This paper discusses a field test that placed computer vision cameras on rigs' shale shakers to measure drilled cuttings and cavings. The images from the camera were processed using proprietary computer vision software and sent to the drilling team, where it was used to assess drilling performance, hole cleaning, and wellbore stability. This Proof of Concept (PoC) test was conducted over five months on two drilling operations in the Marcellus shale, using three different camera types placed in two shaker locations. Over 200,000ft of video data from 11 wells were collected from the West Virginia / Pennsylvania regions. This paper describes the research and development of this technology's hardware (camera) and software (computer vision).

To the author's knowledge, this is the industry's first use of computer vision to measure drill cuttings autonomously and continuously at the rig site. Computer vision applied at the shaker will increase overall ROP, greatly improve the speed and accuracy of wellbore instability detection, reduce the exposure in a hazardous zone and improve the detection of pending downhole problems.

Keywords:
upstream oil & gas, neural network, wellbore design, machine learning, shaker, classification, shale shaker, annular pressure drilling, wellbore integrity, accuracy
Subjects:
Wellbore Design, Pressure Management, Information Management and Systems, Wellbore integrity, Artificial intelligence
Copyright 2022, Society of Petroleum Engineers DOI 10.2118/210369-MS
You can access this article if you purchase or spend a download.