Data-Driven, In Situ, Relative Sensor Calibration Based on Waveform Fitting Moment Tensor Inversion

Ismael Vera Rodriguez, Sergey Stanchits, Jeffrey Burghardt

Research output: Contribution to journalArticlepeer-review

22 Citations (Scopus)

Abstract

This article presents a deconvolution-based methodology for the relative calibration of acoustic emission sensors. The method uses the observations from multiple acoustic emission arrivals to estimate correction functions that account for instrument response and site effects. The corrections act on the measurements reshaping the wave arrivals, in many cases allowing visible discrimination between compressional and shear phases. The methodology is illustrated with the application on a dataset obtained during the hydraulic fracturing at triaxial stress conditions of a Colton sandstone block. A workflow for the estimation of full moment tensor solutions, including the generation of quality control attributes complements the presentation. The moment tensor solutions display fracture geometries mostly aligned with the main hydraulic fracture’s plane. The style of activation can be associated with an extensive stress regime before breakdown and to a compressional stress regime after breakdown, when the pumps were reversed to withdraw fluid from the block. The source types commonly deviate from the pure double-couple type; however, the deviations are under 30° in the case of opening fractures, and under 20° in the case of closing fractures. Therefore, the acoustic emissions in this experiment are the result of predominantly shearing dislocations.

Original languageEnglish
Pages (from-to)891-911
Number of pages21
JournalRock Mechanics and Rock Engineering
Volume50
Issue number4
DOIs
Publication statusPublished - 1 Apr 2017
Externally publishedYes

Keywords

  • Acoustic emission
  • Hydraulic fracturing
  • Moment tensor inversion
  • Sensor calibration
  • Triaxial experiment

Fingerprint

Dive into the research topics of 'Data-Driven, In Situ, Relative Sensor Calibration Based on Waveform Fitting Moment Tensor Inversion'. Together they form a unique fingerprint.

Cite this