FOD Detection Using Time-Series Anomaly Filtering
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Gas turbine engines face challenges in detecting Foreign Object Debris (FOD) damage effectively due to high false alarm rates from existing detection systems, which complicates periodic inspections and maintenance.
Innovation Solution
A system and method utilizing a controller to process time-series data from debris monitoring sensors, involving pre-processing, anomaly detection, and a FOD damage model to accurately determine FOD events and generate health reports, reducing false alarms by analyzing sensor integrity and data alignment.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Loss of time
If FOD detection systems and sensors are used to reduce inspection frequency, then inspection time is reduced, but false alarm rate increases
Solution Approach 1:
The detection system is segmented into multiple independent components: debris monitoring sensors capture raw signals, pre-processing filters clean and align the data, anomaly detection algorithms identify deviations, and FOD damage models confirm actual damage. This segmentation allows each component to specialize, reducing false alarms while maintaining continuous monitoring capability that reduces inspection time.
Solution Approach 2:
Multiple intermediary processing layers are introduced between the sensor and the final detection decision. The pre-processing step acts as an intermediary to clean and align raw sensor data, the anomaly detection model serves as an intermediary to identify potential issues, and the FOD damage model acts as a final intermediary to confirm actual damage. These intermediaries filter out false signals while preserving true FOD events.
2Measurement precision
If complex periodic inspections are conducted to mitigate FOD damage risk, then detection accuracy is improved, but maintenance time increases
Solution Approach 1:
The system performs preliminary continuous monitoring and anomaly detection between scheduled inspections. Debris monitoring sensors continuously capture data, and anomaly detection algorithms preliminarily identify potential FOD events before they cause significant damage. This preliminary action allows inspections to be targeted and reduced in frequency while maintaining high detection accuracy.
Solution Approach 2:
The patent replaces complex mechanical inspection procedures with automated sensor-based monitoring and algorithmic analysis. Debris monitoring sensors continuously monitor for FOD events, and computer algorithms automatically analyze the data to detect anomalies and confirm damage. This substitution eliminates the need for frequent manual disassembly and physical inspection, reducing maintenance time while improving detection accuracy through continuous data collection.
Data Source
AI summary
A system and method for detecting foreign object debris damage is disclosed. A method for foreign object debris detection in a gas turbine engine may include receiving, by a controller, a first time-series data from a database, wherein the first time-series data comprises a feature, pre-processing, by the controller, the first time-series data to generate a second time-series data, generating a third time-series data via an anomaly detector model, sending, by the controller, the third time-series data to a foreign object debris (FOD) damage model, and determining, by the controller, that a FOD event has occurred based on data received from the FOD damage model. In various embodiments, the method may further comprise generating, by the controller, a health report (HR).