Italian Group Fracture (IGF): E-Journals / Gruppo Italiano Frattura
Not a member yet
2800 research outputs found
Sort by
Machine learning-assisted fracture prediction: Integrating synthetic and experimental data for quasi-static notch failure analysis
Full text linkMachine learning has emerged as a powerful tool in various scientific fields for developing data-driven models, reducing the need for extensive physical testing. In this study, the fracture load was first predicted using Theory of Critical Distances (TCD) for experimental data. Further, this study presents a machine learning-based framework for the prediction of quasi-static fracture loads of U-notched polycarbonate specimens using a combination of experimental and synthetic data. Experimental data was obtained for eleven notch configurations, while synthetic data was generated through fracture mechanics-based simulations using PYMAPDL. XGBoost was trained on experimental and synthetic training datasets, evaluated against a fixed randomly selected experimental test set. Model performance was evaluated using MAPE, MAE, RMSE values as R2 varied significantly with test set sampling and random state selection. This study systematically evaluates how varying the proportion of synthetic data influences model performance, offering a scalable strategy to minimize experimental dependence without compromising accuracy. Experiment-only dataset provided the highest accuracy, while hybrid models performed reasonably well. The full dataset model combining all experimental and synthetic data, achieved the most robust and accurate predictions, with errors ranging within ±5% and yielding the lowest MAPE of 1.18% and MAE of 78.73 N. It is suggested that synthetic data can significantly enhance the training of machine learning models, but cannot completely replace the experimental data, especially in critical applications
Utilizing cylindrical and cubical specimens with edge notch to determine size-independent fracture quantities of rock materials
Full text linkThe compliance method was first applied to short rod specimens to determine the nonlinear fracture toughness of rock materials by ISRM (International Society for Rock Mechanics) in the 1980s. In this study, utilizing the techniques of the J-integral and the crack closure integral (CCI), crucial linear elastic fracture mechanics expressions for straight-notched disk bending (SNDB) specimens, whose tests are simpler than those for short bar specimens, and single-notch cube bending (SNCB) specimens are initially derived to estimate crack propagation states in rock samples. Andesite-based SNDB specimens from the literature are examined using the compliance approach, and a strong correlation is observed between the compliance approach and the nonlinear approach reported in the literature. Subsequently, limestone-based SNCB specimens and beams containing cracks are produced and tested under bending. The fracture test data are estimated using the peak load approach, and the results of the comparative analysis are found to be satisfactorily consistent for both beams and SNCB specimens. The findings of this study reveal that the non-Hookean fracture quantities of rocks can be adequately determined using SNDB and SNCB specimens of a single size
Application of the thermography method for determining the fatigue limit of a nickel alloy produced by wire‑arc additive manufacturing
Full text linkThe development of additive technologies for manufacturing safety‑critical components operating under vibration must be accompanied by a careful analysis of the material’s resistance to high‑cycle fatigue (HCF). This paper presents a technique for the rapid assessment of the fatigue limit of nickel alloys fabricated by wire‑and‑arc additive manufacturing (WAAM). The technique grounded in infrared thermography (IRT), utilizes the self‑heating effect during cyclic loading. The technique realisation involves choosing specimen design, test equipment, the number and parameters of loading blocks, self‑heating indicators, and result‑processing procedures that account for material specifics. It is shown that the rate of temperature rise at the specimen surface at the start of each loading block can serve as an indicator of self‑heating. Experimental data on the fatigue limit of specimens made from the heat‑resistant alloy Inconel 625 produced by WAAM are obtained. Validation of the developed method is performed by comparing the fatigue limit derived from IRT with the results of conventional fatigue testing and the corresponding S–N curve. For additive nickel alloys, the proposed accelerated fatigue‑limit assessment allows a substantial reduction in the number of specimens and the time required to select technological parameters and refine additive manufacturing processes compared with traditional fatigue testing
Comparative assessment of the acoustic activity and the Pressure Stimulated Voltage in marble specimens under compression
Full text linkThe temporal evolution of the electric activity generated in marble specimens under uniaxial compression is analyzed and quantified in terms of the Pressure Stimulated Voltage (Electric Potential) developed. The evolution of the electric activity is considered in juxtaposition to that of the respective acoustic one, quantified either in terms of the average frequency of generation of acoustic signals or of their Cumulative Energy content. Two classes of specimens were tested, differing with respect to the loading rate imposed. It is concluded that the electric activity is very weak, or even negligible, until the critical instant designated by the entrance into the stage of thermodynamically irreversible response of the material. Beyond this instant the electric activity starts increasing very rapidly almost until the instant at which the load attains its peak value. A few seconds before fracture, the electric signal exhibits an abrupt drop. The temporal evolution of the electric activity and that of the acoustic one are in excellent agreement, independently of the parameter used for their quantification. The study revealed that both activities provide clear pre-failure indices, early warning about upcoming disastrous fracture. Moreover, it was highlighted that the loading rate diversifies the results only from a quantitative point of view, “translating” the stress interval within which the pre-failure indices are located: The higher the loading rate the lower the stress level at which the pre-failure indices are detected
Studying the strength and damageability of composite element in looped metal-composite joint under tensile loading
Full text linkThe article presents the results of computational and experimental studies of the strength and damageability of a connecting composite part in a metal-composite joint of isogrid and anisogrid structures made of polymer composite material under tension. The assessment of the minimal cross-sectional area of the connecting part was performed based on design calculations, taking into account strengthening during extrusion of the excess binder. The onset of failure in the contact zone with the steel element of the metal-composite joint was predicted based on experimental studies using model samples. A comparison was made between the calculation results for the tensile loading diagram, considering the physical nonlinear behavior of composite material in the joint zone, and the readings of strain gauges after testing the metal-composite joint. Damages and deformation of the connecting composite part under the tensile load was imaged using acoustic microscopy
