation between calculated and measured bond strength was obtained.
