An adaptive pattern recognition method for early diagnosis of drillstring washout based on dynamic hydraulic model
Abstract During drilling operations, the drillstring washout is a common form of drillpipe failure and needs to be detected in time. In this paper, an improved pattern recognition method for early detection of drillstring washout is proposed. The integrated diagnosis model includes three parts: (1) dynamic hydraulic model and standard washout modes; (2) adaptive pattern recognition algorithm with piecewise approximation; and (3) synthetic probability method for washout detection. The proposed method is applied to detect the washout event in real time based on the common measurements (standpipe pressure, pump rate and pit gain) of a field well. The washout is successfully diagnosed, and the detection time is 16 min lesser than that for field detection. As compared with other diagnosis methods, the proposed method shows a high diagnosis sensitivity with good anti-interference ability. This method provides an improved strategy for the early detection of drillstring washout.
Paper
Full text
An adaptive pattern recognition method for early diagnosis of drillstring washout based on dynamic hydraulic model
Semantic Scholar · Engineering · 2019
Abstract
Abstract During drilling operations, the drillstring washout is a common form of drillpipe failure and needs to be detected in time. In this paper, an improved pattern recognition method for early detection of drillstring washout is proposed. The integrated diagnosis model includes three parts: (1) dynamic hydraulic model and standard washout modes; (2) adaptive pattern recognition algorithm with piecewise approximation; and (3) synthetic probability method for washout detection. The proposed method is applied to detect the washout event in real time based on the common measurements (standpipe pressure, pump rate and pit gain) of a field well. The washout is successfully diagnosed, and the detection time is 16 min lesser than that for field detection. As compared with other diagnosis methods, the proposed method shows a high diagnosis sensitivity with good anti-interference ability. This method provides an improved strategy for the early detection of drillstring washout.