ABSTRACT

ABSTRACT: This paper presents the work to investigate the feasibility of using AI-based acoustic emission technology to monitor bit-wear during drilling process. Field drilling experiments were carried out on two different rock formations, namely, Southwick Sandstone and Chicopee Shale. Regular earbide rotary drill bits with three different conditions (new, worn and bearing failure) were used, providing several well-definedrilling situations. Acoustic emission (AE) were monitored during the drilling and then characterized. Analysis on the AE signalshowed that significant differences in AE properties can be found among different drilling situations such as: new bit, worn bit, etc.; and that these differences may be identified by pattern recognition techniques.

INTRODUCTION

Drilling is an activity with many applications of strategic or societal importance. The fields using drilling as a key technology include exploration for and extraction of oil, gas, geothermal and mineral resources; environmental monitoring and remediation; underground excavation and infrastructure development; and scientific studies of the earth's surface. Efficient and effective drilling technologies have become a critical element in a robust and healthy economy. Therefore, improvements in fundamental technologies applicable to the drilling of rock will benefit the U.S. economy and strengthen the competitive position of the United States in the world wide drilling, excavation and eomminufion industries (Anon., 1994).

A research was conducted to investigate the feasibility of using artificial intelligence (AI) based acoustic emission (AE) technique for bit-ware monitoring during drilling. Field drilling experiments were conducted. Acoustic emission signals generated from the bit-rock interface were monitored during the experiments. Further analysis of the AE signals showed that AE characteristics are different in different drilling situations, and the differences may be found in either time domain through AE count, event number, duration, and other waveform parameters, or frequency domain through frequency distribution or component values, etc. (features). These features can be conveniendy extracted by using statistical pattern recognition techniques and in turn may be used to identify drilling situations such as the extent of bit wear, impending bit failure (damaged bearings) or formation changes.

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