Pairwise Critical Point Detection Using Torque Signals in Threaded Pipe Connection Processes

Abstract

The quality of threaded pipe connections is one of the key quality characteristics of drill pipes, risers, and pipelines. This quality characteristic is evaluated mainly by a pair of critical points, which are corresponding to the mechanical deformations formed in the pipe connection process. However, these points are difficult to detect because of nonlinear patterns generated by latent process factors in torque signals, which conceal the true critical points. To address this problem, we propose a novel three-phase state-space model that incorporates physical interpretations of connection process to detect pairwise critical points. We also develop a two-stage recursive particle filter to estimate the locations of the underlying critical points. Results of a real threaded pipe connection case show that the detection performance of the proposed method is more powerful than that of other existing methods.