What's New in Spotter R2.11: Fewer False Alarms, Real-Time ECD Detection, and Clearer ROP Feedback
The short version
R2.11 is focused on signal quality. Fewer false alerts on stuck pipe and torque monitoring, a new real-time ECD risk detection capability for hole cleaning, clearer feedback from the ROP Agent when it can't provide a recommendation, and a smoothed stick-slip probability curve that's easier to read over long intervals. On the platform side, configuration is simpler and access control is now role-based with group-scoped visibility.
Stuck pipe and torque: fewer false alarms, more context
This is the update most users will feel immediately.
Casing safe-guard zone labeling. Stuck-pipe warnings triggered near casing shoe depth (configurable, e.g. within 100m) now carry a label: "in casing safe-guard zone." This doesn't suppress the alert — it explains it. The crew still sees the warning, but now it's immediately clear whether the flag is near casing or in open hole. That distinction matters when deciding whether to act or keep drilling. Warning types that are always suppressed within casing remain fully suppressed — no change there.
Smarter torque spike detection. The torque spike detector now accounts for axial motion reversals — the kind of torque behavior you'd expect when starting or stopping a trip, or reversing direction. Previously, those spikes could trigger a risk alert even though the torque response was operationally normal. Now, if the spike correlates with a direction change, it's suppressed. The result is fewer false torque spike alerts and more reliable risk signaling during actual drilling.
Stronger DS input quality checks. False static friction warnings often traced back to bad DS input samples — hookload readings taken during connections, or anomalous values that didn't reflect actual downhole friction. R2.11 validates each DS_INPUT sample against two checks: a minimum value threshold (confirming the string is connected and loaded) and a configurable relative change limit versus the previous DS_INPUT in the same activity (default tolerance ~50%). If a sample fails, it's skipped and buffered. When the next valid sample arrives, missing values are linearly interpolated and inserted into the sequence. Static friction risk is only evaluated once the input sequence is clean and consistent. The result is far fewer false DS alerts caused by connection samples or anomalous hookload values.
String weight test reminders. An optional feature — when enabled, Spotter flags when pick-up or slack-off weight measurements are overdue. The default threshold is 500m since the last test. If both are overdue, a combined reminder is published. These are INFO-level messages, not alarms — practical nudges to keep reference weights current, which directly improves the accuracy of friction and drag calculations. A configurable cooldown (default one day) prevents the same reminder from repeating too frequently.
Hole cleaning and ECD: real-time risk detection
Spotter now estimates ECD in real time based on surface parameters and flags increasing-ECD risk situations as they develop — on timescales of seconds to minutes.
This matters because it fills a gap. The existing hole cleaning detection worked on longer timescales (tens of minutes), using forecasting models that needed a complete input sequence before they could evaluate risk. That meant short-duration stands could finish before the model had enough data to work with.
The new real-time ECD capability activates as soon as mud flow rates stabilize. It runs alongside the existing models and catches risk patterns on the stands that are too short for the forecast-based approach. The generation frequency is configurable, with the upper limit set by source data sampling rate. When risk is detected and passes internal downlinking checks, Spotter raises a hole cleaning warning and publishes a message. Real-time ECD estimates are written to WITSML under the curve ECD_estim in the Exebenus Spotter log. A configurable cooldown controls how frequently these alerts can fire.
For systems running without live mud density, the configured default density feeds into the ECD calculation, so the outputs stay consistent regardless of sensor availability.
ROP Agent: telling you when it can't help
Previously, when the ROP Optimization agent couldn't provide a recommendation — because configured limits were too narrow or current parameters fell outside the optimization range — the system went quiet. No output, no explanation.
R2.11 changes that. When current drilling parameters fall outside configured ROP optimization limits, Spotter now publishes a message to the ROP_message curve identifying which parameters are out of range (e.g. RPM, or WOB and Flow). If the optimization service determines that the configured limits are too restrictive to find a solution, it publishes a message directing the user to check the configuration UI. A cooldown ensures these messages aren't repeated more than once per 60 minutes — enough to notify without spamming.
This is a transparency improvement, not a performance change. The ROP Agent works the same way it always has — but now, when it has nothing to recommend, it tells you why instead of staying silent. For crews monitoring Spotter's output in real time, the difference between "no recommendation because nothing is better" and "no recommendation because the limits are wrong" is significant.
Simplified ROP status curves. The individual WOB_STATUS, RPM_STATUS, and MFIA_STATUS curves are no longer written to WITSML. ROP_STATUS is now a simple binary: 0 means the ROP agent is not activated, 1 means it is. Higher status codes are retired. At a glance, you can see whether ROP optimization is running or not.
Stick-slip: a longer-term view
Stick-slip probability estimates now include a second, long-term moving average alongside the existing short-term probability. This gives a smoother curve that's easier to interpret over extended drilling intervals — making it practical to spot gradual increases in vibration severity that might signal developing stuck pipe risk in certain operational scenarios.
The smoothed probability is published with stick-slip status messages and written to WITSML under the curve name StickSlipProbMovAvg in the Exebenus Spotter ROP Status log.
Simpler configuration
The configuration UI has been cleaned up to reduce noise and get wells set up faster.
Single stuck pipe toggle. Instead of separate DS/MS/HC checkboxes, there's now a single "Stuck Pipe Risk" option. When enabled, all associated functionality runs automatically — DS, MS, HC predictive alerts, SPP/torque/breakover torque detection, downlinking, string weight tracking, alert aggregation and escalation. Individual sub-functions can still be disabled via other configuration flags if needed.
Context-aware field display. Configuration pages now hide fields that aren't relevant to the agents selected for a wellbore. Stuck pipe settings (casing depths, risk zones, warning sensitivity, depth-based log configuration) only appear when Stuck Pipe Risk is enabled. ROP-specific configuration only shows when the ROP Optimization agent is selected. Default mud density remains visible for all jobs since it feeds into multiple agents. Less clutter, fewer mistakes.
New configuration landing page. After login, Spotter now opens to a configuration overview page showing all wellbore configurations with status indicators and group names. Filter and search options are built in. Clicking a wellbore opens its full configuration detail.
Access control and user management
Spotter now supports role-based access with group-scoped visibility. This is relevant for operators running Spotter across multiple wells and teams.
Two roles. Administrator and User. Administrators manage users, groups, wells, and WITSML server connections. Users create, update, and delete wellbore configurations only within their assigned groups, and can start/stop wellbores they're authorized for.
Group-based access. Wells and wellbores can be assigned to one or more user groups. Users can belong to multiple groups, and access to configurations is determined by group membership. If a user tries to access a wellbore that's restricted to a group they're not part of, the UI explains why and directs them to contact their administrator. Administrators get a shortcut link from that same dialog into Group Management to make the assignment.
New management pages. Dedicated pages for User Management (listing users with role and group filters, user detail views) and Group Management (create groups, assign wells, manage membership). Regular users can only access their own user page. Only Administrators see the Group Management menu.
Bug fixes
Two fixes in this release. A bug where restarting Spotter agents on an existing well configuration could delete output records outside the intended time range has been resolved — deleted records now consistently match the unprocessed interval of the well dataset. Separately, internal warning and comment generation for certain stuck pipe alerts has been hardened, reducing cases where an alert fires without a corresponding explanatory message.
