METHOD RESEARCH OF FORECASTING OIL AND GAS USING HIGHER-ORDER STATISTICS
XIONG Xiao-Jun YIN Cheng ZHANG Bai-Lin DING Feng LI Da-Wei Southwest Petroleum Institute, Chengdu 610500, China
Based on the higher-order statistics property being insensitive to Gaussian noise and containing more information than power spectra, and higher-order spectral attributes of seismic signal, we propose a new approach to forecast oil and gas with the higher-order spectral parameter. The results of the theoretic model and the test on practical data not only prove the geological meaning of higher-order spectra, but also draw a change role of higher-order spectra between water-containing and oil-containing sandstone. It can be used to forecast oil and gas directly. Further more, its prediction results are more effective than that of regular amplitude attributes.