CRPP - Comprehensive Rock Properties Predictor - A New Method for Predicting Rock Properties from the Seismic Data

The presentation is devoted to a new method of rocks properties predicting by seismic attributes. This stochastic method is the effect of considerations of how to maximize the use of multi-dimensionality of data, both seismic and geologic. An integral part of these considerations were: what sort of data should be analyzed and how the results can be effectively verified. The most important element of method is specifically configured neural network designed to predict several geological properties. The power of the method is that it simultaneously takes into account all possible geological information available at the locations of wells and maximally utilizes the information available within the seismic data. The result is a set of maps of several geological parameters predicted not one after another but all at once. All of this is done by almost fully automated process that minimizes interpreter's subjective influence on the result. The verification of the crucial properties prediction is based on relationship that allows us to check the quality of analysis in the whole area of study. This method can be described as a coherent, intelligent and independent system having a comprehensive look at the relationship between the complex geology and its seismic response.