The procedure of flow detection is done using both ambient noise measurement and the analysis of Stoneley waves. To illustrate the procedure, we use the data recorded at well C1. At each step of the procedure, we indicate the signal processing tools which are used.

For VSP acquisition, The seismic source is a light weight dropper and the borehole sensor is a hydrophone.  The sampling interval in depth is 2.5 m. Before each shot, the ambient noise has been recorded. The VSP are highly corrupted by Stoneley waves (tube waves). Conversions of down going P-wave in down and up going Stoneley waves have been observed at the level of the karstic bodies. This phenomenon occurs in highly permeable formations.

• play
  Step 1: Ambient noise and ambient noise factor.
  Assuming that flow circulation introduces some changes in ambient noise properties, the analysis of the ambient noise has been done in order to detect the presence of flows. For that purpose, the average and the variance of the amplitude spectrum of each noise trace have been computed. We have noticed that an ambient noise factor defined as the average to variance ratio increases significantly at the level of karstic bodies. The analysis of the ambient noise shows that the variations of the ambient noise factor are correlated with the level of conversion of P-wave in Stoneley waves.


• play
  Step 2: VSP with hydrophones.
  We can notice that the first arrival which is the down going P wave  is strongly attenuated at 60 m depth. At that depth, the P-wave is partly converted to a down going Stoneley wave which is reflected at the bottom of the well. The VSP data have been processed in order to extract the down-going and up-going Stoneley waves. The down going Stoneley mode can be extracted by a velocity filter, the velocity being the apparent velocity of the Stoneley wave. A narrow band wave number filter applied after velocity correction is equivalent to a velocity filter in the f, k domain.


• play
  Step 3: Amplitude of upgoing Stoneley waves.
  The Hilbert transform has been applied to the different wave fields in order to estimate their amplitude ( instantaneous envelope). Figure  shows the amplitude variation of the upgoing Stoneley waves. We can notice a phenomenon of conversion at a depth of 57m.


• play
  Step 4: P-wave to Stoneley wave conversion factor and VSP flow index.
  The instantaneous amplitudes of the upgoing converted Stoneley waves have been stacked in a small corridor located after the arrival time of the downgoing P-wave, in order to obtain a body-wave to Stoneley wave conversion factor which points out a karstic level at a depth of 57m.
  The attribute, named VSP flow index, defined as the product of the ambient noise factor by the body wave to Stoneley wave conversion factor has been used to detect both karstic bodies and flow.