川西北双鱼石地区栖霞组地震资料优化处理及裂缝预测技术应用

Application of Seismic Data Optimal Processing and Fracture Prediction in the Shuangyushi Block,Northwest Sichuan

川西北双鱼石地区栖霞组气藏勘探开发潜力大,是目前四川盆地加快天然气增储上产的重要领域。目前该地区已经进入开发阶段,面对勘探阶段的地震资料品质及裂缝预测结果尚不能满足开发的需求,需要在已有资料的基础上进一步优化处理,提高资料精度。将该区用于构造解释的叠后地震资料进行了解释性处理,通过扩散滤波和反射系数反演技术,去除了噪声并提高了分辨率,改善了地震资料的品质;通过多属性RGB(red-green-blue)融合技术定性预测了裂缝发育特征,利用BP(back propagation)神经网络和DFN(discrete fracture network)离散建模技术量化预测了裂缝的密度、长度和发育方向。预测结果表明,双鱼石地区栖霞组裂缝总体为北东-南西向伴随断裂走向呈条带状展布,局部发育有北西-南东向的裂缝。北部山前带由于构造活动影响,裂缝发育密度高,南部裂缝发育密度稍低。地震预测结果与成像测井资料吻合,验证了方法的有效性。

There is a great potentiality for exploration and development of Qixia Formation in the Shuangyushi Block,Northwest Sichuan.It is an important area for accelerating the increase of natural gas reserves and production.At present,the region has entered the stage of development.The seismic data quality and fracture prediction results at exploration stage are unable to meet the needs of development stage,which needs further optimize processing based on existing data to improve the data accuracy.Therefore,the interpretative processing technologies such as diffusion filtering and reflection coefficient inversion were used to improve the seismic data quality of structural interpretation.Those methods could remove the noise and improve the resolution,and optimize the quality of seismic data.The fracture characteristics were qualitatively predicted by multi-attribute red-green-blue(RGB)fusion technology.And the density,length and direction of cracks were well quantitatively predicted by back propagation(BP)neural networks and discrete fracture network(DFN).The results show that the cracks are generally developed in the northeast-southwest direction along with the fault in the Shuangyushi Block,while partial area is northwest-southeast.Due to the influence of tectonic activities,the northern piedmont belt has a high density of crack development.The density of crack development in the south is slightly lower.The results of seismic prediction are in good agreement with the imaging log data,which proves the effectiveness of the method.