Multi-parameter Reservoir Prediction And Fluid Identification Method Study

Now, with the AVO theory of continuous improvement and development, particularly prestack simultaneous three-parameter (velocity, shear velocity, density) inversion of the mature, three-parameter profiles can be through the physical relationship of rock fluid properties have direct access to other Parameters, fluid parameters set, however, many different types of fluid can be identified is also different, it is difficult to choose the right according to the actual parameters of fluid properties. Therefore it is necessary to carry out multi-parameter flow identification method based on systematic research.In this paper, the rock based on the theory of physics, the first of the rock containing different fluids change their properties characteristic set of parameters for various parameters in different media to show the characteristics of the intersection based on a quantitative method to select More sensitive to the target reservoir parameters, and then verify using cross plot methods and analysis of parameter sensitivity analysis includes one-dimensional cross plot analysis, two-dimensional and three-dimensional parameters parameters intersection intersection of three aspects. Finally, based on BP neural network of fluid identification method. Paper is used throughout the study Castagna and Smith's theoretical model for the data source.The paper analyzes the theoretical model of the process, through the quantitative optimization of the intersection of the fluid out of the high sensitivity of the parameters, high sensitivity to parameter optimization, multi-parameter identification to improve the efficiency of fluid, avoiding the low sensitivity Parameters involved in forecasting and uncertainty of multiple solutions. Identified in the fluid analysis, one-dimensional, two-dimensional intersection has a more clear distinction between the effect of target reservoir, intersection of this particular three-dimensional parameters in the proposed new method, making the distinction between the reservoir effect is more evident. Neural network recognition in the fluid used to study the BP neural network pattern recognition technology, the use of pre-fair method of selection by a good sensitivity parameters, as input, by comparison, the sensitivity analysis highlights the superiority of parameters as the input set .