System failure prediction through rare-events elastic-net logistic regression

Predicting failures in a distributed system based on previous events through logistic regression is a standard approach in literature. This technique is not reliable, though, in two situations: in the prediction of rare events, which do not appear in enough proportion for the algorithm to capture, and in environments where there are too many variables, as logistic regression tends to overfit on this situations; while manually selecting a subset of variables to create the model is error- prone. On this paper, we solve an industrial research case that presented this situation with a combination of elastic net logistic regression, a method that allows us to automatically select useful variables, a process of cross-validation on top of it and the application of a rare events prediction technique to reduce computation time. This process provides two layers of cross- validation that automatically obtain the optimal model complexity and the optimal mode l parameters values, while ensuring even rare events will be correctly predicted with a low amount of training instances. We tested this method against real industrial data, obtaining a total of 60 out of 80 possible models with a 90% average model accuracy.

Effect of ion flux on helium retention in helium-irradiated tungsten

Helium retention in irradiated tungsten leads to swelling, pore formation, sample exfoliation and embrittlement with deleterious consequences in many applications. In particular, the use of tungsten in future nuclear fusion plants is proposed due to its good refractory properties. However, serious concerns about tungsten survivability stems from the fact that it must withstand severe irradiation conditions. In magnetic fusion as well as in inertial fusion (particularly with direct drive targets), tungsten components will be exposed to low and high energy ion (helium) irradiation, respectively. A common feature is that the most detrimental situations will take place in pulsed mode, i.e., high flux irradiation. There is increasing evidence on a correlation between a high helium flux and an enhancement of detrimental effects on tungsten. Nevertheless, the nature of these effects is not well understood due to the subtleties imposed by the exact temperature profile evolution, ion energy, pulse duration, existence of impurities and simultaneous irradiation with other species. Physically based Kinetic Monte Carlo is the technique of choice to simulate the evolution of radiation-induced damage inside solids in large temporal and space scales. We have used the recently developed code MMonCa (Modular Monte Carlo simulator), presented in this conference for the first time, to study He retention (and in general defect evolution) in tungsten samples irradiated with high intensity helium pulses. The code simulates the interactions among a large variety of defects and impurities (He and C) during the irradiation stage and the subsequent annealing steps. In addition, it allows us to vary the sample temperature to follow the severe thermo-mechanical effects of the pulses. In this work we will describe the helium kinetics for different irradiation conditions. A competition is established between fast helium cluster migration and trapping at large defects, being the temperature a determinant factor. In fact, high temperatures (induced by the pulses) are responsible for large vacancy cluster formation and subsequent additional trapping with respect to low flux irradiation.

Effect of ion flux on helium retention in helium-irradiated tungsten

Helium retention in irradiated tungsten leads to swelling, pore formation, sample exfoliation and embrittlement with deleterious consequences in many applications. In particular, the use of tungsten in future nuclear fusion plants is proposed due to its good refractory properties. However, serious concerns about tungsten survivability stems from the fact that it must withstand severe irradiation conditions. In magnetic fusion as well as in inertial fusion (particularly with direct drive targets), tungsten components will be exposed to low and high energy ion irradiation (helium), respectively. A common feature is that the most detrimental situations will take place in pulsed mode, i.e., high flux irradiation. There is increasing evidence of a correlation between a high helium flux and an enhancement of detrimental effects on tungsten. Nevertheless, the nature of these effects is not well understood due to the subtleties imposed by the exact temperature profile evolution, ion energy, pulse duration, existence of impurities and simultaneous irradiation with other species. Object Kinetic Monte Carlo is the technique of choice to simulate the evolution of radiation-induced damage inside solids in large temporal and space scales. We have used the recently developed code MMonCa (Modular Monte Carlo simulator), presented at COSIRES 2012 for the first time, to study He retention (and in general defect evolution) in tungsten samples irradiated with high intensity helium pulses. The code simulates the interactions among a large variety of defects and during the irradiation stage and the subsequent annealing steps. The results show that the pulsed mode leads to significantly higher He retention at temperatures higher than 700 K. In this paper we discuss the process of He retention in terms of trap evolution. In addition, we discuss the implications of these findings for inertial fusion.

