As the global demand for energy and natural resources continues to increase so does our interaction with Earth's near surface through resource extraction and waste injection. In monitoring these interaction, seismology plays a central role. The focus of this work is on improving the detection and localization of seismic sources, a fundamental problem in seismology.
After discussing the strengths and limitations of existing methods for source detection and localization, I develop a solution based on a beamforming approach that uses cross-correlation functions in a maximum likelihood search for sources of seismic energy. I call this method InterLoc, short for `interferometric locator', and apply it to data recorded at two active underground mines to demonstrate its effectiveness in monitoring both impulsive sources and persistent sources. Next, I demonstrate how persistent seismic sources, typically seen as contaminants, can be used directly to measure small changes in the medium between a source and either source-station pairs. This method relies on the ability to locate and monitor source activity and then use this information to identify and select cross-correlation functions to isolate each source of interest. From the resulting cross-correlations, it is possible to measure small temporal changes in the waveforms. To demonstrate this method, I show how ore-crushers can be used to track the growth of a block cave by measuring changes in traveltimes due to ray paths having to circumvent the growing cave.
In the final chapter I focus on the development of a processing framework for the detection and location of microseismic events recorded on dense (or large-N) surface arrays. The proposed framework involves: (1) data reduction; (2) dividing the array into smaller sub-arrays; (3) waveform processing within fixed time windows; (4) stacking of time windows selected based on each potential origin time and source location; and (6) combining the output from all sub-arrays to infer detections and locations of sources. This methodology is validated with synthetic data built to emulate a real dataset from a 10,050 node survey to evaluate the suitability of land for carbon sequestration. Based on the presence of very strong coherent contaminating sources and low rock quality, I am only able to detect sources with moment magnitude greater than -0.5. In the five hours of data processed there is no positive detections suggesting this could be a good site for carbon storage. More work is needed to improve the detection threshold and quantify risk based on event location and magnitude.
In summary, my work demonstrates how the interference (via cross-correlation) and stacking of seismic waveforms can be combined in different ways to create effective solutions for problems faced by today's industries.
| Identifer | oai:union.ndltd.org:uottawa.ca/oai:ruor.uottawa.ca:10393/38303 |
| Date | 19 October 2018 |
| Creators | Dales, Philippe |
| Contributors | Audet, Pascal |
| Publisher | Université d'Ottawa / University of Ottawa |
| Source Sets | Université d’Ottawa |
| Language | English |
| Detected Language | English |
| Type | Thesis |
| Format | application/pdf |
Page generated in 0.0017 seconds