| With the increasing demand for energy,oil and gas exploration and development of unconventional reservoirs becomes more and more important.However,the rock structure of unconventional reservoirs is complex and the physical properties change greatly,and the rock physical response mechanism is complex,which makes logging interpretation and comprehensive evaluation very difficult.The experimental test method of rock physics is the main method to study the physical properties of reservoir rock,but it is difficult to quantitatively and systematically analyze the influencing factors and laws of rock physical properties.As a new method,digital core technology can be used as an effective supplement to rock physics experiments,and has achieved many results in the simulation of rock physical properties.However,for shale reservoirs,due to resolution limitations,it is difficult for single-scale digital cores to fully characterize the microstructure inside shale cores.Therefore,it is particularly important to construct a multi-scale digital core model of shale containing all scale pore information.At the same time,this model plays an important role in reservoir description,rock physics numerical simulation and so on.In this thesis,the shale samples of Qingshankou Formation in Songliao Basin are taken as the research object,and the construction methods of multi-scale digital core voxel model and multi-scale pore network model are studied.The multi-resolution CT scanning imaging and SEM imaging techniques are used to comprehensively and systematically study the core pore structure.In order to achieve image matching between cores with different resolutions,this thesis adopts a hybrid registration method,which is suitable for registration between twodimensional and three-dimensional images from millimeter plungers,centimeter plungers to full-diameter core images of different scales.At the same time,in the aspect of image segmentation,a cross-scale calibration method is adopted to realize the digitization of lowresolution images by establishing the gray value-component ratio curve of different components.Compared with the traditional core analysis method,the digital core voxel model can obtain higher spatial resolution and improve the accuracy and visualization of core structure.The establishment of multi-scale digital core voxel model can more comprehensively reflect the multi-scale structural characteristics of rock.Compared with the image-based digital core voxel model,the pore network model represented by sphere(pore)tube(throat)only retains the topological structure characteristics of core pores,and greatly reduces the amount of data,which can realize the full-scale pore structure characterization.In this thesis,the maximum sphere algorithm is used to extract the pore network models of different scales and combined with the porosity distribution obtained by the voxel model,the correlation between pore parameters is used to realize the cross-scale superposition between pore network models of different scales,so as to realize the construction of multi-scale pore network model.At the same time,the model also introduces experimental data of high-pressure mercury injection and low-temperature nitrogen adsorption and uses random reconstruction method to supplement nano-scale pore information. |
