Periodic Reporting for period 1 - NOSTRADAMUS (NOn-contact STRucturAl DAMage for fUture Safety and lightweight)
Berichtszeitraum: 2021-10-01 bis 2023-09-30
The decarbonisation of transport requires the design of lighter components for electric vehicles. To ensure that new components meet safety requirements in terms of their susceptibility to vibration fatigue (VF), and to facilitate shorter development times, a new VF identification tool is required. Nostradamus aims to deliver tool that will enable fast and non-contact damage identification caused by VF using the High-speed (HS) camera. Optimisation of digital replicas of the actual products (digital twins) will become easier, to deliver of Safer, Lightweight, Green and Quiet (SLGQ) products.
The MSCA NOSTRADAMUS action has advanced the current SoA in the field of remote sensing by developing a new scientific method for modal identification from high-speed video recordings. This has made the use of optical methods in the field of vibration fatigue easier and more reliable. The training components enabled the Fellow to expand his expertise and the Fellow’s previous work led to ground-breaking findings in these areas, so there was a real two-way transfer of knowledge. The NOSTRADAMUS project led to three publications in high impact journals and the presentation of the work at eight scientific conferences worldwide (EU, USA, Brazil). The action had a positive impact on the Fellow and strengthened his professional maturity and independence.
The MSCA NOSTRADAMUS action has advanced the current SoA in the field of remote sensing by developing a new scientific method for modal identification from high-speed video recordings. This has made the use of optical methods in the field of vibration fatigue easier and more reliable. The training components enabled the Fellow to expand his expertise and the Fellow’s previous work led to ground-breaking findings in these areas, so there was a real two-way transfer of knowledge. The NOSTRADAMUS project led to three publications in high impact journals and the presentation of the work at eight scientific conferences worldwide (EU, USA, Brazil). The action had a positive impact on the Fellow and strengthened his professional maturity and independence.
The NOSTRADAMUS begun with training on the high-speed camera by repeating the work on frequency-based triangulation for measuring 3D deflection shapes. In the studied part, the deflection shapes were improved by using the dense spatial information of the high-speed camera to improve the identification of material damping. To obtain the strain mode shapes, further processing was then performed in the frequency domain using linear elasticity theory. This work was disseminated as a journal publication, and it was presented on three conferences and a poster presentation.
Research on Vibration Fatigue based on the training course consisted of replicating two papers from the host on VF using modal decomposition and short-time estimation of VF life non-stationary loading. During the training, experimental tests were conducted on the laboratory test cases. Then, the short-time life estimation procedure was extended to measuring the vibration response of the structure with a high-speed camera instead of using the limited number of classical sensors (accelerometers like in the repeated research). To enable cycle counting in the frequency domain, the strain-mode-shapes identified with the high-speed camera were converted into stress-mode-shapes. In this process, a new method for modal identification was developed, which is suitable for modal identification from high-speed camera measurements. The developed method was based on the Fellow’s previous research, and it is transferred to the host. Developed method was disseminated in the journal publication, and it is developed as the opensource python package (MorletWaveModal) to allow easy exploitation by others. Also, the presentation of the method and the opensource effort was communicated via scientific conferences and YouTube video presentations. To convert the measured strain to stress a linear elasticity theory was used as implemented in the FEM model (developed from the CAD model of the structure). The FEM model was mixed with the experimentally obtained strain-mode shapes using a sub-structuring method. Cycle counting using the multiaxial VF criteria was performed only for the dominant modes leading to damage assessment.
Research on Vibration Fatigue based on the training course consisted of replicating two papers from the host on VF using modal decomposition and short-time estimation of VF life non-stationary loading. During the training, experimental tests were conducted on the laboratory test cases. Then, the short-time life estimation procedure was extended to measuring the vibration response of the structure with a high-speed camera instead of using the limited number of classical sensors (accelerometers like in the repeated research). To enable cycle counting in the frequency domain, the strain-mode-shapes identified with the high-speed camera were converted into stress-mode-shapes. In this process, a new method for modal identification was developed, which is suitable for modal identification from high-speed camera measurements. The developed method was based on the Fellow’s previous research, and it is transferred to the host. Developed method was disseminated in the journal publication, and it is developed as the opensource python package (MorletWaveModal) to allow easy exploitation by others. Also, the presentation of the method and the opensource effort was communicated via scientific conferences and YouTube video presentations. To convert the measured strain to stress a linear elasticity theory was used as implemented in the FEM model (developed from the CAD model of the structure). The FEM model was mixed with the experimentally obtained strain-mode shapes using a sub-structuring method. Cycle counting using the multiaxial VF criteria was performed only for the dominant modes leading to damage assessment.
The most important scientific achievements of NOSTRADAMUS are the field of remote sensing in the transportation sector. A new method is introduced for identification of structural damping by high-speed camera measurements alone. Structural damping was identified to within 1/1000 of the pixel displacement. Up to this point, this was not possible with classical methods and represents a groundbreaking achievement. The most important step in estimating the lifetime of vibrations is the identification of the modal parameters: Natural frequencies, damping ratios and mode shapes. The development of the advanced method of modal identification from noisy, low dynamic and short signals is the second breakthrough of NOSTRADAMUS. With this method, the identification of structural damping became more reliable and accurate, and it was extended to the identification of natural frequencies and mode shapes. Such a method was not available, and NOSTRADAMUS made it possible. This method was developed as a ready-to-use software Python package, the source code is made available via an online repository. The quality of the developed method can be seen in the analysis of the Python package repository (https://pypistats.org/) which recorded an average of 4 downloads per day until its release in March 2023. The third achievement is the increased efficiency of full-field modal identification from high-speed video recordings of the vibrating structure. This is an important achievement in the field of remote sensing, where NOSTRADAMUS performs a direct conversion of pixel-based signals measured in the time domain into the modal domain. Full-field modal identification increases the accuracy of modal parameter identification. However, the current SoA required an inefficient video processing method to identify the displacement. NOSTRADAMUS enabled efficient identification of modal parameters directly from pixel intensities down to 1/1000 of the pixel, while the current SoA requires a hybrid approach with an additional high dynamic range sensor to achieve this accuracy.
The work carried out extends the limits of identifying structural changes using high-speed cameras. The methods and software developed/upgraded (MorletWaveModal, PyIDI, FLife) are based on the bottom-up approach, i.e. they can be used independently of each other, which allows a broader application. The main contribution is made in the transport sector for the remote sensing of vehicle components from a vibration fatigue perspective. The work of remote sensing can be extended for the fault detection of bridges and in the energy sector for the blades of wind turbines.
The work carried out extends the limits of identifying structural changes using high-speed cameras. The methods and software developed/upgraded (MorletWaveModal, PyIDI, FLife) are based on the bottom-up approach, i.e. they can be used independently of each other, which allows a broader application. The main contribution is made in the transport sector for the remote sensing of vehicle components from a vibration fatigue perspective. The work of remote sensing can be extended for the fault detection of bridges and in the energy sector for the blades of wind turbines.