The Automatic Division And Correlation Methods For Sequence Stratigraphy And Its Applications

With the growth in demand for resources,and the development in the exploration technology,a large amounts of geology data should be systematic analyzed,extracting more useful geology information for a more effective exploration and utilization of oil and gas resources.Sequence stratigraphy correlation is one of the necessary steps in oil and gas exploration,we can study the change in reservoir,and identify the distribution of the connectivity layer of oil and gas situation,search for beneficial oil and gas areas from sequence stratigraphy correlation.This study will provide useful evidence for effective exploitation of oil and gas.Although the sequence boundary's identification and stratigraphy correlation methods have made the certain progress,but the technology has been not mature enough by now.At present,in the study of sequence stratigraphy,the division of sequence stratigraphy and the wells tie is mainly accomplished by the geology analyzer,but the quality of the results of the analysis depends on the the knowledge structure and experience of the analyzer,the result is subjective.The same block in the same area,the resultes from the different geology analyzers are different,and the working strength is very big.Therefore,it is necessary to develop a practical sequence stratigraphy auto correlation methods,the methods will have important theoretical and practical value.Based on the modulo maximum principle about wavelet multi-scale edge detection, multi-logging curves data fusion is realized.In order to enhance the accuracy and reliability of the sequence stratigraphy division and correlation,we need to extract the advantages of different logging curves,and fuse the advantages into one curve,use it to automatically divide and correlate sequence stratigraphy.The research analyzed several types of logging curve that is frequently used,in which natural gamma ray curve(GR) and spontaneous potential curve(SP) can be a very good reflection about the change of clay content in stratigraphy,it is beneficial to divide and correlate sequence stratigraphy.We need to integrate the two curves together,using the advantages of GR in the high-frequency and the advantages of SP in low-frequency.This research is based on the wavelet decomposition and reconstruction,using the wavelet multiscale edge detection, decomposites GR and SP after de-noising and normalization by the dyadic wavelet,askes the approximation coefficient matrix and detail coefficient matrix after decomposition in every scalees modulus maxima and location,and compare GR and SP's detail coefficient modulus maximum maxima,using the maximum method we can get the fusion detail coefficient modulus maximum maxima,then,using the optimal weighted method we can get the fusion approximation matrix to the approximation coefficient modulus maximum maxima in the two curves' same scale,the last,based on approximation coefficient modulus maximum maxima and detail coefficient modulus maximum maxima,applicate the Mallat alternation foldover algorithm to wavelet reconfiguration,we can get the different scalees fusion curves.Using the message entropy and variance as the yardstick,illustrates different scalees fusion curves'selecting principle and the optimal weight's selecting principle after data fusion.The message entropy and variance are the important target to measure the signal message abundance degree.The message entropy and variance are bigger,the message is more.Comparing the message entropy and variance in different scalees after the fusion, along with the decomposition scale increasing,the message entropy and variance is less,it illustrates that the scale increases then the fusion curves' message is less.The minimal scale fusion curve should be selected as a standard for dividing the stratigraphic sequence.In the process of the data fusion,for the reasonable assigning weight to GR and SP's approximation coefficient,in accordance with the certain grading step size,we can assign weight to GR and SP's approximation coefficient matrix to get the fusion curve. Through calculating the different weight fusion curves' message entropy,variance and time-frequency color spectrum,we discover that the message entropy and variance increase along with increasing the GR's value and decreasing the SP's weight.That illustrates that we can finally choose the GR's weight is 0.7,the SP's weight is 0.3 as the fusion curve's optimal weight when the fusion curve's message increases and combinate the time and the black contour's message in the frequency color spectrum.This research proposes normalization maximal energy spectrum to define the best scale,and use the optimal scale correspondence wavelet coefficient to divide stratigraphic sequence.In the process of using the wavelet transform dividing the stratigraphic sequence,it is important to choose the