Field Test of a Low-frequency Sparker Source for Acoustic Waveform Logging

Low-frequency acoustic-energy sources for waveform logging have important applications in: 1) Verifying theoretical calculations; 2) generating tube waves in large-diameter boreholes; and 3) providing larger sample volumes in cases where borehole effects are important. A new low-frequency source was fabricated by modifying an existing acoustic-waveform logging system to discharge multiple capacitors in series with an automobile spark plug. The sparker source was tested in boreholes of 15and 8-centimeter diameter in homogeneous granite containing isolated fractures. The sparker source produced repeatable waveforms with frequencies centered on 5 kilohertz in the 8-centimeter-diameter borehole, and 7 kilohertz in the 15-centimeter-diameter borehole, compared to frequencies near 15 kilohertz for the same system using a lOW-frequency magnetostrictive source. The lower-frequency sparker source excited consistently measurable tube waves, in agreement with theory. Test results also confirmed that lower-source frequencies greatly decreased sensitivity to borehoie effects. Observed differences in frequency content and extent of shear-mode excitation in the two different diameter boreholes are probably related to differences in mode-excitation functions. The data confirm theoretical predictions that optimum shear-mode excitation occurs for source frequencies near normal-mode cutoff. Refiection of low-frequency tube waves appears to be an effective means for distinguishing between isolated open fractures and intervals containing extensive alteration around nearly impermeable fractures. INTRODUCTION Acoustic well logging has been an accepted method for measuring the primary porosity of sedimentary rocks in the petroleum industry for many years (Pickett, 1960). Compressional velocities determined from acoustic well logs also have become important in verifying velocity distributions used in processing seismic-reflection data (Narvarte, 1949). Recent acoustic wen-log research has focused on attaining greater .spatial resolution through sophisticated processing of acoustic-log data (Willis and Toksoz, 1983), and through the use of logging proves with as many as 12 receivers at variable offsets from the soUrce (Arditty and others, 1982). However, other important applications for acoustic well-log data do not demand such a precise degree of resolution, but do require etrective compressional velocities averaged throughout larger columns of rock. For example, the interpretation of seismic-refiection coefficients depends on the recognition of relatively large contrasts in acoustic impedance. The major source of uncertainty introduced