Carbonate reservoirs dominated by secondary storage space: Key issues and technical strategy

Carbonate reservoirs of the Tarim Basin in China are buried deeply, with strong later diagenesis, and seldom are controlled by geologic facies. The effective storage spaces are mainly dissolution caves and fractures with few primary pores. Fluid distribution in the reservoir is extremely complicated. All these characteristics make it a world-class challenge for oil and gas exploration and production. Conventional poststack methods ignore the complexity of the reservoir medium with low accuracy of prediction results, by which the location of the dissolution cave is determined roughly. Only faults and fracture zones can be described to a certain degree instead of quantifying the fracture detection. Aiming at the key problems of the complex carbonate reservoirs in the Tarim Basin, a series of techniques has been developed to improve the successful rate of reservoir prediction, including the following aspects: the wide-azimuth and high-density data-acquisition technique to improve the quality of the original seismic data; a series of migration techniques including prestack reverse time migration; and amplitude-preserved Q migration to improve the quality of seismic gathers and the imaging precision with the matching precision of AVO gathers. The target reservoir data dominant frequency increased by more than 15% and by nearly 20 Hz. In addition, the effective DEM-Gassmann rock-physics model is proposed for velocity prediction, by which relative error of estimated shearwave velocity reaches to less than 5%. Furthermore, an advanced inversion algorithm is introduced to improve the accuracy of extracted elastic information. By using new prestack inversion and frequency-dependent AVO inversion, the tie rate of the dissolution-cave reservoir and fluid prediction increased by more than 15%. Furthermore, the fracture-prediction tie rate can be enhanced by 20% through limited-azimuth anisotropic inversion, on the basis of which the quantitative identification of fracture fluid for actual seismic data can be realized.