next up previous [pdf]

Next: Elastic modeling Up: De Ridder and Haines: Previous: De Ridder and Haines:

Introduction

Ongoing research at a site outside Durango, Colorado, is focused on characterizing and modeling a coal bed that has been burning underground for a number of years. The goal is to develop an understanding of how coal fires burn for the purpose of designing optimum solutions for extinguishing coal fires around the world. The unique aspect of this particular fire is that it is small-scale and easily accessible. We were asked whether seismic methods could potentially provide useful information at the site, ideally as a way to map the burned and unburned coal. Any related information would be useful, such as mapping only the unburned coal, and/or locating fissures and other subsidence features that might not reach the surface.

We conducted elastic modeling to simulate a P-wave seismic survey at the site, to assess the likelihood of successfully applying seismic methods to this problem, and to determine optimal parameters for any field data acquisition. The model is highly simplified given that it neglects the fissures and other overburden heterogeneity. The main target is the partially burned coal layer.

The site is roughly 300m by 100m. Researchers of the Global Climate & Energy Project at Stanford University (GCEP, 2008) have been studying the site using a number of different approaches to understand the progression of the fire over time and try to predict its likely future trends. One approach has been the mapping of surface anomalies across the site - mainly fissures resulting from the compaction of the ash as the coal burns. Researchers have also drilled a total of 14 wells across the site, in areas where they know the coal has burned, where they think the burn front is located, and in unburned areas. A 100 ft core was recovered from one of these wells. Five wells have geophysical logs (caliper, gamma and density), and rock cutting samples were taken at 5-ft intervals in all the wells.

The coal bed is approximately 5 m thick where it is unburned and closer to 2-4 m thick where it has burned. It outcrops northwest of the site and dips to the southeast. In the area of the site itself, the coal bed ranges from approximately 10 m depth to approximately 30 m. Wells showing ash are predominantly in the northwest (up-dip) end of the site, with unburned coal located down-dip. The coal bed is embedded within a sandstone formation that also includes some shale layers. A thin unconsolidated layer composed of rocky soil up to approximately 1.5 m thick overlies parts of the site.

We present specifics of our elastic model, and then present the resulting synthetic data. We examine the data in various domains, and show a brute stack of the data set. Finally, we discuss our findings in relation to the original question, and consider the impacts of neglected aspects of the modeling.


next up previous [pdf]

Next: Elastic modeling Up: De Ridder and Haines: Previous: De Ridder and Haines:

2009-04-13