| Up to the end of 2000, the world's proved recoverable reserves of oil were 402×10~8 t,with the annual output being 34×10~8 t. In this year and the year after, the growth of averagedemand is estimated to be 1.5％—1.8％, while current petroleum resource can only support 35to 40 years' usage. Moreover, in both ocean and land, there are less oversize oil-gas fields thansmall size oil-gas fields or oil reservoir. The petroleum resource in our country is extraordinaryindigence, with even less per capita. In the end of 1999, the amount of proved reserves ingeology sense was 205×10~8 and the petroleum annual output was 1.6×10~8, being the 18thin the world. During the last 20 years, with the economic development, China was experiencingever growing demand for petroleum. However, domestic petroleum couldn't meet such demand.Up to 2003, the actual exploitation amount of all the oil-gas company was less than one third ofthe proved reserves. One important reason is that the proved reserves are far less than theobjective existence of oil. Therefore, the traditional reconnaissance method, analysis techniqueand computing method exact improvement greatly. And we need to enhance our knowledge ofunderground petroleum storage layer.As our problems become too complex to rely only on one discipline and as we findourselves at the midst of an information explosion, multi-disciplinary analysis methods and datamining approaches in the petroleum industry become more of a necessity than professionalcuriosity. To tackle difficult problems ahead of us, we need to bring down the walls we havebuilt around traditional disciplines such as petroleum engineering, geology, geophysics andgeochemistry, and embark on true multi-disciplinary solutions. Our data, methodologies andworkflow will have to cut across different disciplines. As a result, today's 'integration' which isbased on integration of results will have to give way to a new form of integration, that is,discipline integration. In addition, to solve our complex problems we need to go beyondstandard mathematical techniques. Instead, we need to complement the conventional analysismethods with a number of emerging methodologies and soft computing techniques such asexpert systems, artificial intelligence, neural network, fuzzy logic, genetic algorithm,probabilistic reasoning, and parallel processing techniques. Soft computing differs fromconventional (hard) computing in that, unlike hard computing, it is tolerant of imprecision,uncertainty, and partial truth. Soft computing is also tractable, robust, efficient and inexpensive.In this overview paper, we highlight role of soft computing techniques for intelligent reservoircharacterization and exploration.Some study consider that the oil prospecting information management is a dynamic process, through multiple prove methods (i.e. geology, earthquake and well logging), which includes dataacquisition, information extraction from the data and the transformation from information toknowledge of oil storage layer (hereinafter we call it Data-Information-Knowledge chainD-I-K-CHAIN). This article is to realize the scientific computing of D-I-K-CHAIN byintegrating multiple soft computing techniques. It includes the portrait oil layer identificationwhich is based on geology data and log data and the storage layer landscape tracing which isbased on earthquake data, geology data and log data. And it further provides assistant decisionsupport for the prove personnel to obtain the precise knowledge of the oil reservoir.
|
PDF Full Text Request