Study On The Analytical Method Of Sequence Stratigraphy By Curve Symbolic Deduction

Sequence stratigraphy plays an important role in the field of oil exploration.In order to ensure the accuracy of the exploration,a large amount of manual labor are used to partition the sequence based on seismic data and logging data.However,the labor division method is very inefficient in dealing with large quantities of data.In order to solve this problem,This thesis proposes a logistic curve analysis method to divide of the sequence stratigraphy.In this thesis,the outliers of logging curve are detected by applying the Grubbs anomaly detection method.Moreover,the abnormal values are corrected by using the anomaly substitution rule,and the detection and correction results can meet the expectation.In order to eliminate the effects of different scale and dimension of the logging curve data,a preprocessed method named extreme value normalization is utilized.Moreover,logging curve symbolization method is originally proposed to classify the sequence stratigraphy based on the shape features of the logging curve.In this method,three symbols UP,DOWN and ZERO are defined.In addition,symbolic rules and symbolic filtering rules are used to deal with the logging curve.The symbolic results could maintain the shape features of the curve in a great extent,and the depositional cycle are identified by the shape features of the curve symbols.Finally,the sequence stratigraphy is naturally divided.A sequence stratigraphy automatically division system is implemented following the method proposed in the thesis.Some experiments are conducted to verify its effectiveness,the Gamma(GR)curve of CB25 well in ShengLi Oilfield is analyzed by using this system,The experimental results are compared with manual method and the result shows owns a certain accuracy and reliability in analyzing sequence stratigraphy and can be used for analysis and division of sequence stratigraphy.