Fast inversion of array laterolog measurements in an axisymmetric medium

• Published in: Applied geophysics Vol. 16; no. 4; pp. 539 - 548
• Main Authors: Hu, Song, Chen, Long, Wang, Jin
• Format: Journal Article
• Language: English
• Published: Beijing Chinese Geophysical Society 01.12.2019; Springer Nature B.V; Petroleum Exploration & Production Research Institute, SINOPEC, Beijing 100083, China%The Third Gas Production Plant of ChangQing Oilfield, CNPC, Xian 710000%The Fifth Gas Production Plant of ChangQing Oilfield, CNPC, Xian 710000
• Subjects: Algorithms; Arrays; Complex formation; Computer simulation; Data; Data logging; Depth; Earth and Environmental Science; Earth Sciences; Electrical resistivity; Geophysics/Geodesy; Geotechnical Engineering & Applied Earth Sciences; Inversions; Jacobi matrix method; Jacobian matrix; Logging; Mathematical models; Numerical simulations; Parameters; Rock; Rocks; Saturation; Thickness; 1D inversion; windowing treatment; 2D inversion; invasion profile; array laterolog
• ISSN: 1672-7975; 1993-0658
• DOI: 10.1007/s11770-019-0767-0

The array laterolog is an important tool for complex formation logging evaluation due to its high resolution and large detection depth. However, its logging responses are seriously affected by leakage events due to the surrounding rock and by mud invasion. These factors must be considered when inverting array lateral logging data, so that the inversion results reflect the true formation conditions as much as possible. The difficulties encountered in the inversion of array lateral logging data are: too many inversion parameters cause the calculation of the Jacobian matrix to be difficult and the time required to select the initial inversion values due to the slow forward-modeling speed. In this paper, we develop a fast processing method for array laterolog data. First, it is important to clearly define the main controlling factors for the array laterolog response, such as thickness, the surrounding rock, and invasion. Second, based on a depth-window technique, processing the array laterolog data for the entire well is transformed into multiple 2D inversions of the layers using a series of continuous depth windows. For each formation in a depth window, combined with the 1D equivalent fast-forward algorithm, rapid extraction of the radial resistivity profile of the formation is achieved. Finally, the 1D inversion result is used as the initial state to further eliminate the influence of surrounding rocks and layer thicknesses on the apparent resistivity response. Numerical simulation results show that the factors affecting the response of the array laterolog are the invasion properties, the layer thicknesses, and the surrounding rocks; the windowing technique greatly reduces the number of inversion parameters needed and improves the inversion speed. A real application of the method shows that 2D inversion can rapidly reconstruct the actual resistivity distribution and improve the accuracy of reservoir saturation calculations.