Development of Drill Pipes Failure Prediction Models and Operational Management System Using Real-Time Data Analytics and Ai

Abstract

The drop on the daily rates for the Drilling Rigs in the recent years has pushed Drilling Contractors in the industry for innovative solutions. Industry 4.0 is bringing many features and technologies to overcome these challenges and help the companies to meet this new scenario. This paper will present how a partnership between Ocyan, an ultra-deep-water Drilling Contractor and RIO Analytics, an A.I. technology company that develops solutions for failure prediction of industrial assets, is using artificial intelligence and Data Analytics to manage and control drill pipes operation and prevent failures, correlating different sources of information. Drill pipe is one of the most critical equipment on a deepwater Drilling Rig and Drill pipes incidents are one of the biggest causes of nonproductive time and unplanned costs in the drilling industry. In most cases, the lack of information about the drill pipes, such as historical and operational efforts related to their individual use make it very hard to investigate an incident that occurred, and consequently, to predict a pipe failure. Also, some operational limits (such as make-up torque and elevator capacity) that are driven by dimensional inspection results are often not used correctly for operational planning, leading to unnecessary risks. To be able to apply failure prediction algorithms and correlate operational and historical information for each individual drill pipe, a web-based software was developed building a valuable database and management system, allowing users to easily navigate for drill pipes information, generate reports, and simulate operational scenarios by providing operation planned tally (list of drill pipes). Warnings are generated as the results for the simulations indicating any risk for operations. Critical situations are made available to the rig crew, immediately transmitted to the Ocyan's Decision Support Center (CSD) and management team onshore, while less critical alerts are recorded in the system for further investigation. Software integrates with different inspection reports formats and automatically updates critical information on drill pipe's database, allowing also to identify invalid or wrong information on these reports, upon inspection criteria used. With the implementation of this predictive maintenance solution, companies aim to increase Operational and Process Safety, avoid NPT and reduce maintenance cost regarding the Drill Pipes. Based on the integration with real-time data from rig sensors and identification of active operational tally, it has been possible to automatically control drilled meters and rotating hours for each drill pipe, which triggers inspection requirements, generating automated work orders for the CMMS. Also, an algorithm was developed to calculate real-time damage in each drill pipe during operation, considering the most significant parameters (such as torque, tension, drilling depth, wear, pressure, dog leg severity, jarring, etc.), using it to provide valuable information for failure prediction.