About the course
The progress in key-technologies such as Artificial Intelligence and Internet of Things opens new horizons for developing the lightweight structures of the future, where the structures will be able to perform self-diagnostic checks of their integrity, self-estimate their lifespan and communicate with each other valuable information.
This course aims to provide the fundamental knowledge for enabling these key-technologies to transform conventional structures to cyber-physical assets. The lectures will focus on the application of machine learning for design and failure analysis of lightweight structures, AI-based structural health monitoring, diagnostics and prognostics strategies, state awareness capabilities and digital twins.
This course is aimed for young researchers (PhD’s and PostDoc’s) who want to become independent scientists developing their research in the field of design and failure analysis of lightweight structures using machine learning, structural health monitoring, diagnostics and prognostics and state awareness. The attendants are expected to:
- have strong knowledge of mechanics of materials (preferably FRP materials)
- be able to apply theory of probability and statistics
- understand the fundamentals of the finite element method
- be familiar with machine learning techniques
During the course, the attendants should expect interactive lectures consisting of classroom activities and concept exercises. A special ‘meet-your-lecture’ session will take place where the attendants will consult the lecturers about their research topics.
We will provide 3 scholarships to 3 applicants and cover their registration fees. The attendants are welcome to submit their CVs and a motivation letter to Dr. Dimitrios Zarouchas (email@example.com), addressing the relevance of their research to the content of the course. A committee will assess their applications.
All attendants will receive a certificate of attendance.
Can We Learn Without Failure? – a critical reflection on education and the learning process
Dr. Calvin Rans received his PhD in Aerospace Engineering from Carleton University in 2007 for his work in the field of fatigue crack growth in riveted lap joints. Subsequently he continued his research efforts in fatigue and damage tolerance studying fatigue crack growth in hybrid metal-composite laminates and additively manufactured materials at Delft University of Technology where he currently holds an associate professor position. Applying what he has learned in research to education, Calvin has become an avid educational specialist, earning the title of Docent van het Jaar (Teacher of the Year) for the entire Netherlands in 2019 and becoming visible in open and online education both nationally and within Europe as a whole.
Hybrid neural networks for damage predictions of composite substructures
I joined the group of Aerospace Structures and Computational Mechanics in TU Delft Aerospace Faculty as an Assistant Professor in April 2017. Before coming to TU Delft, I had worked as a Scientist in the Institute of High Performance Computing in Singapore and as a Research Fellow in the Department of Mechanical Engineering of National University of Singapore (NUS). I obtained my PhD jointly from NUS and Imperial College London in March 2014. I also hold the Diplôme d'Ingénieur from Ecole Polytechnique and the Bachelor of Engineering from NUS under a Double Degree program.
My research focuses on the numerical modelling of damage and failure in fibre-reinforced composites. My main work in the past is the development of Floating Node Method (FNM) for the modelling of crack propagations in composites. FNM has been adopted by many organizations, such as NASA Langley Research Center, National Institute of Aerospace and Boeing in the US, Volvo Cars and Swerea SICOMP in Sweden, etc. Recently, I have been exploring the multi-disciplinary combination of machine learning and finite element analysis for the substructuring modelling of composite structures.
Open-source Framework for Data-Driven Design and Analysis of Structures and Materials (F3DASM)
Miguel Bessa is an Associate Professor in the Materials Science and Engineering Department at the Delft University of Technology. He is the Director of an inter-faculty Artificial Intelligence lab called MACHINA, dedicated to machine intelligence advances for materials design. He is also the recipient of a Veni personal grant (2019). Prior to coming to the Netherlands, he was a postdoctoral scholar in Aerospace at the California Institute of Technology, and he received his PhD (2016) in Mechanical Engineering at Northwestern University as a Fulbright scholar. He envisions a new era of design of materials and structures through artificial intelligence.
