Prediction of formation shear wave velocity using a new hybrid deep learning algorithm

Reservoir, drilling, geology, and petrophysics engineers rely heavily on formation shear wave velocity (FVs), yet traditional methods for obtaining FVs are often time-consuming and costly. Leveraging artificial intelligence (AI) offers the potential for more efficient and accurate predictions. In this study, a novel hybrid approach integrating a convolutional neural network (CNN) with a long short-term memory (LSTM) network was developed for FVs prediction, utilizing 8670 datasets from three wells in a Middle Eastern gas field. This method, the first of its kind applied to geophysical applications, effectively captures the temporal relationships between well logs and input/output data, significantly enhancing prediction accuracy. Pearson correlation analysis reveals negative correlations between FVs and measured depth (MD), caliper (cP), shallow resistivity (RES-SHT), density (RHOB), and formation pressure wave velocity (FVp). In contrast, positive correlations are observed with deep resistivity (RES-DEP), gamma ray (GR), porosity (NPHI), and medium resistivity (RES-MED). Elevated RES-DEP, cP, GR, RES-SHT, and FVs values are associated with higher FV estimates. The CNN-LSTM hybrid model outperforms other algorithms, maintaining high accuracy when applied to an additional well (V3), thereby demonstrating its robustness and effectiveness for shear wave velocity prediction. The dual architecture—combining CNN's spatial feature extraction capabilities with LSTM's ability to model temporal dependencies—underpins the model's stability, adaptability, and applicability to a wide range of geophysical studies.