Application of multiple constraints 3D gravity and magnetic inversion for imaging of volcanics in the basin

Abstract:

The carboniferous volcanic reservoirs have good exploration prospects. The seismic exploration is difficult to identify the lithology for different types of volcanics, but in view of the density and magnetic characteristics of volcanics have obvious difference, thus it can use the high-precision gravity and magnetic data to describe the spatial distribution of the volcanics by three-dimension inversion.

A well known fact, the three-dimension inversion of gravity and magnetic have serious multiple roots and large calculated amount. In order to improve the inversion accuracy and practicability, we realized the inversion method based on multiple constraints and GPU parallel computation. The multiple constraints contains three aspects. First, we adopt the geologic structure constraint by seismic datas to determine the distribution of different stratum. Second, we adopt the qualitative constraint base on probability tomography to reduce the multiplicity roots. The probability tomography method adopt correlation searching algorithm to gain the probability value of invetion grids, the grids of larger probability value have priority in invertion.Third, we adopt the by well datas. According to the well datas and the feature of gravity and magnetic anomaly, we can quantitative determine the physical property of volcanics, and abtain more precise result of invertion.

At last,we apply the multiple constraints invetion method to identifying the lithology of volcanics in the north of junggar basin.The area has complex surface condition and lithology,the seismic datas shows blank or clutter reflection inside carboniferous volcanics , it is difficult to identify the lithology of volcanics. We find that the volcanics have obvious density and magnetic difference through well datas,so we invert the high precision gravity and magnetic data of a area in the north of junggar basin, and abtain the three-dimension spacial density and magnetization distribution of carboniferous volcanics. Further we can analysis the lithology characteristics of volcanics and evaluate the favorable area of reservoirs.

The results show that multiple constraints inversion method based on GPU parallel can improve the inversion accuracy and effective identify the lithology of volcanics.