Abstract
In a pioneering collaboration, a major rig contractor and a Software as a Service (SaaS) provider have jointly created an innovative solution for closed-loop drilling automation. The primary focus of this technology is to achieve cost reduction per foot of drilling by employing a Machine-Learning (ML) based rate of penetration (ROP) optimizer while reducing dysfunctions. This optimizer enables remote control of rig site auto driller setpoints through a seamless cloud-to-cloud connection, without the need for any additional rig devices.
The ML ROP optimizer has been previously published in SPE publications by the SaaS provider, with references (SPE-212568-MS, SPE-208751-MS, SPE-204043-MS, SPE-URTEC-343). A rig contractor developed an Application Programming Interface (API) to receive control setpoints - weight on bit (WOB), ROP, revolutions per minute (RPM), and differential pressure (DPRES) from the optimizer. These setpoints are synchronized between cloud and edge, then to the Human-Machine Interface (HMI), ensuring reliable connection via a heartbeat mechanism. An enhanced auto driller system was deployed to improve setpoint control. Field testing involved a three-well campaign in the Permian Basin, targeting lateral sections in two formations in collaboration with a major Oil and Gas Operator.
Several key performance indicators (KPIs) were developed to evaluate the performance of the automation technology against conventional manual drilling. In Test Well 1, 52% of lateral footage was drilled with the automation technology enabled, resulting in a 13% improvement in ROP and successful completion of the lateral in a single bit trip. For Test Wells 2 and 3, the percentage of lateral footage drilled with the automation technology was 56% and 48%, respectively. On-bottom ROP was improved by 25% on average after filtering out control drilling for rotary steerable system (RSS) downlinks. For both Test Wells 2 and 3, no unplanned trip occurred.
To mitigate the impact of high vibration during drilling, the vibration filter feature was activated for Test Well 3. This feature avoided sending setpoints that were correlated with recent high vibration. This automated mitigation approach effectively reduced vibration without requiring manual intervention. The remote control of setpoints significantly minimized the need for manual intervention, reducing driller fatigue and allowing them to focus on other critical tasks for ensuring safe drilling operations. The paper will present detailed data supporting the observations.
This paper presents an innovative integration of ML and cloud technologies, achieving real-time control of auto-driller setpoints without the need for additional edge devices on the rig. The seamless cloud connection enables scalable and remote deployment of AI and drilling automation, leading to optimized operations and cost reduction. The collaboration between an operator, rig contractor, and SaaS company sets a pioneering example of closed-loop drilling automation.
