ABSTRACT:

Typical open-pit seismic monitoring applications attempt to assess the slip potential behind the rock face during a wall retreat. They employ sub-surface sensor arrays several hundreds of meters in size, localized within the respective wall, for which event locations can be obtained using a homogeneous wave velocity model. More recently, seismic technology is asked to provide a characterization of the seismicity associated with the entire pit. In case of a pit-wide seismic sensor array several kilometers wide, a reliable analysis requires that the mine geometry and the presence of geological strata be accounted for. The shortest 'visible' ray-path technique, originally employed in computer graphics, allows for the use of a homogeneous velocity model with appropriate corrections, thus widening the effective monitoring area and improving accuracy and reliability of event locations. The Fast Marching Method is proposed to resolve event locations using an arbitrary 3D velocity model derived based on the mine geometry and structural geology information. Interestingly, this technique provides a general framework to account for excavations or caves in locating seismicity occurred in underground mining applications.

1. INTRODUCTION

The most important task in monitoring mine seismicity is to provide fast, precise and robust location of seismic events. Event location consists of two steps, the forward solution - which calculates the expected propagation time of a particular seismic wave for a given sourcereceiver position, and the inversion - which minimizes the difference between the expected and observed data. The data are represented by first arrival times of P and S waves at different sensors. Wave propagation depends on the properties of the medium crossed, commonly represented as a simplified wave velocity model. There are two basic approaches for travel time estimation: raytracing and wave-front tracking. The former method uses high frequency ray approximation of the wave equation solution [1], while the latter employs various techniques to obtain a numerical solution of the eikonal equation [2, 3, 4]. In a mining environment, geological structures, the presence of naturally occurring cavities, as well as manmade excavations can be highly complex and require a general 3D velocity model that incorporates layers, blocks and voids, with large wave velocity gradients. With the increasing complexity in the velocity model, the evaluation of the forward solution will be more computationally time consuming, regardless of the specific algorithm employed. Worth noting, given the frequency bandwidth of the passive seismic monitoring, the choice of a homogeneous velocity model can be justified from a seismological point of view in case of underground applications, and is working reasonably well in practice. Based on this simplification, the forward solution is obtained at a minimum computational cost, which is essential for real-time monitoring. In case of open-pit applications, however, the geometry of the free surface, which obviously is dictated by the actual stage of the mining operation, makes the attracted homogeneous velocity model assumption justifiable only for restricted volumes, which allow for source-sensor visibility. This imposes limitations on the overall efficiency of open-pit passive seismic monitoring applications.

This content is only available via PDF.
You can access this article if you purchase or spend a download.