Rock Physics Modeling And Reservoir Parameter Prediction Of Tight Reservoir

With the deepening of petroleum exploration and exploitation,the potential of conventional reservoirs have been decreasing,so the focus of exploration has gradually changed from conventional reservoirs to unconventional reservoirs.Among the unconventional reservoirs,tight reservoirs have become the focus of researches for its large reserves and exploration potential.Compared with conventional reservoirs,tight reservoirs are famous for the low-porosity,low-permeability and complicated pore structures.Therefore,the conventional theoretical rock physics models and methods don't work well in reservoir predictions.In laboratory analysis,well-logging interpretation and seismic inversions,conventional theories and methods may cause the errors of results.Reservoir prediction and description techniques for tight reservoirs are in great need.Conventional methods of reservoir prediction and fluid discrimination need to be developed and improved.In this thesis,based on rock physics theory,researches on reservoir prediction and description have been carried out according to the geophysical characteristics of tight reservoirs.Rock physics modeling is one of the most important methods in rock physics research.In this thesis,based on the research on the basic theory and assumptions,real rocks can be equivalent to ideal rock physics models.Elastic properties of saturated rocks can be linked with the properties of rock matrix,rock skeletons and pore fluids.According to the complicated pore and structure characteristics of tight reservoirs,such as the low-porosity and unconnectivity of tight sandstones and the aligned organic of tight shales,rock physics models of tight sandstones and shales are built to calculate elastic parameters.According to the complicated pore structure and anisotropic characteristics of tight reservoirs,seismic forward modeling was implemented based on a cracked porous anisotropic two-phase media theory.The influences on seismic wavefield of different crack parameters,pore properties and fluid properties have been discussed.This lays the foundations for the subsequent fluid discrimination,reservoir parameter estimation.Reservoir fluid discrimination is an indispensable part for reservoir prediction and description.Reliable fluid indicators help to decrease the risk of exploration and to increase the success ratio of drilling.The regular fluid indicators for conventional reservoirs may lead to errors in tight reservoir discriminations.Fluid indicators suitable for tight reservoirs are needed.In this thesis,we have proposed an alternative strategy for the combination of fluid indicators for tight reservoirs.An alternative fluid indicator RPTI(Fluid indicator based on rock physics templates)has been derived to combine the advantages of different single indicators.The combination is implemented based on the characteristic of the fluid trend in the rock physics template,which improves the discrimination accuracy.In order to achieve further quantitative reservoir information,reservoir parameter estimation is essential.Based on the rock physics theories proposed in this thesis,we have proposed a rock physics inverse modeling method with model constraints to estimate tight reservoir parameters.In a reservoir parameter domain,reservoir parameters can be estimated by the isosurface intersections of elastic parameters.An extended rock physics inverse modeling method has also been proposed.Inverse modeling should be implemented in two different reservoir parameter domains.Then,with the domain transformations of the inverse modeling results,four different reservoir parameters can be estimated simultaneously.Different kinds of elastic parameters make different contributions to the accuracy of inverse modeling.An applicability evaluation method has been proposed for the choice of elastic parameters as input data.Satisfied results of tests on model data and real data have proved the rationality and applicability of this method in the prediction of reservoir parameters.This method provides datum foundation for tight reservoir prediction and evaluation.