Aircraft Maintenance Message Prediction for Early Fault Planning

Resolve Bottlenecks,
Find Innovative Solutions
Generate Solutions

Solution Overview

Problem

Current aircraft maintenance systems fail to provide sufficiently early warnings of impending undesired conditions, limiting the flexibility of maintenance responses and often requiring immediate action without allowing for proactive planning or substitution of aircraft.

Innovation Solution

A method that stores and processes maintenance messages from multiple flight legs to predict future maintenance needs using a machine learning algorithm, providing several days' lead time for maintenance crews to prepare and allowing for more flexible fleet management options such as aircraft substitution or rerouting.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Loss of time

If constant monitoring and early detection systems are implemented, then maintenance personnel can be notified of faults prior to the next flight, but the warning time is insufficient to allow for flexible maintenance responses such as aircraft substitution or rerouting

Engineering Contradiction:
Improvemaintenance preparation timeVSAvoidprediction accuracy
Core Design Contradiction:
Loss of timeVSReliability

Solution Approach 1:

The system performs preliminary analysis of maintenance messages and flight parameters to predict future faults before they occur. By analyzing historical data and current trends, the system generates predictions that allow maintenance personnel to take preparatory actions days in advance, such as scheduling maintenance during planned downtime, arranging aircraft substitution, or rerouting flights, rather than reacting to faults at the last minute.

Inventive Principle:
Principle #10Preliminary action

Solution Approach 2:

The system transitions from traditional real-time fault detection (single time dimension) to multi-dimensional prediction by incorporating historical maintenance messages, flight parameters, and temporal patterns. This allows the system to predict faults not just for the next flight but several days in advance, adding a temporal dimension that enables more flexible maintenance planning and response options.

Inventive Principle:
Principle #17Another dimension (Dimensionality change)

2Adaptability or versatility

If traditional fault detection systems are used, then faults are detected when they occur or immediately before the next flight, but maintenance personnel cannot perform proactive planning or substitute aircraft

Engineering Contradiction:
Improvemaintenance response flexibilityVSAvoidground time
Core Design Contradiction:
Adaptability or versatilityVSLoss of time

Solution Approach 1:

The system performs preliminary analysis of maintenance messages and flight parameters to predict future faults before they occur. By analyzing historical data and current trends, the system generates predictions that allow maintenance personnel to take preparatory actions days in advance, such as scheduling maintenance during planned downtime, arranging aircraft substitution, or rerouting flights, rather than reacting to faults at the last minute.

Inventive Principle:
Principle #10Preliminary action

Solution Approach 2:

The system dynamically adjusts prediction horizons and maintenance recommendations based on changing flight schedules, aircraft utilization, and emerging fault patterns. This dynamic approach allows the system to optimize the balance between early warning and operational flexibility, adapting maintenance strategies to current fleet conditions and enabling more versatile response options.

Inventive Principle:
Principle #15Dynamics

Data Source

PatentEP3462266B1Aircraft maintenance message prediction
Publication Date: 2022.05.04 THE BOEING CO
  • EP3462266B1 patent drawingFigure 1
  • EP3462266B1 patent drawingFigure 2~3
  • EP3462266B1 patent drawingFigure 4

AI summary

A method and apparatus for maintaining a vehicle (104), such as an aircraft (100). A plurality of maintenance messages (120) generated during operation of the vehicle (104) are stored to form a plurality of stored maintenance messages (134). The stored maintenance messages (134) are filtered to remove from the stored maintenance messages (134) those maintenance messages (120) that are correlated to minimum equipment list actions (124) to form filtered stored maintenance messages. A predicted maintenance message (130) is generated from the filtered stored maintenance messages by applying a machine learning algorithm (144) to the filtered stored maintenance messages. The predicted maintenance message (130) may be used to perform a maintenance operation (128) on the vehicle (104).