Research On Sedimentary Microfacies Identification Based On Rough Set Theory

Reservoir sedimentary facies analysis is the basic work for reservoir description and evaluation, the key to evaluate exploitation reservoir is to analyse microenvironment and microfacies. Nowdays, the sedimentary facies identification method is mainly artificial identification according to core and logging data. This method is not only time-consuming, laborious, expensive, but also the results influenced by subjective factors. The main research method is comprehensive study of all kinds of information, analysis of core and logging features, extraction of parameters, combined with mathematical and computer for quantitative identification.One of the keys to solve the issue of sedimentary facies identification is to select the appropriate algorithm. Microfacies types are random, fuzzy and variability, and restricted by the extent of the available information, internal relations, collection methods, so that, quantitative identification is inaccurate, incomplete and uncertain. The existing methods have two types:Based on the classical set theory that using the assumption of mathematics to eliminate or avoid inaccuracy, incompleteness and uncertainty, the result is not satisfactory; Based on fuzzy set theory can resolve the information inaccuracy, incompleteness and uncertainty, but the method can not calculate the exact number of ambiguous elements, can not determine which elements are useful or redundant,and it needs some additional information or prior knowledge, but this information is not always easy to get, or there is a big subjective level, these shortcomings limit the application of the method. Rough set theory is a mathematical tool used to characterise incompleteness and uncertainty that can effectively analyze inaccurate, incomplete and uncertain information, you can also analyze data and do reasoning, discovering hidden knowledge and revealing the potential,so it is a good theory for the identification of sedimentary facies.In the paper, rough set theory is used to analyse the imprecise, incomplete and uncertain features of information effectively, and also its analysis, reasoning, mining capacity about the theory,at last the exploratory research on whether the theory can be carried out in identification of sedimentary facies. The main contents consist of:1 sedimentary microfacies identification parameters extraction. According to petroleum geology, sedimentology, logging sedimentary and other disciplines, use of regional geological data, analyse coring data and contrast logging data, to determine the microfacies types, determine the logging typical, quantitative calculation, and then establish a method of extracting identification parameters.2 Sedimentary microfacies identification parameter discretization. Discussed the need for discretization of continuous attributes and the primary method of dispersing attributes by rough set theory, this paper applies Boolean logic combined with rough set theory to discrete sedimentary microfacies identification parameters.3 Identification parameter reduction and pattern recognition. Apply reduction algorithm based on rough set theory to reduction of sedimentary facies identification parameters, according to features of the parameters, eliminate the redundant parameters, and strengthen the effective to exclude the subjective factor; and then pattern recognition method based on rough set theory will be introduced to the identification of microfacies types. From a new perspective, the algorithm starting with the use of data with inaccurate, incomplete and uncertain characteristics, explore the new methods and new technologies of the identification of sedimentary facies.4 Software programming and analysis. The preparation of automatic recognition software, the use of the software to identification of sedimentary facies of the study area and the analysis of the advantages and disadvantages.The application shows, the sedimentary facies identification algorithm based on rough set theory has higher accuracy, which is an effective and potential method. The method also can be applied in such as basin type partition, lithology identification and other geological identification. Finally analyze the deficiencies of the method and improvement in order to further improve recognition accuracy and efficiency.