Abstract This article presents a cutting-edge approach to automated drilling, integrating an auto-driller system with advanced ecosystem powered by dynamic digital twin, to optimize drilling performance in real time. The methodology enables continuous adjustment of operational parameters based on actual wellbore conditions, significantly enhancing drilling efficiency, safety, and risk mitigation. By identifying operational sweet spots and automating critical processes, the system delivers substantial improvements in key performance indicators (KPIs). The paper includes implementation results and a comparative analysis, demonstrating the effectiveness of this intelligent ecosystem in modern well delivery. The ecosystem provides functionality for pre-drilling planning and real-time execution phases. During the planning phase, a detailed drilling roadmap is created within the ecosystem, where operational parameters are validated through risk assessments and mechanical/hydraulic modeling to ensure wellbore and drillstring integrity. These verified parameters serve as target values for the auto-driller during execution. In the operational phase, a dynamic digital twin continuously assesses wellbore conditions, calculates optimal operational parameters, and updates the roadmap accordingly. Real-time outputs—including setpoints for on- and off-bottom activities—are transmitted directly to the OEM Programmable Logic Controllers (PLCs), enabling fully unmanned and automated drilling. The system proactively detects the severity of drilling malfunctions and anticipates potential hazards, ensuring that key risks are mitigated at an early stage. Additionally, the auto-driller operates under a predefined control algorithm, which includes: automated operations sequence with continuous monitoring and adaptation based on current wellbore conditions. The drill-ahead module automatically defines the targeted toolface orientation for slide mode execution to follow the planned trajectory. Implementing pre-drill planning within the ecosystem enables the selection of downhole equipment, identification of optimal drilling parameters, and estimation of expected ROP, taking into account the analysis of efficiency from previous offset wells drilling operations. Leveraging the ecosystem to control an auto-driller guarantees the execution of technological operations under optimal conditions. Wellbore, drilling equipment states are continuously and automatically evaluated in real-time by dynamic digital twin. Rock cutting process and Wellbore treatment efficiency, safe operational envelop and drilling dysfunction risks are automatically evaluated along the execution. Well construction cycle improvements exceed 30% of the conventional drilling and predefined lookup auto-drilling techniques for similar wells. Using the drill ahead module has reduced sliding time by up to 10%. NPTs and ILTs losses were decreased by operating in the actual safe operational zone and mitigating at an early state downhole hazards such as: micro influxes and mud losses, pack-offs, stuck pipe, premature drilling tools wear, etc. and BHA overloads due to excessive torsional and lateral vibrations stick. The digital twin 1hz frequency automated analysis identified all precursors of upcoming hole condition deterioration: deviation of slack off weights and micro overpulls, automatically identifies FFs and generates sensitivity trend analysis, notifying key personnel, generating reports, and setting values to swiftly achieve optimal states. The applied methodology proves its viability and effectiveness in well construction, particularly in settings with constrained operational pathways. Integrating the ecosystem that merges well design and real-time analysis with auto-driller systems presents an opportunity to deploy innovative process management techniques to achieve consistent improvements throughout available equipment. The integration of the digital ecosystem can seamlessly span across a diverse range of drilling systems, offering unprecedented flexibility to harness hybrid AI and physics-based insights.
Карпов et al. (Mon,) studied this question.