Abstract
Japan Agency for Marine-Earth Science and Technology (JAMSTEC) operates the scientific drillship Chikyu for scientific research. However, Chikyu has experienced problems with respect to drill pipe failure. This may be due to the limitation of indication of anomalies such as drill pipe failure in conventional drilling data monitoring.
As one of primary aims of scientific drilling is to recover core samples from sediment layers under the seabed, improving core recovery rate is very important because it can enhance the operation efficiency. Obtaining the information of lithology of drilling layer is also helpful for both scientific and operational aspects. However, there is no direct information regarding the core recovery rate and lithology. The recovery rate and lithology can be determined after retrieving a coring tool. Therefore, this study applies a machine learning technique to identify the drilling states, which includes anomaly detection of the drilling torque assuming the drill pipe failure, the prediction of core recovery rate as well as lithology.
This study aims to achieve real-time drilling state identification. Accordingly, a drilling data acquisition and distribution system was developed. The drilling data distributed from the system is read by data analysis systems in several languages for real-time analysis.
The drilling state identification models created in Python include the anomaly detection model and the prediction models of the core recovery and lithology. The models were installed in the real-time drilling data analyzing system, and real-time drilling state identification was attempted during the operation to confirm the health of the real-time drilling data analyzing system and to demonstrate identification with machine learning.
