SIGNAL PROCESSING FOR IN-SEAM SEISMIC BASED VOID DETECTION TECHNIQUE

Open Access
Author:
Wang, Hongliang Henry
Graduate Program:
Mining Engineering
Degree:
Doctor of Philosophy
Document Type:
Dissertation
Date of Defense:
August 15, 2007
Committee Members:
  • Maochen Ge, Committee Chair
  • Margaret Hardy, Committee Member
  • Vladislav Kecojevic, Committee Member
  • Derek Elsworth, Committee Member
  • Sridhar Anandakrishnan, Committee Member
Keywords:
  • Signal Processing
  • In-Seam Seismic
  • Void Detection
  • Wavelet Transform
Abstract:
ABSTRACT Underground coal mining is often carried out in the vicinity of abandoned mine voids filled with hazardous air or water. The aim of this research was to study the problem of underground coal mine void detection by using the in-seam seismic (ISS) technique. This technique refers to methods that utilize artificially generated channel waves trapped in coal seams to locate geologic disturbances and mine voids. ISS is one of the basic geophysical methods for underground survey. The advantage of ISS is that seismic energy in the form of channel waves is better preserved, and therefore seismic waves can be detected over much larger distances compared with those radiating three-dimensionally. There are many technical problems and theoretical challenges related to the development of the ISS-based void detection technique, especially the acquisition of high quality seismic signals and the interpretation of these signals. First, there are practical issues related to data acquisition in underground coal mines. One has to face the problems of field tests with limited space and irregular source-receiver geometry. Second, there are challenges with data processing because of the underground environment and complex seismic signals. It was necessary to reduce the ambiguity conventionally associated with routine geophysics methods and to find an efficient data processing and void mapping method that is suitable for underground surveys. An important outcome of this research is the investigation of a set of comprehensive data analysis approaches that are suitable for ISS signals processing. The main contribution of this research is the development of wavelet analysis based signal processing methods. The wavelet transform, which studies a signal in both the time domain and the frequency domain, was successfully applied to time-frequency analysis, dispersion analysis, and decomposition and reconstruction of channel waves. Wavelet transform, a unique approach introduced in this thesis for the ISS technique, proved to be an efficient tool for signal analysis, including detecting newly merged signals and enhancing reflection signals. ISS basically uses reflection surveys to detect mine voids, and such reflection signals are much more difficult to identify than transmission signals. Using the Fourier transform, wavelet transform, and other tools, such as the gain controlled stacking method, the reflection signals were enhanced and detected with clear arrivals. Void mapping was realized by a simple and robust method called the elliptical mapping method which is inherently compatible with the complex underground mine conditions and supports simultaneous data processing. Geophysics methods have been widely used for underground surveys. Among several geophysics approaches for mine void detection, ISS has several distinctive advantages, including high resolution and reliability, long survey distance, and low cost. Thus, the ISS technique and associated signal analysis methods developed in this research appear to provide a very promising technique for mine void detection.