Aircraft Engine Balancing via Neural Network Vibration Prediction
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
High-bypass turbofan engines in commercial aviation face significant challenges in minimizing vibrations due to inherent imbalances, leading to excessive cabin noise and vibration during flight, necessitating frequent balancing operations that are time-consuming and costly.
Innovation Solution
The method employs a multiplicity of artificial neural networks to predict engine vibrations based on design variable inputs, processing these predictions with a pattern matching algorithm to identify optimal balance weight configurations that minimize vibrations, allowing for efficient reconfiguration of balance weights to reduce engine vibrations.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Productivity
If traditional balancing methods are used with manual selection of balancing screws, then the balancing process can be performed, but it requires multiple flight tests and is time-consuming
Solution Approach 1:
The patent replaces manual mechanical balancing procedures with an automated computer-based system that uses artificial neural networks to predict vibrations and calculate optimal balance weight configurations. The system automatically processes flight data, runs simulations, and determines balancing solutions without requiring multiple manual flight tests, thereby significantly improving productivity and reducing time loss.
Solution Approach 2:
The patent creates a virtual copy of the aircraft engine through computational models and artificial neural networks. This digital replica allows simulation of various balancing scenarios and prediction of vibration characteristics without requiring physical flight tests for each configuration, enabling efficient optimization of balance weight settings.
2Object-affected harmful factors
If multiple flight tests are conducted to achieve acceptable vibration levels, then vibration reduction can be achieved, but operational costs increase
Solution Approach 1:
The patent performs preliminary vibration prediction and balancing optimization using artificial neural networks and computational models before actual flight tests. By pre-calculating optimal balance weight configurations and predicting vibration characteristics, the system identifies the minimum number of flight tests required to achieve acceptable cabin noise and vibration levels, thereby reducing operational costs while maintaining effectiveness.
Solution Approach 2:
The patent implements a feedback loop where flight data is collected, processed through artificial neural networks, and used to refine vibration predictions and balancing solutions. This iterative feedback process optimizes the balancing configuration with minimal flight tests, reducing the energy and resources required while achieving effective vibration reduction.
3Object-affected harmful factors
If balance weights are reconfigured based on flight test results, then vibration reduction is achieved, but the process becomes complex and time-consuming
Solution Approach 1:
The patent replaces complex manual balancing operations with an automated computer system that uses artificial neural networks to process flight data, predict vibrations, and calculate optimal balance weight configurations. This substitution simplifies the balancing process by automating calculations and decision-making, reducing operational complexity while effectively reducing engine vibrations.
Solution Approach 2:
The patent systematically varies balancing parameters such as balance weight magnitude, position, and phase angle through automated calculations driven by artificial neural networks. By optimizing these parameters based on predicted vibration characteristics rather than trial-and-error flight tests, the system reduces the complexity of the balancing operation while achieving effective vibration reduction.
Data Source
AI summary
A method for balancing an aircraft turbofan engine using a respective artificial neural network for each flight of multiple reference aircraft to predict the engine vibrations that would be produced in response to input of flight parameter and balance weight data acquired from a flight of a test aircraft of the same type. The resulting sets of predicted vibrations are then processed to identify and collect those engine vibration predictions which match or nearly match the engine vibration measurements acquired from the test aircraft during its flight test. For each matching modeled flight of the reference aircraft, the magnitude and phase angle of the mass vector for any commonly used balance weight configurations are determined and included in a list of recommended balance weight configurations. A technician can then reconfigure the balance weights attached to the engine on the test aircraft in accordance with a selected recommendation.