The fleet — eight wells in a high-pressure dry-gas system.
The operator has enough wells to create context, but not enough engineering hours to watch every pressure trace between bi-weekly reviews.
The E&P team runs eight horizontal Cotton Valley/Haynesville wells from one operations group. They are dry-gas wells in a high-pressure environment, so the operating envelope is not interchangeable with the liquid-rich Permian story or the GOR signal in the Bakken walkthrough. Wellhead pressure, flowing tubing pressure, choke position, gas rate, and the separator readings arrive through the WITSML store the team already uses. The process is familiar: an engineer pulls the traces, compares them in a workbook, and holds a fleet review every other week. That cadence is enough to discuss broad production movement. It is not enough to catch a quiet change that begins the day after the workbook closes.
Review 1 — the drop gets buried.
Well H-06 loses 180 psi of wellhead pressure over a shift, but the bi-weekly review sees a busy high-pressure system rather than a clear event.
On H-06, the wellhead-pressure trace falls about 180 psi during one operating shift. The drop is not an emergency trip, and the gas-rate response is muted by a choke adjustment made earlier that day. The value remains inside the broad alarm limits the team inherited with the control system. When the engineer reaches the next bi-weekly review, the event is one valley among eight high-pressure traces, and the surrounding production context does not make it visually distinct. The review closes with a note to watch H-06. No one can say precisely when the pressure moved, which other signals moved with it, or whether the well recovered afterward.
Within hours — Orogen finds the shape, not just the value.
Orogen compares H-06 with its own rolling high-pressure baseline and detects the pressure change before the next review can become the source of truth.
Once connected, Orogen keeps a separate baseline for each well and each relevant channel. H-06 is not compared with a generic basin threshold; its wellhead pressure is compared with the operating range it has established across recent clean hours, alongside choke position and gas rate. A sustained 180 psi departure crosses the rolling envelope within hours. The agent then checks the signal shape: the pressure fall is stepwise, the choke movement explains only part of it, and the gas-rate response does not return to its prior relationship. That combination is different from a routine control adjustment. Orogen surfaces the event with a timestamp, baseline comparison, related channels, and the raw trace window so an engineer can investigate the well instead of searching eight histories by hand.
Same shift — the alert becomes an operator decision.
The team receives a briefed anomaly while the operating context is still available, not a retrospective graph two weeks after the change.
The alert reaches the operations queue the same shift. It does not claim to diagnose the well or replace the person responsible for the call. It says what changed, when it changed, how far it moved from H-06's normal high-pressure range, and which adjacent curves deserve a look. The engineer can ask the field team to confirm gauge health, review the choke change, and decide whether to stabilize or continue production. If the pressure recovers, that recovery is part of the same event; if it does not, the next review starts with a trace window and an owned question. For an eight-well fleet, that is the difference between monitoring a dashboard and keeping a live operating memory of the wells.