Global instability analysis and control of three-dimensional long open cavity flow

The three-dimensional wall-bounded open cavity may be considered as a simplified geometry found in industrial applications such as leading gear or slotted flats on the airplane. Understanding the three-dimensional complex flow structure that surrounds this particular geometry is therefore of major industrial interest. At the light of the remarkable former investigations in this kind of flows, enough evidences suggest that the lateral walls have a great influence on the flow features and hence on their instability modes. Nevertheless, even though there is a large body of literature on cavity flows, most of them are based on the assumption that the flow is two-dimensional and spanwise-periodic. The flow over realistic open cavity should be considered. This thesis presents an investigation of three-dimensional wall-bounded open cavity with geometric ratio 6:2:1. To this aim, three-dimensional Direct Numerical Simulation (DNS) and global linear instability have been performed. Linear instability analysis reveals that the onset of the first instability in this open cavity is around Recr 1080. The three-dimensional shear layer mode with a complex structure is shown to be the most unstable mode. I t is noteworthy that the flow pattern of this high-frequency shear layer mode is similar to the observed unstable oscillations in supercritical unstable case. DNS of the cavity flow carried out at different Reynolds number from steady state until a nonlinear saturated state is obtained. The comparison of time histories of kinetic energy presents a clearly dominant energetic mode which shifts between low-frequency and highfrequency oscillation. A complete flow patterns from subcritical cases to supercritical case has been put in evidence. The flow structure at the supercritical case Re=1100 resembles typical wake-shedding instability oscillations with a lateral motion existed in the subcritical cases. Also, This flow pattern is similar to the observations in experiments. In order to validate the linear instability analysis results, the topology of the composite flow fields reconstructed by linear superposition of a three-dimensional base flow and its leading three-dimensional global eigenmodes has been studied. The instantaneous wall streamlines of those composited flows display distinguish influence region of each eigenmode. Attention has been focused on the leading high-frequency shear layer mode; the composite flow fields have been fully recognized with respect to the downstream wave shedding. The three-dimensional shear layer mode is shown to give rise to a typical wake-shedding instability with a lateral motions occurring downstream which is in good agreement with the experiment results. Moreover, the spanwise-periodic, open cavity with the same length to depth ratio has been also studied. The most unstable linear mode is different from the real three-dimensional cavity flow, because of the existence of the side walls. Structure sensitivity of the unstable global mode is analyzed in the flow control context. The adjoint-based sensitivity analysis has been employed to localized the receptivity region, where the flow is more sensible to momentum forcing and mass injection. Because of the non-normality of the linearized Navier-Stokes equations, the direct and adjoint field has a large spatial separation. The strongest sensitivity region is locate in the upstream lip of the three-dimensional cavity. This numerical finding is in agreement with experimental observations. Finally, a prototype of passive flow control strategy is applied.

Numerical analysis of mixe-mode fracture of brickwork masonry with embedded discontinuity elements

One of the common pathologies of brickwork masonry structural elements and walls is the cracking associated with the differential settlements and/or excessive deflections of the slabs along the life of the structure. The scarce capacity of the masonry in order to accompany the structural elements that surround it, such as floors, beams or foundations, in their movements makes the brickwork masonry to be an element that frequently presents this kind of problem. This problem is a fracture problem, where the wall is cracked under mixed mode fracture: tensile and shear stresses combination, under static loading. Consequently, it is necessary to advance in the simulation and prediction of brickwork masonry mechanical behaviour under tensile and shear loading. The quasi-brittle behaviour of the brickwork masonry can be studied using the cohesive crack model whose application to other quasibrittle materials like concrete has traditionally provided very satisfactory results.

Mode II fracture energy in the adhesive bonding of dissimilar substrates: carbon fibre composite to aluminium joints

The end-notched flexure (ENF) test calculates the value of mode II fracture energy in adhesive bonding between the substrates of same nature. Traditional methods of calculating fracture energy in the ENF test are not suitable in cases where the thickness of the adhesive is non-negligible compared with adherent thicknesses. To address this issue, a specific methodology for calculating mode II fracture energy has been proposed in this paper. To illustrate the applicability of the proposed method, the fracture energy was calculated by the ENF test for adhesive bonds between aluminium and a composite material, which considered two different types of adhesive (epoxy and polyurethane) and various surface treatments. The proposed calculation model provides higher values of fracture energy than those obtained from the simplified models that consider the adhesive thickness to be zero, supporting the conclusion that the calculation of mode II fracture energy for adhesives with non-negligible thickness relative to their adherents should be based on mathematical models, such as the method proposed in this paper, that incorporate the influence of this thickness.

