Numerical Simulation of Well Logs Based on Core Measurements: An Effective Method for Data Quality Control and Improved Petrophysical Interpretation

Abstract

Data quality of well logs and laboratory measurements is crucial for accurate petrophysical interpretations in formations with complex solid compositions, thin beds, and adverse geometrical conditions . In this paper, we introduce a new method to calibrate and verify the reliability of core data and well logs acquired in spatially complex rocks. The method is based on the numerical simulation of well logs to reproduce the effects of borehole environmental conditions and instrument physics on the measurements. Additionally, high-resolution (HR) core data combined with rock typing and multiwell measurement analysis techniques enable the construction of multilayer formation models. We document the successful application of the new core-well-log calibration method to two wells penetrating a clastic formation in the North Sea. While the numerically simulated well logs match the available borehole measurements in the first well, large measurement discrepancies were observed in the second well. Normalization of nuclear logs in the second well based on core data and numerically simulated well logs improved the assessment of bulk density and neutron porosity by 5% and 20%, respectively, while unnormalized nuclear logs overestimated formation porosity. Multiwell comparisons of well logs also confirmed that measurement accuracy was compromised. The problem with data quality was attributed to a probable inadequate tool calibration, although the log header did not indicate any notable issues. Additionally, numerical simulations of nuclear magnetic resonance (NMR) porosity logs indicated a prominent depth mismatch among well logs. The numerical simulation of well logs based on HR core data enables the detection of inconsistent, noisy, and inaccurate measurements, including cases of abnormal borehole environmental corrections causing biases in petrophysical interpretations.