Why leak detection moved from patrol to data
Pipeline patrol used to mean boots, trucks, and right-of-way walks. Those sweeps still catch new construction encroachments and visible spills, but small leaks below ground do not wait for a quarterly walk. Once liquid or vapor leaves the line, the damage starts before a patrol crew sees it. Regulators now expect operators to explain detection latency when a spill report is late.
The better systems use pressure, flow, and volume comparisons. When the mass in does not equal the mass out within a defined window, the system raises an alarm. The window can be minutes, not days. That shift from patrol to telemetry is the reason leak detection became an engineering problem instead of just a field problem.
How OLO and RTM differ
Offline leak assessment is a batch calculation. It compares a completed time window and writes a conclusion after the event. Real-time monitoring watches the same flow and pressure signals during operation and raises alarms before the window closes. OLO is useful for audit trails, trend analysis, and post-event review. RTM is useful when a line cannot afford the volume loss before the next shift.
Operators using both often let RTM protect the active window and use OLO to validate suspected events. The combination is stronger than either alone because OLO can catch slow rate changes that RTM suppresses as noise, while RTM catches fast events before OLO finishes its window.
Sensor, SCADA, and rule-set choices
SCADA provides the signals, but the rule set decides whether those signals look like a leak or normal operation. A rule set that is too loose alarms during normal line pack changes. A rule set that is too tight misses small leaks. The common approach is a layered model: flow imbalance, pressure step change, and acoustic or negative pressure wave. Layering reduces false positives without removing fast detection.
Sensors should match the commodity. Crude lines behave differently from refined products and from gas. Density, pressure wave speed, and hydrate risk all change the calibration. Sensor placement matters too. Close spacing improves detection but raises cost. The usual mistake is installing sensors to protect equipment instead of protecting the line itself.
Closing the loop with dispatches and tickets
An alarm is only useful if the right person responds with the right check. The best programs convert an alarm into a dispatch note with location, magnitude, suspected product, and isolation point. That note becomes a field ticket. The ticket becomes a supervisor sign-off. The sign-off becomes compliance evidence.
Paper tickets lose that chain. Start with an operations audit if tickets still live in notebooks and email threads. The same discipline applies here as in LDAR and methane leak repair and ground disturbance and utility strikes. Close the loop once, and the next alarm gets faster.