wavelet coefficient,choosing the wavelet coefficient is the same as choosing the scale correspondence wavelet coefficient.From the viewpoint of Fourier series expansion and spectrum energy,the optimal scale and the optimal scale factor correspondence spectrum energy accounts for the entire frequency band's total energy's proportion is the largest,that is the "normalization the maximal energy spectrum".Using the normalization wavelet coefficient frequency spectrum,we can search the maximal value of the energy in the scope of the wavelet scale.The maximal value correspondence wavelet scale is the optimal scale factor.Firstly,choose a wavelet to transform the logging signal,secondly,square the wavelet transform coefficient's modulus maxima,differentiate and eliminate DC components,calculate the normalization wavelet coefficient frequency spectrum,finally,search the optimal scale in the scope of the wavelet scale.By comparing the results of Morlet,DMeyer and Mexihat wavelet carrying on the sequence stratigraphy unit division among three kinds of wells with typical sedimentary characteristics,it can be concluded that using Mexihat to distinguish is accurate,and the performance is stable and suitable to carry on the sequence stratigraphy unit division. Using the above three wavelets to transform the fusion curve of the three kinds of wells with typical sedimentary characteristics,based on the time frequency color spectrum information of wavelet coefficient and normalization the maximal energy spectrum the most superior criterion and the corresponding wavelet coefficient can be determined,and the sequence stratigraphic units can be distinguished based on wavelet coefficient.When Morlet and DMeyer wavelet are used to distinguish the sequence stratingraphic units,the extrem point of the wavelet coefficient criterion is used as the demarcation of the sequence stratigraphic units,While Mexihat wavelet uses the slope throught the zero of the wavelet coefficient to distinguish the sequence stratigraphic units.By using the methods mentioned above to correlation the results of three wavelets carrying on the sequence stratigraphic unit sedimentary characteristics,it can be concluded that Morlet wavelet can not make an accurate distinction among the wells with the characteristics of the longer mudstone,and DMeyer wavelet can not distinguish the wells of interbreeded thin sandstone and mudstone accurately,while Mexihat wavelet can distinguish the sequence stratingraphic units accurately among the three kinds of wells with different sedimentary characteristics,which can be used to distinguish the sequence stratigraphic units among the related wells in the research area.Applying the dynamic waveform matching algorithm realizes well tie correlation.Fit the fusion curves of two wells respectively,and then extract curve characteristic parameters from fitting curves which correspond to each distinguished sequence stratigraphic units.The characteristic parameters of logging well curves usually include depth,thickness mean of curves,standard deviation,the curve slope fitted by straight lines,curves patterns,and so on.Using these characteristic parameters can realize the correlation among the different stratigraphy in the dy(?)ic wave matching way.During the process of correlation,a 'matching cost' is required when contrasting every two layers. If the total matching cost reaches the minimum number,the matching relationship between the layers of the two wells is the perfect one on such an occasion.Apply the above fusion methods of the logging well curves to obtain the fusion curves,distinguish the sequence stratigraphic units by the multi-criterion wavelet transformation and normalization the maximal energy spectrum,and auto correlate the sequence stratigraphic units with the dynamic wave matching way.First,choose three sections in the area of the Wu Maying—She Nvsi,and distinguish and correlation the sequence stratigraphys of the three parts automatically.Then,illustrate the correctness of the method based on the auto division and correlation results according with the sedimentary process of the three parts in the area of the Wu Maying—She Nvsi.Finally, on the seven framework section in the Shahe jie Formation of Kongnan area of Huanghua depression,apply the methods mentioned above to establish isochronstratigraphy correlation framework,and analyze the cycle correlation in that area,which deepens the cognition to the gas resource developmental mode.The method of auto division and correlation of sequence stratigraphy units mentioned above in this paper are an effective supplements,which can strengthen the reliability, objectivity of the stratigraphic correlation,and improve the stratigraphic correlation efficiency.These explorations provide a new way of thinking and an effective way to auto division and correlation the sequence stratigraphy and establish isochronstratigraphy correlation framework。...