Spatial anomaly detection with optimal transport

TL;DR

This paper presents an automated spatial anomaly detection framework for jet engines using optimal transport theory, effectively identifying temperature anomalies across engine fleets with high accuracy.

Contribution

It introduces a novel anomaly detection method based on optimal transport for Gaussian measures, tailored for spatial temperature data in jet engines, with demonstrated cross-engine applicability.

Findings

Successfully detects anomalies in engine temperature data

Reduces false positives and negatives in anomaly detection

Applicable to other thermodynamic measurements

Abstract

This manuscript outlines an automated anomaly detection framework for jet engines. It is tailored for identifying spatial anomalies in steady-state temperature measurements at various axial stations in an engine. The framework rests upon ideas from optimal transport theory for Gaussian measures which yields analytical solutions for both Wasserstein distances and barycenters. The anomaly detection framework proposed builds upon our prior efforts that view the spatial distribution of temperature as a Gaussian random field. We demonstrate the utility of our approach by training on a dataset from one engine family, and applying them across a fleet of engines -- successfully detecting anomalies while avoiding both false positives and false negatives. Although the primary application considered in this paper are the temperature measurements in engines, applications to other internal flows and related thermodynamic quantities are made lucid.