Prediction of Remaining Useful Lifetime in Composite-based Products

Collaborative research with Hamburg University of Technology & Swinburne University

The number of products and structures made of composite materials is set to keep increasing in the upcoming years, due to their low weight and high-performance mechanical properties. Composite materials have become the foremost choice to increase the capability of systems with less weight in many industries: robotics, aviation, hydrogen storage …

With such a massive shift, ensuring the safety and reliability of composite-based products is becoming paramount. With an ever-increasing demand for efficiency and performance in high-end applications, the integrity of these often-critical products cannot be compromised. For such high-performance applications, end-users are increasingly in need real-time monitoring and access to product life data for determining remaining life. The University of Melbourne, Hamburg University of Technology and Swinburne University of Technology recognized and have research underway to provide solutions.

A key issue is predicting when composite components will fail, to optimally schedule their maintenance or replacement. However, the time-to-failure of a composite exhibits extreme variability depending on its manufacturing, operation and environment. By keeping track of different physical properties of the products in real-time, remaining useful lifetime can be accurately predicted, and maintenance or replacement decisions can be taken accordingly.

Our research explores the crossroad of composite product design, manufacture, big data and artificial intelligence in predicting remaining useful life. The Universities’ team is working on the design of a sensing network collecting structural data in real-time. Collected data is evaluated by machine learning and deep learning algorithms, which compute in real-time the remaining useful lifetime of the product.

Moreover, the collected data will be utilized to identify and diagnose product failures, offering end-users deeper insights into the condition and performance of their products.

This research is currently supported by the German Federal Ministry of Education and Research.