Improvement Of Wavelet Transform For Well Log Curve And Application To Stratigraphic Correlation

Stratigraphic correlation is very important in the research of formation and it plays a key role in understanding the geological phenomenon and resource exploration. Wavelet analysis method is excellent in signal processing methods. The characteristics of well log curve are high sampling rate, good continuing, rich information of geology and so on. The purpose of the study is to improve the wavelet transform method for well log curve and then apply the improved method to do the stratigraphic correlation work.In this paper, we compare results of different wavelet basis functions applied to wavelet transform for well log curve. Furthermore, we analysis many extension methods to take a toll on boundary effect, including periodic extension, symmetry extension and so on. Moreover, we analogy sedimentary process. Using the model, we do the wavelet transform. From the practice, we get the method of distinguishing layer interface. Last but not least, we apply corrected wavelet transform method, to achieve the purpose of the stratigraphic correlation of the formations in Qilian mountain regions of Qinghai province.The Result shows that:(1)morlet wavelet function is excellent in all kinds of wavelet basis function in processing logging data to differentiate cycle information. Because it has high correlation coefficient in inversion of reconstruction of well log data with original simulating GR well log data and the concentration of its spectrum diagram is very high.(2) Wavelet transform with well logging data has boundary effect: Periodic extension method is very useful in undermining boundary effect. The length of periodic extension is controlled by wavelet transform maximum scale.(3) A certain frequency components of well log curve, has good response in a depositional period. Thus, the amplitude of wavelet coefficient is larger than other depositional period. Thus, we can classify the layer interface.(4) In Qilian Mountain area, the optimum scales are 15, 50, 100, 150, and 300. The GR curves contains Milankovitch cycles.All in all, the improved wavelet analysis method is well applied in the Qilian mountain formation stratigraphic correlation.