Reservoir Parameters Inversion Based On The Soild/Fluid Interaction Model

Reservoir parameter inversion based on the rock physics model is one of the mostimportant problems in petroleum exploration. Rock physics model, which links elasticparameters such as impedances and velocities to reservoir parameters such as porosity,saturation and permeability, plays the important role in the reservoir parameterinversion. The reservoir parameter inversion from seismic data is a complicated andnonlinear problem with nonunique solutions and multi local optima. Therefore, theinversion algorithm must have a strong search ability and high accuracy.The Biot mechanism and the squirt-flow mechanism are the two most importantmechanisms of solid/fluid interaction in partially saturated porous rocks. TheBiot-Squirt (BISQ) model, which can simultaneously consider both the Biot mechanismand the squirt-flow mechanism, is an important rock physics model of fluid/solidinteraction. Based on the BISQ model in the partially saturated porous medium, therelationship between the reservoir parameters and the elastic parameters is detailedlyinvestigated by numerical examples, and some rules are concluded such as thebehaviors of the velocity of P-wave versus porosity are very different in high and lowfrequency ranges.In order to deal with the reservoir parameter inversion, a new hybrid GeneticAlgorithm, which combines genetic algorithm, self-adaption, and simulated annealingalgorithm, is developed as the inversion algorithm. The hybrid Genetic Algorithm welldeals with the problem of nonunique solutions and trapping in the local optima duringthe process of searching for the global optima. Numerical inversion experimentsindicate the superiority of the hybrid Genetic Algorithm over the traditional GeneticAlgorithm in terms of solution quality, convergence rate and convergence speed.Based on the BISQ model, the presented hybrid Genetic Algorithm has beenapplied on two groups of real experimental data to show the reliability and applicability.The result shows that the velocities, which are based on the inversion results, matchwith the experimental data very well. Application to two reservoir study has provided asa main result the porosity and saturation of the study areas. Areas with high porosityand low saturation match with areas with high gas probability. This coherence demonstrates the applicability of the new hybrid Genetic Algorithm.