Artificial intelligence-based corrosion sensing and prediction for aircraft applications (AICorrSens)

Abstract

Corrosion causes enormous damage to mechanical structures in many industrial sectors, and the aviation industry is no exception. To extend the lifetime of airframes without compromising safety, it is very important to have a clear picture of the state of corrosion (SoC) of the aircraft. Thus, it is essential to develop methodologies suitable for real-time monitoring of the SoC and subsequent reliable notification when a structure has been compromised by corrosion. Published results so far suggest that the ultrasonic (e.g. acoustic emission, guided waves) as well as electrochemical sensors (e.g. electrochemical noise, impedance spectroscopy) are suitable for monitoring aircraft-relevant corrosion but lack the technological readiness to be applied in commercial aircraft yet. A huge issue in achieving reliable monitoring systems is the correlation between corrosive phenomena and (typically) noisy sensor data. The AICorrSens project addresses these issues by developing a multisensor setup for monitoring the SoC based on ultrasonic, electrochemical, and environmental sensors coupled with AI algorithms. Training data shall be generated by performing accelerated corrosion tests with coupons and demonstrator parts equipped with sensors. Using AI for the subsequent data analysis, one can overcome operational noise, and thus, allow today’s corrosion detection methods onboard real- time evaluation of the SoC in terms of detection, localization, quantification, and typification. The ambition of the project is to transform the created continuous stream of data into classifications of the SoC that are intuitively understandable through a human-machine interface, including a qualified corrosion prediction by the AI models generated from test campaigns. The project results shall lead to increased aircraft safety and reliability and deliver a clear economic benefit for aircraft operators as it allows a switch from regular inspection intervals to condition-based maintenance. Funded by: Austrian Research Promotion Agency Program: Take Off, Call 2019 Consortium: CEST Competence Centre for Electrochemical Surface Technology (CEST), Johannes Kepler University Linz – Institute of Structural Light-weight Design (IKL), Danube University Krems – Department for Integrated Sensor Systems (DISS), Senzoro GmbH (SENZ). Project duration: 10/2020 – 09/2023. EWGAE 35, Ljubljana, Slovenia, 13th – 16th Sep. www.ewg