TY - JOUR A1 - Schackmann, Oliver A1 - Márquez Reyes, Octavio A. A1 - Memmolo, Vittorio A1 - Lozano, Daniel A1 - Prager, Jens A1 - Moll, Jochen A1 - Kraemer, Peter T1 - Machine learning strategies with ensemble voting for ultrasonic damage detection in composite structures under varying temperature or load conditions N2 - In recent years, the development of machine learning (ML) techniques has led to significant progress in the field of structural health monitoring with ultrasonic-guided waves. However, a number of challenges still need to be resolved for reliable operation in realistic settings. In this work, we consider the complex problem of experimental damage detection under varying temperature or load conditions where damage locations are not included in the training set. The ML techniques proposed here include supervised and unsupervised methods originally developed for image and time series classification combined with ensemble voting. A performance demonstration of the ML techniques is presented using benchmark datasets from the open-guided waves platform. The unsupervised approach is then applied to a new dataset from an experimental campaign carried out on a composite over-wrapped pressure vessel used for hydrogen storage with real defects. Results show that ensemble voting enables the effective combination of the predictions of multiple transducer pairs, even with a limited number of strong individual classifiers. When applied to unsupervised learning, this returns high accuracy also when real damage over the structure is considered. KW - Structural-Health-Monitoring KW - Ultrasonic guided waves PY - 2025 UR - https://nbn-resolving.org/urn:nbn:de:kobv:b43-631571 DO - https://doi.org/10.1177/14759217251333066 SN - 1741-3168 SP - 1 EP - 19 PB - Sage Publications AN - OPUS4-63157 LA - eng AD - Bundesanstalt fuer Materialforschung und -pruefung (BAM), Berlin, Germany ER - TY - CONF A1 - Lozano Duarte, Daniel Hernando T1 - Modeling guided waves interacting with arbitrarily shaped inhomogeneities using SBFEM in combination with an extruded quadtree decomposition for 3D plates N2 - For structural health monitoring systems or non-destructive testing, it is crucial to study the interactions of ultrasonic-guided waves with inhomogeneities or damage in structural components. Simulation of these interactions poses a challenge. After the wave-damage interaction, the numerical model must reproduce features like mode conversion or wave scattering. Mathematically, damages are discontinuities in a computational domain, and each interacts diversely depending on its geometrical features. We propose a quadtree-based meshing in the current contribution to capture these geometrical characteristics. This kind of discretisation approach requires special techniques to couple cells of different sizes since, after a quadtree decomposition, the problem of so-called hanging nodes may arise. Transition elements based on high-order transfinite shape functions are used to solve this issue. The advantage is that these elements allow the correct coupling of the cells while retaining a high-order interpolation. Moreover, the reuse of the dynamic stiffness matrices can be exploited based on the similarity of the cells. This procedure renders the approach very efficient. Examples show the scattering characteristics of different guided wave modes after interacting with inhomogeneities and discontinuities in a plate. T2 - DAGA 2023 CY - Hamburg, Germany DA - 06.03.2023 KW - SBFEM KW - Quadtree KW - Image-based models KW - Wave-defect interaction PY - 2023 AN - OPUS4-59661 LA - eng AD - Bundesanstalt fuer Materialforschung und -pruefung (BAM), Berlin, Germany ER - TY - CONF A1 - Lozano Duarte, Daniel Hernando T1 - Octree Meshes for Ultrasonic Guided Waves: Mortar coupling combined with the Scaled Boundary Finite Element Method in Prismatic Structures N2 - Ultrasonic guided waves interacting with complex geometrical features are of high interest in the SHM field. For example, investigations into the extent to which a small change in defect geometry affects wave propagation have applications in Probability of Detection and defect reconstruction. These fields have to be supported by numerical tools aiming to retrieve accurate and reliable results and, thus, more realistic modeling. Due to the short wavelengths of ultrasonic waves and the relatively large domains in which the waves propagate, fine grid resolution for the entire domain is usually not possible due to limited computational resources. One solution is to partition the domain into different subdomains and use a specific, efficient approach for each subdomain. This contribution shows a combination of methods that aims to study these wave defect interactions in an elastic media. Firstly we introduce high-order 3D meshes that can resolve