Downhole trajectory automation for rotary-steerable-system (RSS) tools has achieved Level 3 automation, enabling drilling-curve profiles from 0° inclination. Case studies demonstrate the benefits of downhole automation, such as increased rate of penetration (ROP), consistency of delivered dogleg severity (DLS), and reduced surface interactions. Downhole automation minimizes dependencies on variations of formation type, drilling parameters, and directional drillers’ experience levels and delivers a faster operation with a smoother trajectory closer to the well plan, reducing collision risks and easing completion operations while contributing significantly to global sustainability goals.
Advancement of Downhole Automation
Downhole automation was launched with the invention of RSS tools in the late 1990s. Before that, wells were drilled with either pendulum assemblies or mud motors. This was defined as Level 0 for downhole automation because the tools in the bottomhole assemblies (BHA) did not feature electronics and were controlled from the surface using only drilling parameters (Fig. 1 above). With the invention of the RSS tool, downhole automation moved to Level 1 because the direction and inclination sensors and controllers could execute commands.
In 2003, the first downhole automation method, hold inclination, was simultaneously introduced by a few service companies. This moved downhole automation to Level 2 as well paths began to be drilled in straight lines. The hold vertical automation method was invented in 2004. In 2008, the method of hold inclination and azimuth was invented to reduce the risk of the “walk” rate of the RSS.
The first curve-control methods were introduced in 2020, first for push-the-bit RSS and later for point-the-bit RSS tools, thereby reaching Level 3. The curve-control methods were enabled to work only from 10° inclination because of the accuracy of measurements at low inclination.
In 2024, automatic kickoff from vertical was deployed to the field and accomplished successful field tests. This mode also proved that automatic transition between the modes downhole was possible and enabled further development of downhole automation to Level 4, when all modes were transitioned automatically upon reaching a certain point of the well path. The vision for Level 5 is defined as drilling the well automatically with minimal supervision from the surface.
Curve Control
The RSS tool is preprogrammed at the base with a plan DLS target and sent to the job. The controller uses the ROP to produce a series of set-point changes for inclination and azimuth targets and inputs them to the attitude controller. This method heavily relies on the quality of continuous measurements, which can be risky at low inclination. The tool measures continuous inclination and azimuth; then, the controller automatically adjusts the steering ratio and tool‑face demand, and continuous inclination and azimuth from the control unit are sent to the surface, confirming the delivery of desired DLS and tool-face targets. The directional driller can send correction commands for DLS and tool-face changes as needed.
Downhole control reacts faster to formation changes, with the ability to change the steering ratio downhole as quickly as in 1 minute. Downhole control minimizes the disturbances dependent on a change of drilling parameters, BHA design, formation tendencies, drill-bit design, shocks, and vibrations, thus improving ROP and delivering a consistent curvature rate. It also saves time on decision-making because the controller reacts faster downhole.
Controlling DLS rate downhole eliminates the risk of unwanted high DLS hampering casing-running operations or BHA twistoff. The system is closed-loop, using high-frequency measurements and eliminating delays in actuations.
From Vertical
The automated trajectory control is based on a multilayered cascaded approach that allows different sections of a well path to be automated independently. Because curve-control mode can be used only from 10° of inclination, the need to automate the section of the well path kicking off from vertical was still not fulfilled. When kicking off from vertical, the magnetic tool face (MTF) is used. When the directional driller is ready, they can switch to the gravity tool face (GTF). It was decided to develop an automatic-kickoff mode wherein the tool can internally transit from MTF to GTF and then automatically transition to a curve-control algorithm. Automatic kickoff is a downhole-control mode to transition from the vertical section to the curve-control mode as soon as the error between the MTF and target azimuth is less than 20°. It uses inputs for the target MTF direction and DLS target to kick off the RSS tool automatically, creating target inclination and azimuth to follow and then switch to curve-control mode.
Data analyses were performed on all available manual-kickoff-accuracy data to set goals and key performance indicators (KPIs) for the algorithm. The x-scale is normalized to –1 and 1, and the y-scale shows the normalized frequency of runs used in the analyses. The distribution error around the target azimuth can be more than one quadrant ±45° from the intended target. The automatic-kickoff mode in a closed loop would reduce this error by adjusting steering parameters such as steering ratio and tool face in the control unit and automatically transitioning to curve-control mode or hold-tangent mode. The parameters also were defined from the study of a massive database of runs to see at which inclination directional drillers would switch to gravity mode manually. During automatic-kickoff mode, the directional driller can adjust parameters such as DLS and tool face and perform a 3D kickoff if necessary.
The control software was validated at the product center using automated tests on a hardware-in-the-loop system. Several drilling scenarios for various inclinations and azimuths and tool configurations were specified and implemented in an automated software-testing suite. Electronic stimulus signals that simulated sensor measurements and directional-driller commands were sent to tool electronics from a real-time computer running mathematical models of the drilling-system components. The responses from the tool electronics to the various stimuli were validated against predefined criteria.
This practice demonstrated a lower risk in deploying new firmware, reducing risks and uncertainties. Modeling before field release also reduces the number of iterations during the firmware-development phase. It eliminates the need for a third-party testing facility, thus reducing budget.
The complete paper presents two case studies and overall results from the field testing.
Case Studies and Results
Case Study 1. The 6¾-in. section was drilled with a push-the-bit rotary steerable motorized BHA using most of the automated modes available to date. The well plan called for 9°/100 ft DLS with a target MTF of 345°. The tool was preprogrammed at the base with the initial settings for the desired MTF and DLS. A field test demonstrated successful results. The BHA was washed to the bottom of the hole, where the automatic kickoff was engaged at 5° of inclination. The directional drillers continued drilling as normal, and the tool switched to GTF mode at 8° of inclination, transitioning automatically to curve-control mode. The RSS azimuth measurement was 1° difference from the target MTF, which was within the target KPIs. The number of downlinks was reduced by 67%. The directional drillers continued to drill in curve-control mode and, after landing the well, switched to hold-inclination and azimuth mode.
Case Study 2. The 8¾-in. section was drilled with a push-the-bit rotary steerable motorized BHA using most of the automated modes available to date. The well plan called for 8°/100 ft DLS with a target MTF of 180°. The tool was preprogrammed at the base with the initial settings for the desired MTF and DLS. The automatic kickoff was engaged in the vertical hole; the directional drillers continued drilling as normal, and the tool switched to GTF mode at 5° inclination, transitioning automatically to curve-control mode. The RSS azimuth measurement was 5° difference from the target MTF, which was within the target KPIs. The number of downlinks during the kickoff phase was reduced by 30%. The directional drillers continued to drill in curve-control mode and, after landing the well, continued with hold-inclination and azimuth mode, drilling one of the longest horizontal sections on record.
Results
The field test of the automatic-kickoff mode was completed successfully, with a total of 21 runs kicking off wells from vertical in the desired direction and switching automatically to curve-control mode. Sixty percent of the runs were performed in unconventional areas. The average number of downlink reductions was 33%. The automatic-kickoff mode was run with different types of BHAs with push-the-bit RSS, standalone and motorized. Eight runs were performed with third-party measurement-while-drilling tools and limited information transmitted to surface. No limitation to perform only 2D automatic kickoff exists because the correction of the tool face is possible by downlinks.
The kickoff distance is a relatively short part of the well path; however, correct kickoff is crucial for directional control and tortuosity reduction at the start of the section. The well path must be as smooth as possible, and downhole control is one way to achieve this goal.
Per the global analysis of a database of more than 10,000 runs during the kickoff phase in manual mode, an average of 4.8 downlinks was performed. The data were chosen based on similar BHA design and RSS type used during the field test. The interval was taken up to 8° of inclination.
Download complete technical paper here. Available until 31 March 2026.
This article, written by JPT Technology Editor Chris Carpenter, contains highlights of paper SPE 225656, “Achieving Level 3 Automation in Rotary-Steerable-Systems Downhole Control: One Step Closer to Self-Steering RSS,” by Katerina Brovko, SPE, Victor Marquinez, and Maja Ignova, SLB, et al. The paper has not been peer-reviewed.