THE PROJECT

Description

The project aims to create a physical and virtual laboratory for the monitoring and diagnosis of light aircraft. Heterogeneous measurements are processed through Data Fusion techniques and integrated with simulations of a Digital Twin. Data classification algorithms provide indicators of the aircraft’s health status to promptly identify any faults. The case study is an ultralight helicopter, for which a simplified test bench is created to validate the procedure.

Objectives

  • Select appropriate sensors for the type of component being monitored
  • Build Data Fusion algorithms that integrate heterogeneous signals from aircraft
  • Develop Machine Learning techniques for monitoring, diagnosis, and prognosis of aircraft by integrating Data Fusion techniques
  • Develop the Digital Twin of the aircraft, integrating a finite element model, a lumped parameter model of the transmission system, and a motion control model under various flight conditions

Phases

The project is divided into several phases:
Selection and installation of additional sensors: Analysis of the system to identify the sensors to be integrated with those already on-board to acquire all the necessary data for comprehensive aircraft monitoring.

Integration of experimental data through Data Fusion techniques: The monitored signals are heterogeneous, as they are acquired through different types of sensors, and must therefore be processed to obtain a homogeneous dataset on which classification algorithms can be effectively applied.

Development of Machine Learning techniques: Definition of classification algorithms and ad-hoc indicators capable of monitoring and diagnosing the health status of the aircraft, based on representative features.

Creation of the aircraft's Digital Twin: Advanced mathematical models, used to simulate the aircraft in various domains, are integrated to create a Digital Twin of the aircraft. This will allow real-time simulations to diagnose and prognosticate structural issues, simulate damages, and optimize design.

Setup of the test bench: The effectiveness and reliability of the developed techniques and systems are validated through tests conducted on a test bench specifically created as a simplified mockup of the helicopter chosen as the case study.

Results

  • Sensorize the aircraft
  • Develop Machine Learning techniques for monitoring, diagnosis, and prognosis of aircraft by integrating Data Fusion techniques
  • Develop the Digital Twin of the aircraft, integrating a finite element model, a lumped parameter model of the transmission system, and a motion control model under various flight conditions
  • Prototype test bench and validated Digital Twin of the test bench
  • Simulated dataset of signals in the presence of damaged components
  • Real-time diagnosis and prognosis system based on the Digital Twin
  • Design methodology for optimizing aircraft design to achieve weight reduction and lower fuel consumption