Studying The Inversing Technique And Method On Thin Mutual Layers Of Sand,Mud And Coal

With the increasingly deep study on the oil and gas exploration, identifying and interpreting the thin sand layers(2-8m) has been more and more important. How to obtain the thickness of the thin sand layer, that is the thickness predicting technique of thin layers or thin mutual layers, has been a significant studying task on the hydrocarbon seismic prospecting nowadays. Aiming at the geological characteristic of sand, mud and coal mutual layers, this paper uses neural network arithmetic to design a inversing method to enhance the resolution of the thin mutual layers.First, this paper introduces international developing conditions currently and trends on inversing method of the thin mutual layer, and illuminates the practical significance of studying this method on the thin mutual layers of sand, mud and coal. As a result of the actual limitation of inversing technique, we design a series of researching thinking which aims at the inversing method of the thin mutual layer, that is, 1) analyzing alternately the lithologic and electronic characters from the finished wells to optimize the sensitive physical characteristic for the thin reservoir, such as density, GR and electrical resistivity; 2) correcting the physical characteristic curves which are susceptive to the thin reservoir with lithology data from the wells, to make sand, mud and coal separated clearly; 3) on the basis of regular wave impedance inversion, developing the inversion of sensitive physical characteristic curves ( such as density inversion, GR inversion, electrical resistivity inversion and so on ) aims at the inversion for the mutual layers.To test whether the study is effective, in the essay, practical data is selected to test the method. In the researching area, there is the 3D seismic data of low resolution and low dominant frequency and thin and mutual stratum containing sand, mud and coal, which produces certain difficulties for the inversion. With the above mentioned inversing method and the practical data on the thin mutual layers of sand, mud and coal, in the essay, wave impedance inversion and reflection coefficient inversion are utilized to improve the resolution of inversing result. In the studying, we discover that density is the most sensitive to the lithologic characters in the studying area as physical characteristic. Therefore, according to the studying thinking, on the basis of regular wave impedance inversion, we make use of lithologic data to correct density and sound wave curves, utilizing Jason and STRATA's EMERGE module nerve network to proceed the density inversion.The inversing result shows that using STRATA's EMERGE module nerve network to predict the thin reservoir is more effective, but too much workload; the inversing effect is more effective after using lithologic data to correct density and sound wave curves; considering reflection coefficient is an effective way to improve resolution in the arithmetic of nerve network. This method improves the longitudinal resolution of the reservoir prediction in this area, enforces the identifying ability to the sandstone of the mutual layers, provides consolidating base for the proceedingly carrying out reservoir prediction and lithologic trapping, obtains better geological effect, and provides more meaningful exploration for the hidden hydrocarbon reservoir.