Research On Key Technology Of Geosteering Interpretation And Prediction While Drilling Based On Multi-source Information

The geosteering drilling can improve the well trajectory in time according to various real-time information of the stratum and the borehole,so as to improve the drilling rate of the reservoir,so it is of great significance.It is an important research topic that whether or not it can be quickly and automatically layered based on multi-source information.However,traditional geosteering drilling needs the underground geosteering instrument to upload the data of stratum information and well trajectory which obtained while drilling to the ground,after analyzing by the scientific research workers of research institute,then adjust well trajectory according to the decision which downloaded from the ground.This method is slow,prone to various unstable factors.Therefore,it is urgently needed to establish a geosteering interpretation prediction method of automatically interpretation and stratum information prediction based on the block information and adjacent well data combined with multi-source information of the well.Based on the summary of previous research results,the well logging curve during drilling is used as multi-source information,the wavelet transform and wavelet multiresolution analysis are used as a mathematical tool to carry out the following work in this thesis:(1)using the threshold method and the spatial domain method to denoise to logging drilling curve,retain abrupt change feature and edge feature of signal while removing high frequency noise simultaneously;(2)extract a variety of trends along with logging curves,remove the ultra-long stratum periodic trend background;(3)reconstitute several curves which can reflect the characteristics of the strata and do not relate to each other from many curves of correlation through principal component analysis,achieve the purpose of stripping sequence stratigraphic information corresponding by the curves;(4)preferred wavelet scales,analyze the connection between Low frequency and high frequency part and the sequence stratigraphic units,then determine the divide of sequence stratigraphic interface,and finally establish a model which can identify sequence,parasequence set,parasequence and layer group.In addition,further study analysis rules of stratigraphic cycles of multi-level secondary datum,and the relationship between the stratigraphic cycles of the three levels in the long,medium and short term and low frequency and high frequency curve;(5)using wavelet multi-resolution decomposition and edge detection method to fuse the low frequency approximation information and the high frequency detail information of several logging curves during drilling,and reconstruct a fusion curve which can better reflect the formation characteristics;(6)the ARIMA model based on time series method,estimated the low frequency and high frequency curves which have been obtained,thus carry out appropriate forecast to the stratum ahead of the drilling bit.