Indonesia Conference Directory

Pore pressure prediction methods have been developed by several researchers who only use well log data as a basis for determining the relation between velocity and pore pressure. Pore pressure prediction by using laboratory measurements to get the relation with velocity is still rarely conducted. In this paper, the prediction of pore pressure is done in multi-stages or sequentially involving the results of rock physics laboratory measurements for each reservoir lithology (carbonate, basement, and sand) and combining them with field measurement data for each well using statistical rock physics and statistical neural network methods. This method involves all measurement data from the core in the laboratory, well data, including data: lithology, measurement of pressure data in well (RFT, DST, etc.), hydrostatic trend, acoustic and elastic log data, porosity, and mud weight information. The whole data are trained using statistical neural network. Then, the knowledges are used to predict the pore pressure by considering lithofasies of log and also seismic lithology.