Effects of Jointing on Fragmentation : Design and Influence of Joints in Small Scale Testing

Abstract: This thesis has been conducted for the Luleå University of Technology (LTU), Sweden as part of an inter-university collaboration between LTU and the Montanuniversität Leoben (MUL), Austria.   The project has included three master thesis works, all investigating some aspects of the effects of jointing on fragmentation through small scale tests.   The topic of this thesis has been to develop a method for representing/introducing artificial joint planes into concrete blocks and subsequently manufacture a number of specimens with different joint sets for testing, using block dimensions previously used at MUL. The method for manufacturing the jointed test specimens, developed by the author at LTU, has been evaluated through blasting and fragmentation analysis. Comparison of the different produced joint sets has been done to assess the influence of jointing on fragmentation.   A total of 10 magnetic concrete blocks were produced and evaluated. The 10 blocks had an average density of 2485 ± 41 kg/m3 (mean ± standard deviation). Eight of the 10 blocks contained joint sets (JS) with different characteristics, while two blocks were un-jointed reference blocks. A total of four different joint sets were developed.   The four joint sets had the following properties; JS1: joint spacing 95 mm, strike/dip equal to 0/90°, JS2: joint spacing 47.5 mm, strike/dip equal to 0/90°, JS3: joint spacing 47.5 mm, strike/dip equal to 60°/90° and JS4: joint spacing 89 mm, strike/dip equal to 0/70°.   The produced joints have proven to behave as desired, i.e. being able to reflect incident waves and to stop some incident cracks.   It has been found that there is a strong indication of jointed blocks producing a finer median fragmentation size (x50) than that of the reference blocks. This means that by introducing joint sets into the test specimens the degree of fragmentation by blasting has been increased. This was predicted by the Kuz-Ram model.   Each of the 28 blasted rows has been sieved and analysed. All the data has been fitted to the three parameter Swebrec function, producing an average coefficient of determination (an R-square value) of 0.9946 ± 0.0064.