ENLO-SED: an innovative method for large-scale Strain Energy Density (SED) estimation in welded joints using structural stresses derived from Element Nodal LOads (ENLO)
Full text linkWelded joints have always been critical elements of industrial mechanical structures, often being the source of failures related to the presence of fatigue loads. Although the academic world has presented advanced methodologies for the assessment of local fatigue, such as the Strain Energy Density (SED) approach, which offers high accuracy, their high computational requirements hinder their adoption by the industrial world. This paper introduces a new hybrid methodology, called ENLO-SED, which integrates the SED approach by calculating the Strain Energy Density using the element Nodal load approach (ENLO), with the aim of maintaining high accuracy while significantly reducing the computational effort. The proposed method is validated on a complex case study, representative of a real industrial case, demonstrating a prediction error within 8% compared to the application of the classic SED method. Furthermore, the innovative ENLO-SED approach reduces the meshing and solution times by 15 and 5 times, respectively. These results confirm the robustness, efficiency, and scalability of the method, making it suitable for large-scale industrial applications
Implementation of interface damage model with friction to concrete-FRP shear connector
Full text linkThe study investigates the application of fibre-reinforced polymer (FRP) composites in constructions of bridges. It highlights the main advantages of using FRP as a building material and points out its suitability for various structural applications. A numerical analysis was performed on different shape modifications of a jigsaw-puzzle type continuous shear connector. For the interface between concrete and FRP, a bilinear cohesive zone model with friction within a variationally based formulation of interface damage has been chosen and tested. This model captures the load displacement relation accounting for a softening region and offering a continuous response of key variables of stress and damage. Findings illustrate reliability of the cohesive bilinear model as a tool for predicting failure and show a promise for applying it in material design, or in design of FRP composite structures, their members and specifications of their construction details
A digital twin framework with MobileNetV2 for damage detection in slab structures
Full text linkIn this study, a digital twin framework is proposed for damage detection in a civil structure, which consists of a finite element model, neural networks, model updating methods, and signal processing. To verify the proposed framework, we present a case study of slab structure using deflection measurement as input data. The dynamic characteristics of the physical model are used to calibrate the digital twin model. Damage scenarios are created on the digital twin model. The defection of the damaged slab under static loads is analyzed with two-dimensional discrete wavelet theory (DWT), whereas the diagonal wavelets are used to extract images data set used to train the convolutional neural network (CNN). MobileNetV2 uses transfer learning can reduce the number of trained parameters and hence perform fast convergence. The proposed method gives high accuracy about detection of low-severity damage having the severity less than 10%. There is more than 80% accuracy for predicting the damaged location and its severity. The success of using MobileNetV2 and transfer learning helps to improve the methods further on mobile devices and the potential for more applications. Moreover, the proposed framework does not require the data of the intact structures, leading to much wider applications
Structural behavior of GFRP-concrete composite beams
Full text linkGlass Fiber Reinforced Polymer (GFRP) I-sections offer a promising alternative to traditional steel reinforcement due to their reduced weight and maintenance requirements. This study aims to optimize the design of GFRP-reinforced composite concrete beams for cost-effective solutions. Twelve tested beams were tested under four-point bending loading, divided into four groups with varying depths: three conventional reinforced concrete (RC) beams as control specimens; nine beams with GFRP I-sections positioned internally; externally; and internally with exposure to 500°C for 90 minutes. The test results indicate that GFRP-reinforced beams exhibit superior strength and bending resistance compared to conventional RC beams where an increase in maximum load ranging from 62% to 113% and reduced deflection at the same load level. Optimal performance was observed when GFRP I-sections were placed near the tensioned fiber. Exposure to elevated temperatures resulted in minimal performance reductions, not exceeding 5% at yield load and 16% at maximum load comparing with composite tested specimens without exposure to elevated temperature. Theoretical analyses closely aligned with experimental results, providing a foundation for practical guidelines on the economical design of GFRP-reinforced composite
Predicting the strength of 3D-printed conductive composite under tensile load: A probabilistic modeling and experimental study
Full text linkConductive PLA is an innovative composite material that combines the ecological benefits of polylactic acid, a biodegradable thermoplastic, with electrical conductivity properties. Usually used in additive manufacturing for its ease of printing and low environmental impact, PLA remains an insulator, which limits its applications in the electrical field. To overcome this limitation, conductive fillers such as carbon nanotubes or carbon black are being added, opening the way to new functional uses. This study focuses on a specific composite: carbon black-filled PLA (PLA-CB). This material combines the qualities of traditional PLA with enhanced conductivity thanks to the carbon black particles. To assess its performance, a number of mechanical tests were carried out, including tensile tests on samples manufactured by 3D printing using the FFF process. The study focused in particular on the influence of crosshead speed and the impact of different notch shapes on the material's properties. To analyze the durability of PLA-CB, a probabilistic model based on the two-parameter Weibull distribution was used to assess the risk of failure under different conditions. Reliability curves were also established to better understand the tensile stress and strain at break of the material. This approach could also be applied to other 3D-printed polymers to refine their analytical and numerical modeling