Published Online: 19 March 2010
Page Count: 5
Bennett, Richard M.
The Univ. of Tennessee-Knoxville, Knoxville, TN
Sanders, John P.
Clemson Univ., Anderson, SC
General Shale Products Corp., Johnson City, TN
(Received 25 August 2009; accepted 26 January 2010)
Extensive bond wrench testing of clay brick with different void areas has recently been performed. The data set is very comprehensive, testing different types of brick, brick with low and high initial rate of absorption, different mortar types, and voids ranging from 22 to 36 %. All pairwise combinations within each data set were compared with the U-test. The U-test tests for statistically significant difference between means of two data sets. The U-test is distribution-free, is nearly as statistically efficient as the t-test, can handle different sample sizes, and is robust against different variances. Approximately one-half of the comparisons showed no statistical difference in bond strength with different void areas. Of those that showed statistically significant difference, an essentially equal number showed higher bond strength with lower void area as those pairs that showed lower bond strength with lower void area. From the statistical testing, it is clear that flexural tensile strength is not a function of brick void percentage for the range of voids tested.
Paper ID: JAI102707