Oil & Gas Geology ›› 2022, Vol. 43 ›› Issue (6): 1505-1514.doi: 10.11743/ogg20220619

• Methods and Technologies • Previous Articles     Next Articles

Physical simulation and distribution prediction of karst cave collapsing in deep carbonate reservoirs

Xinrui Lyu1(), Xingwei Wu1, Jianfang Sun1, Dongling Xia1, Yanpu Li2, Yanzhi Ding3, Bin Wang3   

  1. 1.Petroleum Exploration and Production Research Institute,SINOPEC,Beijing 102206,China
    2.No. 1 Oil Production Plant,Dagang Oilfield Company,PetroChina,TianJin 300280,China
    3.Geotechnical and Structural Engineering Center,Shandong University,Jinan,Shandong 250061,China
  • Received:2021-04-18 Revised:2022-09-19 Online:2022-11-21 Published:2022-11-21

Abstract:

Large paleokarst caves constitute important storage space in deep carbonate reservoirs. Drilling data reveal that many such caves are filled to some extent by collapsed breccia. A better understanding of their collapse mechanisms, modes and distribution would be helpful in the highly efficient development of such reservoirs. This study applies the similarity theory to the making of unfilled, sand/mud-filled and water-filled spherical caves using the molten wax cavity method to simulate the large karst caves in the Ordovician carbonate reservoirs for geomechanical physical model test of karst cave collapsing. Collapse modes are established based on the revealed collapse mechanism, morphology and damage zone of simulated karst caves of different filling types and the possibility of seismic prediction is also explored based on the seismic reflection characteristics of the collapsed karst caves during the tests. The results demonstrate the effectiveness of using physical model test to analyze the karst cave collapsing in deep carbonate reservoirs under in-situ stress in the Tahe area. Being more visualized and real than traditional methods, the test reveals that compaction and shearing are the main causes of cave collapsing, during which the roof and surrounding rock of the cave sink in, with a shear-slip damage zone of 2 to 3 times larger than the cave diameter. The sand/mud-filled cave is shown to be more stable than the water-filled cave, and the unfilled cave is the least stable. Two cave collapse modes are established: Type Ⅰ being mainly affected by overlying gravity and type Ⅱ by shear stress. Their seismic reflections show local differences. If combined properly with colored inversion and GR inversion as well as seismic attribute like frequency division energy and coherent energy gradient, and etc., the reflection differences can be used to map the two types of collapsed karst caves. The prediction of collapsed karst caves shows a good correspondence with oil well productivity.

Key words: karst cave collapse, physical simulation, collapse mode, collapse prediction, deep carbonate reservoirs, Tahe oilfield

CLC Number: