Image detection method of drill cuttings. (16th December 2019)

• Record Type: Journal Article
• Authors: Guo, Long; Qian, Jing; Si, Jiangping; Yao, Shuxin
• Is Part Of: Journal of geophysics and engineering. Volume 17:Number 2(2020)
• Page Start: 249
• Page End: 257
• Publication Date: 2019-12-16
• Subjects: digital core -- cuttings -- CT scanning -- grain extraction -- cutting image detection; Geophysics -- Periodicals; Prospecting -- Geophysical methods -- Periodicals; Engineering -- Periodicals; 622.1505
• DOI: 10.1093/jge/gxz103
• Languages: English
• ISSNs: 1742-2132

Abstract

Abstract: With the development of digital core technology, numerical simulation experiments have been used by oilfield geologists and geophysicists in rock physics research. However, conducting numerical experiments, such as fluid mechanic simulations, in uncored parts of wells for unavailable digital models to calculate permeability is still impossible. The algorithm filters a large number of Computed Tomography (CT) scanned cutting images and detects whether they can be used to construct a digital core model. This workflow can be applied when coring fails or is too costly. It also provides as much continuous sample data as possible. Conventional methods of acquiring digital cores require coring after drilling and waiting for several months. The method we present here is the innovation of acquiring digital cores in drilling based on a fast CT scanner for a well site. In the future, obtaining petrophysical parameters in uncored strata may be necessary during drilling time. When sandstone is crushed into debris, most of its matrix grains are usually not destroyed. Instead, they are dispersed in the cuttings as loose sand grains. Some original sandstone grains are well connected by cement, and thus may be kept intact as agglomerated grains. These agglomerated grains are selected by collecting a large number of scanned slices. The artificially broken core samples are used for machine learning, and these training models are subsequently used to determine whether the scanned image can be extracted to unspoiled and cemented grains.