Development of a Methodology for Dynamic Testing of Rock Support : Field Tests and Numerical Analysis

The increasing mining depth in the Luossavaara-Kiirunavaara Aktiebolag (LKAB) mine located in the northern part of Sweden leads to higher stress magnitudes, resulting in increased seismic activity and more seismically-induced damage. The effectiveness of various ground support systems under dynamic loading conditions has therefore become of prime interest to LKAB for successful and safe mining at deep levels. The problems of mining-induced seismicity have necessitated the use of ground support systems which are capable of withstanding strong dynamic loads. Within this respect, a series of seven large-scale dynamic tests of rock support were conducted at LKAB Kiirunavaara mine. Explosives were detonated in boreholes in the pillar between two cross-cuts in order to generate dynamic loads on the rock support system installed on the cross-cut wall. This was done to evaluate the response of rock support subjected to strong dynamic loading. The tests included ground motion measurements, fracture investigation, ground and support motion imaging, as well as the deformation measurements. The design and the results of field Tests 1 to 7 and their associated numerical analysis are presented in this thesis. Field Tests 3 and 7 are excluded from detailed analysis in this study since the burden condition in these tests were different from the rest of the tests. The performance and energy absorption by the support system comprising Swellex Mn 24, 100 mm fibre reinforced shotcrete (40 kg/m3 steel fibre) and 75 mm * 75 mm weld mesh with 5.5 mm diameter is estimated and compared to the large-scale and laboratory test results conducted in different countries. The field tests results indicated that the relation between the burden and the used amount of explosive material and number of blastholes has a vital role in either reducing or involving the effect of detonation gases in test results. Results from field Tests 1, 2, 4 and 5 (considered in this study) showed that in Tests 1 and 2 minimum damage was created on the surface of the wall despite high PPVs were measured on the test wall. In Tests 4 and 5 complete damage of the burden was occurred despite the burden was in the range of Tests 1 and 2. Results from Tests 1 – 5 were used to design an improved burden in Test 6. The evidence of the damage to the test wall in Test 6 showed that the design in Test 6 was successful in terms of minimising the effect of gas in the burden and generating a sub-planar stress wave. The large amount of data recorded during these tests was used to develop the numerical models, and to study the field tests in more detail. The aim was to study the behaviour and response of the burden, free surface of the opening, and the installed rock support to the dynamic loading generated by a nearby blast numerically. A combination of two numerical methods, including the Finite Element code LS-DYNA and the Discrete Element code UDEC was used in the numerical analyses. The LS-DYNA model was used to calculate the dynamic input for the UDEC analysis. This was done by identifying the crushed zone boundary (CZB) developed around the blasthole. The velocity at the CZB (at the moment when it was fully developed) was applied as an internal boundary condition in the UDEC model. The geological conditions of the test site and the installed rock support were modelled in UDEC. The peak particle velocities, the velocity – time graphs, the maximum displacement generated on the surface of the wall, and the ejected thickness of the rock from the wall computed in the UDEC models were studied, and compared to the data obtained from field tests. This was done to identify the models that well represent the field tests behaviour. The identified models were used to study the failure mechanism in the burden and the rock support response to dynamic loading. Results indicated that, the numerical models were able to mimic the behaviour of the jointed rock mass and the rock support fairly well. The difference in behaviour between the numerical models and the field Tests 1 – 5, appeared to be caused by the gas expansion in the field tests, especially Tests 4 and 5. The numerical analysis results confirmed that the design of the burden in Test 6 limited the development of tensile yielded zones and as a result, complete failure of the burden was avoided. Furthermore, the ejection experienced from the wall was similar to that in actual seismic event. Numerical analysis results indicated that the geological structure of the burden and in particular the formed wedges near the opening have a significant role on the damage developed on the surface of the test wall. The presented testing method has the potential to be used in all similar large-scale testing of the rock support. Performing numerical analysis can significantly save time and energy while searching for an optimum design of the burden. By using a calibrated numerical model, it is possible to test the performance of different support systems using the same boundary conditions, in future studies