Data-driven awareness and monitoring of next generation intelligent structures
Professor Fotis Kopsaftopoulos joined the Department of Mechanical, Aerospace, and Nuclear Engineering at Rensselaer Polytechnic Institute as a tenure-track Assistant Professor in October 2017. He received his Diploma and Ph.D. in Mechanical and Aeronautical Engineering from University of Patras, Greece, where he worked on stochastic system identification, statistical signal processing, and vibration-based probabilistic methods for structural health monitoring (SHM). After completing his Ph.D. in 2012, he served as a Postdoctoral Research Associate in the Stochastic Mechanical Systems and Automation Laboratory of the same Department working on research related to novel personal plane concepts, aircraft 4D trajectory monitoring via adaptive time-varying models, and 4D model predictive control for “green” aircraft guidance. Before joining Rensselaer, he served as a Postdoctoral Scholar in the Department of Aeronautics and Astronautics at Stanford University, in the Structures and Composites Laboratory, (April 2013 to September 2017) where we worked on the design, analysis, and integration of intelligent structures and autonomous systems with bio-inspired state sensing and awareness capabilities, as well as novel SHM methods. Professor Kopsaftopoulos has published more than 100 journal and conference papers and 4 book chapters. He has participated in various research projects funded by Government and industrial sponsors (AFOSR, NSF, Army, NASA, ARPA-e, Boeing, Airbus). He serves as an Associate Editor in the Editorial Board of the Structural Health Monitoring International journal (SAGE Publications) and Topic Editor of the Sensors journal (MDPI). He is also a co-Editor of the International Workshop on Structural Health Monitoring (IWSHM) 2015, 2017 and 2019 Proceedings. Professor Kopsaftopoulos is a member of the Organizing Committee of the International Workshop on Structural Health Monitoring (IWSHM), secretary of the Technical Health and Usage Monitoring (HUMS) Committee of the Vertical Flight Society (VFS), and a member of the Program Committee of the InfoSymbiotics/Dynamic Data Driven Application Systems. He has also served as Guest Associate Editor for Special Issues in the Structural Health Monitoring and Aerospace journals.
Coupling of digital-twins and machine learning for model-based diagnosis and prognosis
Claudio Sbarufatti (MSc in Mechanical Engineering, Politecnico di Milano, 2009; PhD in Aerospace Engineering - Rotary-Wing Aircrafts, Politecnico di Milano, 2013) is associate professor (ING-IND/14) at the Department of Mechanical Engineering at the Politecnico di Milano since November, 2019 and is Professor of the courses Machine Design 1 and Digital-Twin for industrial systems management.
His main expertise is focused on the development of methods for health and usage monitoring of structural components and systems. The monitoring methodology leverages on the combination of sensor data and models (Digital-Twin) for the identification of characteristic patterns in the signals, enabling the diagnosis and prognosis of system degradation.
Within this context, his research interests include the development and application of statistical and numerical methods for the solution of inverse problems in the context of structural anomaly and structural load identification, inverse FEM, machine learning, Monte-Carlo methods for Bayesian model updating and system prognosis, damage prognosis of composite structures, sensor network optimization, active and passive impact monitoring and, more recently, the identification of multifunctional composite materials with self-sensing and self-heating functions.
He is head of the research line Structural Health Monitoring and Prognosis in the SIGMALab and is currently involved in the management of national and international projects, the latter in the framework of the European Defence Agency, mainly focused on digital-twin modelling, load and structural health monitoring and residual life prognosis.
He is co-author of more than 100 papers on international journals and proceedings of international conferences.
Contents: (i) the concept of Digital-Twin (DT) (ii) the usage of machine learning for surrogate modelling in view of a fast DT simulation (I might focus on Gaussian Processes) (iii) embedding the surrogate model in a model-based prognostic framework for DT updating (I might focus on MCMC &/or Particle Filter).
