AMLD Ranking Test Platform

The AMLD (Advanced Mobile Leak Detection) Ranking Test is an analytical validation platform that tests a core operational question: can sensor signals from mobile methane survey vehicles predict the severity of a gas leak before a field crew investigates? The platform scores 985 historical indications using a Leak Severity Index and a Dispatch Priority Score, then validates those scores against actual field-grade outcomes.

The answer the data produced was unexpected. Sensor signal strength has near-zero correlation with leak grade. Seventy-eight percent of the most serious leaks — Grade 1 — produced sensor readings below the median on every signal. The sensor finds the gas reliably. It cannot grade the leak from the road. That finding changed the framing of the entire project: from predicting severity to prioritizing dispatch. The platform shifted to a three-tier dispatch framework based on ethane detection and proximity to critical structures — schools, hospitals, fire stations — rather than signal strength.

The platform hosts nine analytical tabs: scoring and ranking, an interactive map, validation analytics, signal analysis, divergence analysis, a dispatch priority view with tier assignments, and a crew assignment module with lasso-select mapping for field supervisors. It accepts both pre-investigation and post-investigation data files. The project includes a full white paper written for gas utility operations managers, consultants, and regulators.