Aircraft Engine Monitoring Data Standardization
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current methods for standardizing aircraft engine monitoring data fail to account for dependencies on external context and relationships between indicators, leading to inefficient anomaly detection and interpretation challenges for engine experts.
Innovation Solution
A method that collects and processes time measurements to calculate specific indicators, identifies exogenous data, and defines a conditional multidimensional model to standardize these indicators, removing dependencies on external context and managing stochastic interdependence relationships between indicators, using projection spaces and regression techniques to normalize data into standardized values.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Productivity
If classic normalization method is used to standardize monitoring data, then calculation is simple and fast, but it cannot manage dependencies on exogenous data nor manage dependency relationships between indicators
Solution Approach 1:
The patent transforms the standardization approach by changing from simple statistical parameters (mean, standard deviation) to a multidimensional conditional model that incorporates exogenous data parameters. This allows the system to account for external conditions (weather, route, piloting) while maintaining computational efficiency through pre-built conditional models.
Solution Approach 2:
The patent introduces conditional multidimensional models as intermediaries between raw monitoring data and standardized values. These models act as mediators that process both indicator dependencies and exogenous data relationships, enabling comprehensive data management without requiring complex real-time calculations.
2Reliability
If multivariate normalization modes from PCA algorithms are used, then dependencies on exogenous data and relationships between indicators are managed, but significant calculation time is required and dimensionless indicators cannot be interpreted by engine experts
Solution Approach 1:
The patent applies preliminary action by pre-building conditional multidimensional models during offline preparation phases. These pre-built models capture the relationships between indicators and exogenous data without requiring complex real-time calculations during actual monitoring, thus reducing computational time while maintaining dependency management capabilities.
Solution Approach 2:
The patent transforms PCA's dimensionless indicators into interpretable standardized values by changing the parameter representation. Instead of abstract principal components, the system produces standardized values that maintain physical meaning and can be directly interpreted by engine experts, while still capturing multivariate relationships.
3Adaptability or versatility
If standardization removes dependencies on external context, then monitoring operates identically under varying conditions, but interpretation by engine experts becomes more difficult
Solution Approach 1:
The patent changes the parameter transformation approach to preserve physical meaning. Instead of creating abstract standardized values, the system produces standardized indicators that maintain their physical interpretation while being adjusted for external conditions. This allows experts to understand the standardized values in terms of actual engine parameters and their deviations from expected behavior.
Data Source
Figure 1
Figure 2~3
Figure 4~5B
AI summary
The invention relates to a method and system for standardizing data used for monitoring an aircraft engine (1) comprising: - means (5) for gathering in the course of time temporal measurements on said aircraft engine (1); - means (5) for calculating on the basis of said temporal measurements a set of indicators Y = (Yi,..., Yi...,Ym) specific to elements of said engine; - means (5) for identifying on the basis of said temporal measurements a set of exogenous data X = (*,,...,*") representative of the exterior context intervening on said set of indicators F; - means (5) for defining a conditional multidimensional model which simultaneously manages the indicators of said set of indicators F while taking account of said set of exogenous data X so as to form a set of estimators Y = (yi,..., y],..., ym) corresponding to said set of indicators Y = (y?,...,yi,...,yn); and - means (5) for normalizing each estimator y} as a function of a reference value of the corresponding indicator y} and of a deviation between said each estimator Yi} and said corresponding indicator y} so as to form a set of standardized values F = (~yi,..., ~yJ,...,~ym).