Estimation of gas hydrate saturation in Ulleung basin using seismic attributes and well logs

Abstract

Gas hydrate has recently received much attention as a potential energy source. To determine the candidate sites for the production of gas hydrate, the distribution and saturation which are closely related its reserves should be preferentially predicted. The saturation of gas hydrate has often been computed by using well logging data, but the results were limited to the well location or the region near wells. On the other hand, neural network technique enables us to predict the hydrocarbon saturation of interest throughout the seismic coverage by combining multiple seismic attributes and well logs in oil and gas exploration. In this study, we identified the distribution of gas hydrate in Ulleung basin, Korea by using seismic attributes such as impedances and elastic parameters extracted from seismic data. In addition, we estimated the saturation of gas hydrate throughout the seismic profile by applying Probabilistic Neural Network (PNN) which uses the extracted seismic attributes and gas hydrate saturation at wells calculated by TPBE as training datasets.