Mapping Cretaceous faults using a convolutional neural network – A field example from the Danish North Sea

The mapping of faults provides essential information on many aspects seismic exploration, characterisation reservoirs for compartmentalisation and cap-rock integrity. However, manual interpretation from data is time-consuming challenging due to limited resolution noise. In this study, we apply a convolutional neural network trained synthetic with planar fault shapes improve in the Lower Upper Cretaceous sections Valdemar Field Danish North Sea. Our objective evaluate performance model post-stack Field. Comparison variance ant-tracking attributes shows that predicts more details may overall geological tectonic understanding study area add potential was previously overlooked. sensitive noise, which can distort predictions. Therefore, proposed should be treated as an additional tool. Nonetheless, method represents state-of-the-art tool useful hydrocarbon exploration CO2 storage site evaluations.