In this paper, we propose an innovative use of MWD and Sonic logging data for better estimation of UCS. Sonic logging data and real-time ROP measurements are integrated into a Bayesian inference framework, employing empirical sonic correlations and an ROP model. Combining the two types of data sets under a single inference scheme can help reduce the uncertainty in UCS estimation and in prediction of ROP.
The prior mean of the parameters employed in the sonic correlations is interpreted from sonic logging in offset or analog well(s). The UCS estimation for the target well is inferred by application of the least square minimization approach to yield approximations for the maximum a posteriori (MAP) estimation. Verification of the inference procedure is demonstrated on actual MWD data and Sonic datasets. The results show that the UCS generated from the empirical correlation of sonic travel time slowness still contains large uncertainty. Joint inversions of ROP data and sonic velocity measurement reduce the uncertainty in the UCS estimation. The results indicate that the calibration of ROP data is critical before use of empirical correlations to estimate UCS. This paper also addresses the possibility of deriving the pore pressure and UCS simultaneously with quality MWD and sonic data.