Finite element simulation of sandwich panels of laminated plaster and rockwool under mixed mode fracture

Sandwich panels of laminated gypsum and rock wool have shown large pathology of cracking due to excessive slabs deflection. Currently the most widespread use of this material is as vertical elements of division or partition, with no structural function, what justifies that there are no studies on the mechanism of fracture and mechanical properties related to it. Therefore, and in order to reduce the cracking problem, it is necessary to progress in the simulation and prediction of the behaviour under tensile and shear load of such panels, although in typical applications have no structural responsability.

Finite element simulation of sandwich panels of plasterboard and rock wool under mixed mode fracture

This paper presents the results of research on mixed mode fracture of sandwich panels of plasterboard and rock wool. The experimental data of the performed tests are supplied. The specimens were made from commercial panels. Asymmetrical three-point bending tests were performed on notched specimens. Three sizes of geometrically similar specimens were tested for studying the size effect. The paper also includes the numerical simulation of the experimental results by using an embedded cohesive crack model.The involved parameters for modelling are previously measured by standardised tests.

Projector Model for Efficient List-Mode Reconstruction in PET Scanners with Parallel Planar Detectors

We have developed a new projector model specifically tailored for fast list-mode tomographic reconstructions in Positron emission tomography (PET) scanners with parallel planar detectors. The model provides an accurate estimation of the probability distribution of coincidence events defined by pairs of scintillating crystals. This distribution is parameterized with 2D elliptical Gaussian functions defined in planes perpendicular to the main axis of the tube of response (TOR). The parameters of these Gaussian functions have been obtained by fitting Monte Carlo simulations that include positron range, acolinearity of gamma rays, as well as detector attenuation and scatter effects. The proposed model has been applied efficiently to list-mode reconstruction algorithms. Evaluation with Monte Carlo simulations over a rotating high resolution PET scanner indicates that this model allows to obtain better recovery to noise ratio in OSEM (ordered-subsets, expectation-maximization) reconstruction, if compared to list-mode reconstruction with symmetric circular Gaussian TOR model, and histogram-based OSEM with precalculated system matrix using Monte Carlo simulated models and symmetries.

Estudio numérico y experimental del efecto de tamaño en modo mixto de los paneles de yeso laminado y lana de roca

Los paneles sándwich de yeso laminado y lana de roca presentan una abundante patología de fisuración debida a flechas excesivas de forjados. Existe, por tanto, la necesidad de avanzar en la simulación y predicción de comportamiento bajo solicitaciones de tracción y cortante de ese tipo de paneles, a pesar de que en las aplicaciones habituales no tienen responsabilidad estructural. El comportamiento de este material puede ser considerado cuasi-frágil, y en base a ello en este trabajo ha sido estudiado haciendo uso de modelos de fisura cohesiva, cuya aplicación a otros materiales cuasifrágiles, como el hormigón, ha aportado resultados muy satisfactorios. En esta comunicación se presenta el trabajo realizado para estudiar el efecto del tamaño del elemento de yeso laminado y lana de roca en su comportamiento mecánico-resistente. Para ello se diseñó una campaña de ensayos en modo mixto sobre probetas de diferente tamaño. Se han realizado ensayos de flexión en tres puntos en modo mixto de unas probetas entalladas, geométricamente similares y de diferente tamaño, obteniéndose las curvas carga-desplazamiento y cargaabertura de la boca de la entalla. Para simular numéricamente el comportamiento en fractura del panel en modo mixto se ha utilizado un modelo de elementos finitos con fisura embebida basado en la fisura cohesiva en el que se introducen como entrada los parámetros obtenidos a partir de la experimentación de trabajos anteriores, obteniéndose un buen ajuste. En función de estos resultados se analiza el efecto del tamaño en los paneles. Sandwich panels of laminated gypsum and rockwool have an abundant pathology of cracking due to excessive slabs deflection. Therefore, it is necessary to progress in the simulation and prediction of behaviour under tensile and shear load of such panels, although in typical applications have no structural responsability. The behaviour of this material may be considered quasi-brittle and, based on this idea, in this work has been studied using a cohesive crack model that has been applied to other quasi-brittle materials, such as concrete, and has provided very satisfactory results. This communication presents the work carried out to study the size effect of the specimen of plasterboard and rockwool in its mechanical and resistant behaviour. The authors designed an experimental campaign under mixed mode composed by testing specimens of different sizes. Assymetrical three-point bending tests have been performed on notched specimens, geometrically similar and of different size, to obtain load-displacement and load-crack moutn opening displacement curves. To numerically simulate the mixed-mode fracture behaviour of the panels we have used a finite element model with embedded crack, based on the cohesive crack model, using as input the experimental parameters obtained in previous work, obtaining a good adjustment. Based on these results we analyze the size effect of the panels

