Aerial Vehicle Engine Health Prediction Using Machine Learning
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing engine health monitoring systems are inefficient in recalculating coefficients for power assurance checks, leading to outdated assessments and delayed detection of engine health issues, which can result in operational challenges for aerial vehicles.
Innovation Solution
A method and system that utilize machine learning techniques to rapidly generate and update coefficients for a power assistance check (PAC) using engine acceptance test procedure data and flight test data, allowing for real-time or near-real-time prediction of engine health, enabling more frequent and accurate assessments.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If traditional methods are used to recalculate coefficients for power assurance checks, then the system is simpler to implement, but the assessment becomes outdated and detection of engine health issues is delayed
Solution Approach 1:
The patent replaces traditional mechanical calculation methods with machine learning models that can rapidly process engine data and generate updated coefficients. The neural network architecture processes flight test data and ATP data to produce recalculated coefficients in minutes rather than hours or days, significantly reducing the time loss while improving assessment accuracy through pattern recognition capabilities.
Solution Approach 2:
The patent changes the computational parameters by using machine learning algorithms that can process and analyze large volumes of engine data much faster than traditional methods. The system transforms the coefficient recalculation process from a time-intensive computational task to a rapid automated process, reducing the time delay while maintaining or improving precision through advanced data processing capabilities.
2Productivity
If machine learning techniques are used to generate coefficients rapidly, then the prediction of engine health becomes more accurate and timely, but the system complexity increases
Solution Approach 1:
The patent extracts the complex machine learning model training and coefficient generation process into a separate, standalone system that can be executed independently. By separating the computational complexity from the core engine monitoring system, the patent enables rapid coefficient recalculation through specialized algorithms while keeping the main system architecture relatively simple and manageable.
Solution Approach 2:
The patent introduces an intermediary layer consisting of pre-trained machine learning models that act as mediators between raw engine data and the power assurance check coefficients. These intermediary models handle the computational complexity of pattern recognition and coefficient generation, allowing the main system to benefit from rapid, accurate predictions without directly managing the complex algorithms themselves.
3Reliability
If traditional coefficient recalculation methods are used, then the system is easier to operate, but engine health issues are detected later
Solution Approach 1:
The patent implements self-service capabilities where the machine learning system automatically processes flight test data and ATP data to generate updated coefficients without requiring manual intervention. The system autonomously performs data collection, processing, model updating, and coefficient generation, improving reliability through continuous automated monitoring while reducing operational complexity through automation of previously manual tasks.
Solution Approach 2:
The patent incorporates feedback mechanisms where the machine learning models continuously learn from new flight test data and update their predictions accordingly. This feedback loop ensures that the system adapts to changing engine conditions and improves its reliability over time, while the automated nature of the feedback process maintains ease of operation by eliminating the need for manual system reconfiguration.
Data Source
AI summary
Systems and methods for modeling engine health are provided. One example aspect of the present disclosure is directed to a method for modeling engine health. The method includes receiving, by one or more processors, engine acceptance test procedure (ATP) data. The method includes receiving, by the one or more processors, flight test data. The method includes generating, by the one or more processors, one or more coefficients for a power assistance check (PAC) based on the engine ATP data and the received flight test data using a machine learning technique. The method includes transmitting, by the one or more processors, the one or more coefficients for the PAC to a vehicle, wherein the vehicle uses the one or more coefficients in the PAC to predict engine health.