Listening to the story of the rock mass : The integration of conventional and unconventional data to understand rock mass behaviour at the Kiirunavaara Mine

Sammanfattning: The need to understand rock mass behaviour has never been greater; as mining globally progresses deeper to extract the precious resources humankind needs to maintain and forward our lives, we encounter rock behaviours associated with deep mining. These environments typically experience seismicity (including fault slip), rock bursting, strain bursting, spalling, aseismic movement along pre-existing discontinuities, and ground falls. These behaviours can have significant consequences to an operation’s safety and profitability. An understanding of the rock mass enables risk mitigating design.However, a number of hindrances exist along the path to gaining this understanding. Limited information in the literature is available on which data to collect, the most appropriate analysis techniques, and how one combines varied data sources into a mine-scale understanding of the rock mass. In essence, we do not yet understand how to listen to the story of the rock mass. This thesis explores these research topics and develops and extends techniques and methodologies to address these issues. Data from Luossavaara-Kiirunavaara AB’s (LKAB) Kiirunavaara Mine was used, however a strong focus was given to developing general methodologies that can be applied to many mining environments.The typical situation at most mine sites is one where many forms of data exist, a subset of which is conventional rock mechanics data. These data typically include, for example, laboratory testing for strength and stiffness of geological units, rock mass characterization from underground mapping, geological and geomechanical core logging, and positions and orientations of mapped discontinuities from underground mapping and/or oriented core logging. These data are often used for empirical design. It is common with particularly damaging seismic and/or ground fall events that additional data be acquired from underground damage mapping, but this information is rarely used on a larger scale. However, there is a subset of data that tends to be overlooked from a geomechanical perspective, even though the potential exists that this data can give us information about the characteristics of the rock mass. In the case of the Kiirunavaara Mine, this includes seismic data (including tomographic velocity structure), a database of information regarding ground falls, and laser imaging data.A variety of conventional and unconventional data from the Kiirunavaara Mine were analysed using existing, and extensions of existing, analysis techniques. A 3-D geomechanical model was developed, incorporating geological and geomechanical core logging data, underground mapping data, and laboratory testing data. The resulting mine-scale model showed some particularly interesting results, including 1) a very large variation in intact rock strength within each statistical grouping of geological units, combined with much overlap between the strengths of geological units, and 2) a 3-D model of volumes of clay alteration built on geological core logging data calibrated favourably to data of underground mapping of clay. Unconventional data sources were also analysed, in particular the use of behavioural data to infer characteristics of the rock mass. Spatial and temporal patterns of ground falls were evaluated, alongside spatial patterns in overbreak, which were identified using new, mine-scale analysis techniques of laser scanning data.Interestingly, correlations were apparent when comparing the analyses of distinct and separate data sources. Ground falls were concentrated in the intact rock between the clay volumes. The rock quality designation (RQD) had lower values in the same volume. Seismicity in the area was concentrated in the intact rock, with very few events in the clay volumes. 3-D velocity tomography models (developed by another researcher within this project) showed 1) no correlation with the RQD model (this result may highlight a limitation of scaling RQD data beyond it’s intended use), and 2) possible correlation with the clay volumes.A methodology was developed using numerical stress analysis to identify, at an early stage of investigation, if a volumetric feature is of significance to seismic rock mass behaviour. With so much surmounting evidence that the clay volumes influence rock mass behaviour at the Kiirunavaara Mine, this feature was well-suited to be tested. Results showed that the clay volumes: 1) influence the mine-scale stress field, 2) have the potential to extend the volume of the rock mass which is expected to experience crack initiation (one underlying cause of seismicity), and 3) result in a stress field that enables slip along many of the orientations of mapped discontinuities (another underlying cause of seismicity).Particularly valuable data were found to share the following characteristics: 3-D coverage, mine-scale, long-term, and multiple and distinct sources. It was important that the different data sources and their analyses led to the same conclusions even though they were separate and distinct. Unconventional data sources that were of particular use in this case include: spatial and temporal ground fall patterns and spatial patterns of seismicity. With further research, other unconventional data sources may show their value to rock mechanics analyses, such as spatial and temporal patterns of: seismicity, focal plane solutions from moment tensors, seismic stress inversion from moment tensors, overbreak, and possibly even tomography (the lack of strong correlation in this case may be specific to this environment and models used). Opening our minds to the possibilities, collaboration with experts in other fields, and a little creativity in our approach will help us in our quest to understand the stories of rock masses.