very fine geometrical details and are based on an Octree discretisation technique that employs special transition shape functions to resolve the hanging nodes that may appear due to the nature of the method. Afterward, the Octree mesh is embedded in a prismatic wave-guide based on the semi-analytical formulation of the SBFEM in 3D. To couple the different domains, we employ the mortar method, which enforces the continuity of the wave field across the internal boundaries of the domains. Examples show the power of the approach especially when it comes to parametric analysis. T2 - YIC23 CY - Porto, Portugal DA - 19.06.2023 KW - Octrees KW - Sbfem KW - Ultrasound KW - Mortar method PY - 2023 AN - OPUS4-59662 LA - eng AD - Bundesanstalt fuer Materialforschung und -pruefung (BAM), Berlin, Germany ER - TY - JOUR A1 - Lozano, Daniel A1 - Gravenkamp, Hauke A1 - Birk, Carolin T1 - Simulations of ultrasonic guided wave scattering using the scaled boundary finite element method N2 - This paper presents a numerical framework employing the Scaled Boundary Finite Element Method (SBFEM) for efficiently solving ultrasonic guided wave modal scattering problems and constructing scattering matrices (S-matrices). The framework integrates hierarchical octree meshing and an enhanced far field formulation. It incorporates boundary integrals, enabling an accurate representation of defect interactions with guided waves. Validation is performed using a benchmark test supported by an error estimation method based on energy balance. Two practical examples are used to showcase the simulation framework, where the scattering of guided waves with surface-breaking and subsurface cracks near rivet holes is studied. KW - Guided waves KW - Scattering matrix KW - Scaled boundary finite element method KW - Elastodynamics PY - 2025 UR - https://nbn-resolving.org/urn:nbn:de:kobv:b43-638295 DO - https://doi.org/10.1016/j.cma.2025.118204 SN - 0045-7825 VL - 445 SP - 1 EP - 23 PB - Elsevier B.V. AN - OPUS4-63829 LA - eng AD - Bundesanstalt fuer Materialforschung und -pruefung (BAM), Berlin, Germany ER - TY - JOUR A1 - Lozano-Martín, Daniel A1 - Kipphardt, Heinrich A1 - Khanipour, Peyman A1 - Tuma, Dirk A1 - Horrillo, Alfonso A1 - Chamorro, César R. T1 - Impact of hydrogen addition, up to 20 % (mol/mol), on the thermodynamic (p, ρ, T) properties of a reference high-calorific natural gas mixture with significant ethane and propane content N2 - Injecting hydrogen into the natural gas grid supports gradual decarbonization. To check the accuracy of equations of state for hydrogen-enriched natural gas mixtures, precise density data from well-characterized reference mixtures are essential. In a prior study, we provided experimental measurements for a natural gas constituted mainly of methane and for two derived hydrogen-enriched mixtures. In the present study, being the second and final part of our investigation, density measurements for a high-calorific natural gas with significant ethane and propane content, along with two hydrogen-enriched variants (10 and 20 mol-% hydrogen) are provided. The mixtures are gravimetrically prepared following ISO 6142-1. Density measurements, conducted with a single-sinker densimeter at temperatures from (260–350) K and pressures up to 20 MPa, are compared with three equations of state: AGA8-DC92, GERG-2008, and an improved GERG-2008. Results indicate that all models perform better for methane-dominant mixtures than for those containing heavier hydrocarbons. KW - Hydrogen-enriched natural gas KW - High-pressure density KW - Reference equation of state PY - 2025 UR - https://nbn-resolving.org/urn:nbn:de:kobv:b43-633371 DO - https://doi.org/10.1016/j.ijhydene.2025.05.173 SN - 0360-3199 VL - 140 SP - 256 EP - 271 PB - Elsevier BV CY - Amsterdam AN - OPUS4-63337 LA - eng AD - Bundesanstalt fuer Materialforschung und -pruefung (BAM), Berlin, Germany ER - TY - JOUR A1 - El Moutaouakil, H. A1 - Heimann, Jan A1 - Lozano, Daniel A1 - Memmolo, V. A1 - Schütze, A. T1 - Feature Extractor for Damage Localization on Composite-Overwrapped Pressure Vessel Based on Signal Similarity Using Ultrasonic Guided Waves N2 - Hydrogen is one of the future green energy sources that could resolve issues related to fossil fuels. The widespread use of hydrogen can be enabled by composite-overwrapped pressure vessels for storage. It offers advantages due to its low weight and improved mechanical performance. However, the safe storage of hydrogen requires continuous monitoring. Combining ultrasonic guided waves with interpretable machine learning provides a powerful tool for structural health monitoring. In this study, we developed a feature extraction approach based on a similarity method that enables interpretability in the proposed machine learning model for damage detection and localization in pressure