Artificial Intelligence-Assisted Seismic Mapping of Mediterranean Complex Turbidite Reservoirs Depositional Architectures for Effective Gas Exploration and Development

Abstract Accurate mapping of internal depositional architectural elements of a Pliocene-aged gas-bearing turbidite reservoir of the Mediterranean Basin into discrete 3D geobodies has been achieved through applying innovative workflow assisted by the convolutional neural network. The mapped reservoir depositional architectures have been integrated to the acoustic and elastic properties inverted form 3D seismic data to build robust multi-realizations reservoir static models. These models have been used to optimize the appraisal and development well locations and accurately assess the gas initial in-place of the discovery. Four wells have been sited utilizing the results of the constructed models. The wells have been successfully drilled and added 175 Million standard cubic feet in a day (MMscf/d) which obviously improve the commercial value of the project. The workflow is a major step in accurate delineation of the internal depositional architectural elements of the deep-water turbidite reservoirs of the Mediterranean Basins, as well as in other locations/basins where similar settings exist. By applying the workflow, the subsurface complexities were revealed with the artificial intelligence algorithms, uncertainties were captured, risks were reduced and project commercial value was uplifted.