神经网络结合聚类分析进行储层油气预测

THE OIL-GAS PROGNOSIS OF RESERVOIR BY THE NERVOUS NETWORK IN COMBINATON WITH THE CLUSTER ANALYSIS

ＢＰ神经网络具有较高的容错性， 具有良好的适应性， 在已知井的情况下， 由已知井出发外推预测储层的含油气性，是其优势项目。聚类分析在没有钻井或者钻井资料较少的工区，用邻区钻井资料约束确定具有代表性地震道的特征参数， 作为神经网络油气预测在学习阶段的期望输出（样本）， 具有很强的生命力， 同时它也是一种模式识别技术。通常， 地震资料油气预测软件在地震道特征参数的提取方面， 只注重地震参数， 而没有考虑地质参数； 只注重常规地震特征参数， 而没有考虑外界条件对地震道特征参数的影响。笔者除了注重地震参数以外， 还考虑地质参数对油气预测的重要性， 除了常规特征参数以外， 还消除外界条件对地震特征参数的影响， 提取出通过特殊变换而得的无名特征参数。

The BP nervous network has high error tolerance and good adaptability. And the superiority of it is in the extrapolation and prognosis of the oil gas potential of a reservoir according to the known well. In the area without or with a little well data, the cluster analysis can be used for determining the characteristic parameter of a representative seismic channel basing upon the well data of a nera area, as an expected output (sample) in the study stage of the oil gas prognosis with nervous network, which has great vitality and is a model identification technique. Usually, in the extraction of the characteristic parameter of seismic channel ,the oil gas prognosis software of seismic data only focuses on the seismic parameter without consideration of the geological parameter, only pays attention to the usual seismic characteristic parameter without consideration of the influence of outside conditions on the characteristic parameter of seismic channel. On the contrary, we have extracted the nameless characteristic parameter with a special variation by the consideration of the importance of geological parameter in oil gas prognosis and elimination of the influence of outside conditions on the seismic characteristic parameter besides paying attention to the seismic parameter and usual characteristic parameter.he Change of Trace Elements in Bore YA in Pudong , Shanghai$$$$ XIA Sheng jun 1, XIE Zhi ren 2, WANG Wen 3 (1 Geo and Ocean Science Department of Nanjing University, Nanjing 210093, China; 2 Institute of Coast and Quaternary, Nanjing Normal University, Nanjing 210097, China; 3 Water Resources Development and Utilization Laboratory, Hohai University, Nanjing 210098, China) Abstract:The analysis of the change of trace element content in the upper part of Bore YA indicates that the content of B and the ratio of B/Ga, Sr/Ba and Cl/Br are sensitive indicators of paleo water salinity The Bore YA palaeo salinity index curve shows, several one hundred scale salinity fluctuations occurred in this area in the last several thousand years, and they are probably relevant to sea water inundations and recessions resulted from concurrent sea level fluctuations The fact indicates that trace elements in sedimental strata in the coastal area are good proxy for conventional sea level changing indicators Key words:trace elements; water salinity; historical periods; sea level change