In burst-prone mines, seismic monitoring is a central component of risk analysis and mitigation efforts. In order to maximize the effectiveness of such analyses, event magnitude and location errors should be minimized. However, minimizing location errors can be difficult due to unknowns in the geologic media through which the seismic energy propagates. The National Institute for Occupational Safety and Health (NIOSH) is conducting research to improve the quality of the event locations contained in a seismic catalog produced by a surface network monitoring a deep underground silver mine in northern Idaho. Travel times of several events with relatively well-known locations, as determined by an in-mine network, were modeled using the fast marching method. The Nelder-Mead simplex method was then employed, in conjunction with global Monte Carlo sampling, to determine a suitable 1-D layered P-velocity model. Using the new model with station delays on the surface network, epicentral location errors were reduced by approximately a factor of three and depth errors were reduced by more than a factor of ten. Significantly lower location errors can support higher-quality analyses to help improve safety in deep underground mines. Planned development of a 3-D model and additional instrument deployments will further increase location quality.
Improving a Deep Metal Mining-Induced-Seismicity Catalog Using Numerical Optimization
Chambers, D. J., and M. S. Boltz. "Improving a Deep Metal Mining-Induced-Seismicity Catalog Using Numerical Optimization." Paper presented at the 51st U.S. Rock Mechanics/Geomechanics Symposium, San Francisco, California, USA, June 2017.
Download citation file: