Research On The Application Of Multiscale Analysis Method Of Well Logging To Sequence Stratigraphy

With the continual development of exploration, the trend of sequence stratigraphic analysis is developing from qualitative analysis to semi-quantification and quantification. Based on the method of well-logging multiscale analysis, the multiscale characters and the fusion of logging curves under different scales are studied systematically to extract the multicycle features of strata and to establish the corresponding relation between logging data and strata and therefore the improved logging data can be better applied to the evaluation of sequence stratigraphy.Based on the multiscale analysis theory of wavelet, single logging curve is processed through multiscale analyzing. According to the multiscale edge detection principle and the reconstruction algorithm of wavelet modulus maxima, different logging curves are fused under different scales through wavelet transform. The method of well-logging multiscale analysis is built based on the multiscale analysis of single logging curve and fusion of different logging curves. The changing characters of depositional cycle can be reflected by the corresponding frequency spectrum response characters of logging curves after multiscale analysis, and then the frequency spectrum response models can be established for identifying different types of parasequences and parasequence sets. The wavelet bases of different properties such as Morlet wavelet, Zero-Crossing wavelet, Gauss wavelet and dyadic quadric spline wavelet are applied to logging data for multiscale analysis, and then the time-frequency features and abrupt changes features of logging data under different levels can be extracted. According to the relation between wavelet scales and its corresponding modulus means and also based on wavelet time-frequency information, the quantitative method of choosing optimal wavelet scale is put forward. Through studying the periodic vibratory characters and the multiscale edge detection property of wavelet coefficients under optimal scales, the relations between wavelet coefficients and sequence boundaries are established and then sequence stratigraphic units'boundaries of different levels are quantitatively demarcated. Choosing suitable comparison index that could describe the inner characters of sequence stratigraphic units and integrating the method of well-logging multiscale analysis with the optimum ordered multielement match, based on these, the quantitative correlation on sequence stratigraphic units between different wells is realized and then the isochronous structure of sequence stratigraphy can be built. All these researches provide a completely new method for the quantitative demarcation and correlation of sequence stratigraphy.