Early warning of drillstring faulty conditions based on multi-model fusion in geological drilling processes

Abstract

The early warning of drilling faults is of paramount importance to prevent or reduce environmental and property damages. In view of the complex drillstring kinematic characteristics and changeable formation environment, this paper proposes a new fault early warning method for the drillstring system based on a multi-model fusion and self-updating strategy. The major contributions are twofold: 1) A hybrid drillstring early warning method is proposed to identify faulty conditions based on the fusion of multi-dimensional prediction models of the drillstring system; 2) An event-triggered model selfupdating strategy is proposed to mitigate the prediction performance degradation caused by formation uncertainties. The effectiveness and practicability of the proposed method are demonstrated by industrial case studies, and the results indicate that the proposed method outperforms other approaches.

Read or Download: PDF