vessels. Furthermore, a systematic optimization was performed to explore and tune the model’s parameters. This resulting model provides accurate damage localization and is capable of detecting and localizing damage on hydrogen pressure vessels with an average localization error of 2 cm and a classification accuracy of 96.5% when using quantized classification. In contrast, binarized classification yields a higher accuracy of 99.5%, but with a larger localization error of 6 cm. KW - Ultrasonic Guided Waves KW - Composite Overwrapped Pressure Vessel KW - Interpretable Machine Learning KW - Structural Health Monitoring KW - Damage Localization KW - Critical Infrastructure KW - Hydrogen KW - Non-destructive Testing PY - 2025 UR - https://nbn-resolving.org/urn:nbn:de:kobv:b43-640249 DO - https://doi.org/10.3390/app15179288 VL - 15 IS - 17 SP - 1 EP - 20 PB - MDPI AN - OPUS4-64024 LA - eng AD - Bundesanstalt fuer Materialforschung und -pruefung (BAM), Berlin, Germany ER - TY - JOUR A1 - Lozano-Martín, Daniel A1 - Tuma, Dirk A1 - Chamorro, César R. T1 - Evaluation of Reference Equations of State for Density Prediction in Regasified LNG Mixtures Using High-Precision Experimental Data N2 - This study evaluates the performance of three reference equations of state (EoS), AGA8-DC92, GERG-2008, and SGERG-88, in predicting the density of regasified liquefied natural gas (RLNG) mixtures. A synthetic nine-component RLNG mixture was gravimetrically prepared. High-precision density measurements were obtained using a single-sinker magnetic suspension densimeter over a temperature range of (250 to 350) K and pressures up to 20 MPa. The experimental data were compared with EoS predictions to evaluate their accuracy. AGA8-DC92 and GERG-2008 showed excellent agreement with the experimental data, with deviations within their stated uncertainty. In contrast, SGERG-88 exhibited significantly larger deviations for this RLNG mixture, particularly at low temperatures of (250 to 260) K, where discrepancies reached up to 3 %. Even at 300 K, deviations larger than 0.4 % were observed at high pressures, within the model’s uncertainty, but notably higher than those of the other two EoSs. The analysis was extended to three conventional 11-component natural gas mixtures (labeled G420 NG, G431 NG, and G432 NG), previously studied by our group using the same methodology. While SGERG-88 showed reduced accuracy for the RLNG mixture, it performed reasonably well for these three mixtures, despite two of them have a very similar composition to the RLNG. This discrepancy is attributed to the lower CO2 and N2 content typical in RLNG mixtures, demonstrating the sensitivity of EoS performance to minor differences in composition. These findings highlight the importance of selecting appropriate EoS models for accurate density prediction in RLNG applications. KW - Regasified LNG KW - Equation of state KW - Density KW - Gravimetric preparation PY - 2025 UR - https://nbn-resolving.org/urn:nbn:de:kobv:b43-646506 DO - https://doi.org/10.1007/s10765-025-03669-4 SN - 0195-928X VL - 46 SP - 1 EP - 25 PB - Springer Science and Business Media LLC CY - Heidelberg AN - OPUS4-64650 LA - eng AD - Bundesanstalt fuer Materialforschung und -pruefung (BAM), Berlin, Germany ER - TY - CONF A1 - Schackmann, Oliver A1 - Márquez Reyes, Octavio A1 - Memmolo, Vittorio A1 - Lozano, Daniel A1 - Prager, Jens A1 - Moll, Jochen A1 - Kraemer, Peter T1 - Intelligent damage detection in composite pressure vessels under varying environmental and operational conditions N2 - Despite proven approaches available in the literature, structural health monitoring by ultrasonic guided waves under varying environmental and operational conditions is still challenging. The use of machine learning approaches is discussed in this work, considering the complex problem of experimental damage detection under varying load conditions in a composite overwrapped pressure vessel for hydrogen storage. Specifically, unsupervised methods originally developed for image and time series classification are combined with ensemble voting to conceive reliable damage detection technique. This enables the effective combination of the predictions of multiple transducer pairs, even with a limited number of strong individual classifiers. A performance demonstration of the technique is presented using a real damage scenario dataset. T2 - IEEE 12th International Workshop on Metrology for AeroSpace (MetroAeroSpace) CY - Napoli, Italy DA - 18.06.2025 KW - Guided ultrasonic waves KW - Structural health monitoring KW - Artificial intelligence KW - Hydrogen storage PY - 2025 SN - 979-8-3315-0152-5 DO - https://doi.org/10.1109/MetroAeroSpace64938.2025.11114628 SN - 2575-7490 SP - 608 EP - 613 AN - OPUS4-64894 LA - eng AD - Bundesanstalt fuer Materialforschung und -pruefung (BAM), Berlin, Germany ER -