Probabilistic mud slowness estimation from sonic array data

There are various ways to get an estimate of the mud slowness. The most common one consists in performing acoustic measurement at the surface using mud sample or using empirical laws linking mud slowness as a function of mud density and type. In this case, the estimation of mud slowness is said to be indirect as the estimated mud value is not the one measured in the well at the time the tool recorded the data. The second approach consists in analyzing dispersive modes contained in recorded data and from this study deriving an estimate of the mud slowness. The drawback of this non automatic method is that it relies heavily on the skill of the person that will analyze the dispersion plots. Therefore, there is a need for a general framework capable to provide reliable and accurate mud slowness estimation with minimum supervision. In order to overcome this need a probabilistic approach combining the high frequency monopole data and the output from the monopole radial profiling to get an estimate of the mud slowness has been developed. Robustness and efficiency of this algorithm will be illustrated on real field data.