ENDPOINT INTEGRITY· pipeline leak detection
Computational Pipeline Monitoring · API 1130 compliant

Detect a pipeline leak in seconds. Locate it on a map. Move on with your day.

Endpoint Integrity samples upstream and downstream pipeline pressure at 100 Hz, runs a cross-stream ML model per segment, and pinpoints leak GPS coordinates from negative pressure wave arrival timing. EI Sentinel field hardware . Deployed at midstream scale across 19 operators.

Schedule a demo How it works
0+
Field devices deployed
0 Hz
Pressure sample rate
0 B/day
Pressure samples ingested
0+
Midstream operators
0%
False-alarm reduction
Live trace · real customer pipeline

This is what the model sees.

Six seconds of real 100 Hz pressure data from a producing pipeline endpoint, followed by a synthetic negative-pressure-wave signature representative of a small leak. The dashed green band is the EI model's expected envelope — pressure outside it triggers an alert with magnitude, duration, and (when both endpoints see it) GPS coordinates.

The cost of detecting a leak late

A single undetected release reorders your year.

PHMSA records hundreds of significant hazardous-liquid pipeline releases every year. The averages are unforgiving — and that's before reputational damage, downtime, and your operating certificate going under enforcement review.

$3.5M

Average release cost

Aggregate operator cost per significant hazardous-liquid release — cleanup, lost product, downtime, third-party damages — per PHMSA 2019–2023 incident data.

$250,438

Daily PHMSA civil penalty

Maximum civil penalty per day per violation (2024 49 U.S.C. § 60122 inflation adjustment). Each undetected day of a release compounds.

CAP

Corrective Action Order

A PHMSA CAP order can require an immediate pressure reduction or full shutdown until a Leak Detection Program meeting API RP 1175 is in place.

What Endpoint Integrity does

Detect. Locate. Integrate.

Three things every midstream control room needs from a leak-detection system, and the only three things we spend engineering on.

⚡

Sub-minute detection

ML models evaluate pressure deviation against learned pipeline dynamics on every sample. Genuine pressure-signature events surface in seconds — not the minutes-to-hours volume-balance systems require.

📍

GPS leak localization

Cross-correlation of negative pressure wave arrival at upstream and downstream sensors converts pressure data into a point on a map. The suspected leak coordinates render directly on the operator's pipeline route.

⇄

SCADA-native integration

Modbus TCP, OPC UA, and REST endpoints — alarm status, pressure values, pipeline health — mapped to the registers your existing control system already polls. No screen-scraping, no replacement console.

Case study (anonymized)

From 6,800 nuisance alerts a day to three actionable ones — in four days.

A Tier-1 midstream NGL operator deployed Endpoint Integrity on a multi-segment pipeline where the legacy SCADA-threshold system was generating thousands of daily nuisance alarms — the textbook setup for an operator missing the one that matters.

Before EI (legacy SCADA thresholds)

6,800 / day
Volume-balance alarms, almost entirely operational noise. Real signals indistinguishable from pump-start transients.

After 4 days of EI deployment

3 / day
Cross-stream ML learned the pipeline's transfer function. Only genuine pressure-signature events surface to the control room.

Equivalent to a 99.96% reduction in operator alert load — and a control-room team that can act on every alert instead of triaging an inbox.

Technology preview

The cross-stream ML architecture you can't replicate with a thresholding rule.

Most leak-detection systems look at a single sensor and flag deviation from a tuned threshold. EI trains a deep model per segment that takes upstream and downstream pressure as input — so it learns the pipeline's transfer function, the way pressure waves actually propagate through this specific pipe at this specific flow.

Pump starts, valve transients, pig launches, and seasonal viscosity changes stop being alarms. Pressure-drop signatures consistent with a real release start being the only thing on the screen.

Architecture deep dive
Negative pressure wave arrival timing diagram
Regulatory posture

Built to the standards your auditor already references.

Endpoint Integrity is a Computational Pipeline Monitoring (CPM) system implemented to the algorithmic, alarm, performance, and integration requirements that PHMSA inspectors and your insurance underwriter expect on day one.

49 CFR § 195.134 / .444

Hazardous Liquid Pipeline CPM

Operator must install a CPM system on category-1 lines and follow API 1130. EI satisfies the CPM algorithm + performance test record-keeping requirements.

API 1130

Computational Pipeline Monitoring

Industry-recognized algorithm classification + design + operation + maintenance + testing standard. EI implements the Algorithm Description + Performance Test obligations directly.

API RP 1175

Leak Detection Program Management

Pipeline Leak Detection Program (PLDP) management lifecycle. EI provides the operator-facing artifacts needed for the audit binder.

PIPES Act 2020

Federal mandate

PHMSA Final Rule extends leak detection requirements to natural gas gathering and transmission lines. EI's pressure-based approach is product-agnostic.

Full compliance map

Show us a leak you couldn't catch.

Send us a recent SCADA pressure trace from a segment you suspect is leaky — or one you know shouldn't be. We'll run it through the EI model and tell you what the algorithm sees within 48 hours, no commitment.

Schedule a demo Run the ROI calculator