Abstract: To address the problem of unclear identification and prediction of effective fractures in volcanic rocks of Chepaizi Uplift on the northwest margin of Junggar Basin, a seismic normalization model was developed for three types of fractures to compare the effects of identifying volcanic rock fractures with coherence, curvature and maximum likelihood attributes, and to quantify the density of developed effective fractures based on the maximum likelihood method by using the modeling technology of micro-scaling effective fractures by core and imaging logging. It was found in the study that the maximum likelihood attribute method could more accurately predict the volcanic fractures and achieve better imaging effect on the details of fracture development in the ault development zone; the high-angle tectonic fracture development zone where the fracture orientation was consistent with the current maximum principal stress direction was more favorable to the formation of volcanic reservoirs; it was predicted that the distribution area of the four effective fracture development zones of volcanic rocks was about 76 km², which can be used as a favorable area for further exploration. The technology in the study can accurately characterize and predict the effective fractures in volcanic rocks, which is of guiding significance to the optimization of favorable targets in volcanic rocks.

Key words: volcanic rock, maximum likelihood method, seismic normalization simulation, effective fracture prediction, Chepaizi Uplift