Duration: 4x 45mins (with some hands-on practices included)
Physics-informed machine learning for structural dynamic
Lizzy Cross is a Professor in the Dynamics Research Group at the University of Sheffield with a focus on advanced data analysis and machine learning for structural health monitoring (SHM) and nonlinear system identification/modelling. She currently holds an EPSRC Innovation Fellowship pioneering physics-informed machine learning for structural dynamics. Lizzy is a co-director of the Laboratory for Verification and Validation, a state-of-the-art dynamic testing facility (lvv.ac.uk). She has published over 140 articles, including 35 journal papers, 5 book chapters. She was recently awarded the Achenbach medal which recognises an individual (within 10 years of PhD) who has made an outstanding contribution to the advancement of the field of SHM.
Structural Health Monitoring in Digital Era
Theodoros Loutas is a faculty member with the department of Mechanical Engineering & Aeronautics of the University of Patras. He holds a PhD in NDT/SHM of composite materials and over than 15 years of research experience in composite materials testing, structural health monitoring and probabilistic SHM data analysis. He has published 40 papers in high-impact international journals and 45 papers in international conferences. He teaches Physics, Dynamics, Non Destructive Testing and Sensors - Signal Processing in the undergraduate program of the Dpt of Mechanical Engineering & Aeronautics of the University of Patras. He has extensive experience in European collaborative projects the last 15 years having participated in more than 15 research European projects. Theodoros is currently the Principal Investigator in 4 European projects (1 Clean Sky-2 project, 2 H2020 in Transport and 1 FET-OPEN project among them).
How to use AI and digitalization for improving maintenance and reliability
Tiedo Tinga is full professor Life Cycle Management at the Netherlands Defence Academy (NLDA) in Den Helder, as well as full professor Dynamics based Maintenance at the Engineering Technology department of the University of Twente. He holds a MSc in Material Science from the University of Groningen, a PDEng in Materials Technology from TU Delft and a PhD in computational mechanics from Eindhoven University of Technology.
Tinga has been working with the National Aerospace Laboratory NLR for 10 years as a senior scientist. He was involved in research projects on structural integrity, fracture mechanics, computational mechanics and life prediction. In 2007 he joined the Netherlands Defence Academy as associate professor Maintenance Technology. There he is involved in educating the future officers of the Dutch Air Force, Army and Navy, and also leads a number of research projects on predictive maintenance. Since 2012, he combines this position with the part-time full professorship at the University of Twente. Since 2020 he also leads the Knowledge Center Smart Maintenance at the Royal Netherlands Navy, aiming to apply new maintenance technologies in practical cases.
His research focuses on improving the predictability of failures, aiming to improve preventive maintenance processes and to develop advanced predictive maintenance concepts. The research has a solid basis in understanding and modelling the physics of failure, which is combined with the development of advanced health and condition monitoring techniques and data analysis procedures. The research is applied to assets in various sectors of industry, including defence, transport, aerospace, maritime, process industry and infrastructure.
Tiedo now leads research programs on maintenance at both institutes and has been (co-) supervising 25 PhD and PDEng students in the past 8 years. Tiedo has published around 100 papers in international ISI journals and conferences. He has also been actively involved in the initiation (funding) and execution of many research projects, which are in close cooperation with industry and scientific partners.
Improving wind turbine blade reliability through application of Digital Twins
Kim Branner, Ph.D., is senior researcher at the Wind Energy Department at Technical University of Denmark (DTU). He is heading the Structural Design & Testing section, which primarily is doing research in the areas of experimental, numerical and analytical design in order to develop more reliable and exact methods and tools for structural design, manufacturing and testing of wind turbine blades and other large composite and metal structures.Kim is also overseeing DTU Large Scale Facility, which is a unique research and demonstration test facility for studying strength and fatigue of large structures exposed to complex loading as well as development and demonstration of sensor, IoT and other digitalization technologies. The facility is in operation since January 2019 and is able to test structures up to 45 m of length with advanced loading and measuring equipment. Kim is currently the project manager for the large project ReliaBlade running until 2023 and with the objective is to improve blade reliability through application of Digital Twins over the entire life cycle. The ReliaBlade project is also unique in the sense that it composes of two national funded projects (Danish and German), which are strongly aligned and coordinated.