Characterization of West-Africa Submarine Channel Reservoirs: A Neural Network Based Approach to Integration of Seismic Data

摘要

Advances in deep water drilling technology have cleared thernpath for new domains of hydrocarbon exploration along thernAtlantic margins. Such deep marine settings are generallyrncharacterized utilizing seismic data while in developmentrnphase. Yet, interpreting facies information from seismicrnamplitude data requires an extensive and tedious work of anrnexpert geologist/geophysicist. This process is not only costlyrnin terms of work hours but also invites human inconsistenciesrninto interpretations. This paper attempts to solve the abovernproblem by proposing a neural network based automationrnmethod that learns how to relate the seismic data to faciesrnobjects like channels. The neural net training is based on thernmanual/CAD interpretation of a small portion of the availablernseismic amplitude data by an expert geologist/geophysicist.rnThe trained neural net is shown to be able to automaticallyrndetect channel body features in the remaining portion of therndata. As the case study, the paper includes a West-Africarnsubmarine channel reservoir where the expert interprets arnseismic amplitude data set of a turbidite sequence and wherernthe neural network manages to estimate channel faciesrnmorphologies from seismic data.