In karstic levels, we observe usually a low velocity of the formation, a low amplitude of the acoustic signal and a low value of the correlation coefficient, as it can be seen in the two depth intervals between 82 and 88 meters and between 95 and 100 meters.

The SVD processing leads to compute a specific attribute used to detect karstic levels. The attribute, named Noise Signal detector, is the product of 3 normalized terms:

• an amplitude term. It is the low amplitude signal detector. Lower is the amplitude of the acoustic signal, higher is the amplitude term;
• a correlation term. Lower is the correlation coefficient between the wavelets, higher is the correlation term;
• a velocity term. Lower is the velocity, higher is the velocity term.

• play
  Step 1: Noise / Signal detector
  Figure shows from left to right:
  
  • the acoustic noise detector : product of the amplitude term by the correlation term;
  • the correlation coefficient log;
  • the Noise Signal detector: product of the acoustic noise detector by the velocity term;
  • the acoustic flow detector.
  Integration in depth of the Noise Signal detector from bottom to the top of the well has been done to mimic a flow measurement. The Noise/Signal detector highlights two karstic levels at depths of 82 - 85 m and of 93 - 100 m.


• play
  Step 2: Acoustic - Flow - VSP.
  Figure shows the comparison between acoustic data, flow measurement with a production logging tool and VSP. The acoustic flow detector shows two productive levels, but the flow measurement obtained using PLT only shows the upper layer. VSP shows that down going body waves give rise to up going Stoneley waves at the depth of the water productive layer.


• play
  Step 3: Acoustic - BHTV.
  The two karstic levels, detected by acoustic logging, are confirmed by Borehole Televiewer logs. The conclusions of the study show that acoustic logging can be fruitfully used to detect productive levels but cannot guarantee flow circulation.