Direct multi-modal inversion of geophysical logs using deep learning
This work addresses real-time decision-making under geological uncertainty for oil and gas drilling, representing an incremental improvement with a novel method for a known bottleneck.
The paper tackles the non-unique inverse problem of geophysical log interpretation for geosteering by proposing a deep learning approach using a mixture density neural network with a novel loss function to output multiple likely stratigraphic solutions and their probabilities in milliseconds, achieving more accurate and realistic results than deterministic methods.
Geosteering of wells requires fast interpretation of geophysical logs, which is a non-unique inverse problem. Current work presents a proof-of-concept approach to multi-modal probabilistic inversion of logs using a single evaluation of an artificial deep neural network (DNN). A mixture density DNN (MDN) is trained using the "multiple-trajectory-prediction" (MTP) loss functions, which avoids mode collapse typical for traditional MDNs, and allows multi-modal prediction ahead of data. The proposed approach is verified on the real-time stratigraphic inversion of gamma-ray logs. The multi-modal predictor outputs several likely inverse solutions/predictions, providing more accurate and realistic solutions than a deterministic regression using a DNN. For these likely stratigraphic curves, the model simultaneously predicts their probabilities, which are implicitly learned from the training geological data. The stratigraphy predictions and their probabilities obtained in milliseconds from the MDN can enable better real-time decisions under geological uncertainties.