Aircraft Engine Defect Identification via Anomaly Vectors
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current methods for identifying failures in aircraft engines face challenges due to varying time measurements in different units, difficulty in applying scoring and classification tools in a multi-varied domain, and the scarcity and expense of large failure databases, making it hard to quickly and cost-effectively identify faults.
Innovation Solution
A method that uses reference vectors with a priori and a posteriori probability calculations, anomaly vector construction, and a decision grid based on expert knowledge to easily identify faults in a physical repository understandable by engine experts, allowing for quick and low-cost fault detection without relying on extensive databases.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If classification or labeling tools are used for statistical control of industrial processes, then fault identification capability is improved, but the requirement for large failure databases increases, which are very expensive and require significant calculation time
Solution Approach 1:
The patent creates simplified reference vectors that copy the essential characteristics of failure patterns without requiring complete failure databases. These reference vectors are constructed from standardized indicator vectors representing typical failure modes, allowing fault identification without extensive training data
Solution Approach 2:
The patent replaces expensive, large-scale failure databases with lightweight reference vectors that can be constructed from expert knowledge and a small number of actual failures. These reference vectors serve as disposable, easy-to-update representations of failure patterns that don't require maintaining large databases
2Adaptability or versatility
If scoring tools are used to substitute quality scores for time measurements, then unit variability is reduced, but the tools become difficult to apply in a multi-varied domain such as aircraft engine monitoring
Solution Approach 1:
The patent transforms diverse time measurements into standardized indicator vectors through normalization and standardization processes. This parameter transformation allows measurements in different units to be compared directly, making the system applicable across the multi-varied domain of aircraft engine monitoring while maintaining ease of operation
Solution Approach 2:
The patent creates a universal framework where standardized indicator vectors serve multiple functions: they represent normal operation, represent failure modes, and enable comparison across different engine parameters. This universal approach simplifies application across the diverse domain of aircraft engine monitoring
3Measurement precision
If real failure databases are collected for training classification tools, then fault detection accuracy is improved, but the cost and time required to build the database increases significantly
Solution Approach 1:
The patent performs preliminary construction of reference vectors using available expert knowledge and standardized indicators before actual fault detection is needed. This preliminary action creates ready-to-use reference patterns that can be immediately applied without requiring time-consuming database collection and processing
Solution Approach 2:
The patent enables the system to serve itself by constructing reference vectors from a small number of actual failures combined with expert knowledge. The system doesn't require extensive external database resources, as it can generate its own reference patterns from minimal data and domain expertise
Data Source
Figure 1
Figure 2~3
AI summary
The invention relates to a method and system for identifying defects in an aircraft engine (1), comprising: means (5) for defining a set of standardized indicators representative of the operation of the aircraft engine (1); means (5) for constructing an anomaly vector representative of engine (1) behaviour as a function of the set of standardized indicators; means (5), used when an abnormality is revealed by the anomaly vector, for selecting a subset of reference vectors having directions belonging to a pre-determined neighbourhood of the direction of the anomaly vector, said subset of reference vectors being selected from among a set of reference vectors associated with aircraft engine defects and determined according to criteria established by expert assessment; and means (5) for identifying the defects associated with said subset of reference vectors.