Prediction of Hydrocarbon Using Gaussian Process for Seabed Logging Application

Seabed Logging (SBL) is a technique that utilizes electromagnetic waves to propagate signals underneath seabed to determine the differences in resistivity levels in order to determine possible oil wells for exploration. This research investigates the potential of a Gaussian process approach to identify the presence of potential hydrocarbon in the deep water environment. Simulations were conducted using Computer Simulation Technology software that replicates the real seabed logging applications to generate various synthetic data. Hydrocarbon is known to have high resistivity, about 30 – 500 ohm-meter if compared to sea water of 1 – 2 ohm-meter and sediment of 2 – 3 ohm-meter. From our simulations, we notice that the depth more than 1,750 m of offset the data is not reliable. Then, from the functions, we determine if it comes from the environment with hydrocarbon or without hydrocarbon. Data collected were processed using Gaussian Process method and focused on squared exponential covariance function types using codes in MATLAB.