Aircraft Malfunction Probability Analysis Through Causal Data Refinement
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing learning systems struggle to establish reliable correlations for rare phenomena, such as aircraft malfunctions, due to insufficient data, leading to ineffective database reduction or distorted statistics when attempting to predict maintenance needs.
Innovation Solution
A method involving data processing steps to refine input data by grouping by root cause, applying causal time windows, weighting, filtering, and Boolean value assignment to determine a probability of malfunction occurrence, enabling accurate maintenance actions.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Productivity
If the database size is reduced to contain only 50% effective incidents, then the learning system can process the data more efficiently, but the database size becomes too small to train the learning system reliably
Solution Approach 1:
The patent segments the database into multiple subsets, each containing a different proportion of effective incidents (e.g., 50%, 60%, 70%, 80%, 90%). This allows the system to process data in manageable segments while collectively maintaining sufficient training data across all subsets to reliably train the learning system.
Solution Approach 2:
The patent introduces a new dimension to the problem by creating multiple database subsets with varying compositions rather than working with a single reduced database. This dimensional transformation allows simultaneous optimization of both processing efficiency (through smaller subsets) and training reliability (through the collective diversity of all subsets).
2Quantity of substance
If cases are artificially cloned to increase their number in the database, then the learning system has more data to analyze, but the statistics of malfunction occurrence become distorted
Solution Approach 1:
The patent dynamically adjusts the composition of database subsets, creating different proportions of effective incidents and non-effective incidents across subsets (e.g., 50/50, 60/40, 70/30 splits). This dynamic approach allows the system to work with sufficient case numbers in each subset while maintaining overall statistical accuracy through the varied composition across all subsets.
3Productivity
If the database contains too few cases to study, then the learning system cannot establish reliable correlations, but reducing the database further to improve processing efficiency makes training impossible
Solution Approach 1:
The patent divides the limited database into multiple subsets, allowing the learning system to process each subset efficiently while collectively utilizing all available data across subsets for training. This segmentation enables both efficient processing of individual subsets and sufficient aggregate data for reliable correlation establishment.
Data Source
Figure 1
Figure 2
AI summary
A method (100) for determining the probability of a malfunction (Mf) occurring in an aircraft that causes a performance defect processes a set of input data to retain only the measurements relevant to determining the probability of the malfunction occurring following a prior event. In other words, the method refines a set of input data to establish relevant causal links between measurement readings and a malfunction, thereby determining the probability of a malfunction occurring based on the occurrence of one or more prior events.