Maximum Likelihood Estimation of Modal Parameters in Structures Using the Expectation Maximization Algorithm

This paper presents a time-domain stochastic system identification method based on maximum likelihood estimation (MLE) with the expectation maximization (EM) algorithm. The effectiveness of this structural identification method is evaluated through numerical simulation in the context of the ASCE benchmark problem on structural health monitoring. The benchmark structure is a four-story, two-bay by two-bay steel-frame scale model structure built in the Earthquake Engineering Research Laboratory at the University of British Columbia, Canada. This paper focuses on Phase I of the analytical benchmark studies. A MATLAB-based finite element analysis code obtained from the IASC-ASCE SHM Task Group web site is used to calculate the dynamic response of the prototype structure. A number of 100 simulations have been made using this MATLAB-based finite element analysis code in order to evaluate the proposed identification method. There are several techniques to realize system identification. In this work, stochastic subspace identification (SSI)method has been used for comparison. SSI identification method is a well known method and computes accurate estimates of the modal parameters. The principles of the SSI identification method has been introduced in the paper and next the proposed MLE with EM algorithm has been explained in detail. The advantages of the proposed structural identification method can be summarized as follows: (i) the method is based on maximum likelihood, that implies minimum variance estimates; (ii) EM is a computational simpler estimation procedure than other optimization algorithms; (iii) estimate more parameters than SSI, and these estimates are accurate. On the contrary, the main disadvantages of the method are: (i) EM algorithm is an iterative procedure and it consumes time until convergence is reached; and (ii) this method needs starting values for the parameters. Modal parameters (eigenfrequencies, damping ratios and mode shapes) of the benchmark structure have been estimated using both the SSI method and the proposed MLE + EM method. The numerical results show that the proposed method identifies eigenfrequencies, damping ratios and mode shapes reasonably well even in the presence of 10% measurement noises. These modal parameters are more accurate than the SSI estimated modal parameters.

Fracture strength of drilled wafers: study of the stress concentration factor and determination of the weibull distribution of the fracture stress

In the photovoltaic field, the back contact solar cells technology has appeared as an alternative to the traditional silicon modules. This new type of cells places both positive and negative contacts on the back side of the cells maximizing the exposed surface to the light and making easier the interconnection of the cells in the module. The Emitter Wrap-Through solar cell structure presents thousands of tiny holes to wrap the emitter from the front surface to the rear surface. These holes are made in a first step over the silicon wafers by means of a laser drilling process. This step is quite harmful from a mechanical point of view since holes act as stress concentrators leading to a reduction in the strength of these wafers. This paper presents the results of the strength characterization of drilled wafers. The study is carried out testing the samples with the ring on ring device. Finite Element models are developed to simulate the tests. The stress concentration factor of the drilled wafers under this load conditions is determined from the FE analysis. Moreover, the material strength is characterized fitting the fracture stress of the samples to a three-parameter Weibull cumulative distribution function. The parameters obtained are compared with the ones obtained in the analysis of a set of samples without holes to validate the method employed for the study of the strength of silicon drilled wafers.

Influence of the Second Flexural Mode on the Response of High-Speed Bridges

This paper deals with the assessment of the contribution of the second flexural mode to the dynamic behaviour of simply supported railway bridges. Alluding to the works of other authors, it is suggested in some references that the dynamic behaviour of simply supported bridges could be adequately represented taking into account only the contribution of the fundamental flexural mode. On the other hand, the European Rail Research Institute (ERRI) proposes that the second mode should also be included whenever the associated natural frequency is lower than 30 Hz]. This investigation endeavours to clarify the question as much as possible by establishing whether the maximum response of the bridge, in terms of displacements, accelerations and bending moments, can be computed accurately not taking account of the contribution of the second mode. To this end, a dimensionless formulation of the equations of motion of a simply supported beam traversed by a series of equally spaced moving loads is presented. This formulation brings to light the fundamental parameters governing the behaviour of the beam: damping ratio, dimensionless speed $ \alpha$=VT/L, and L/d ratio (L stands for the span of the beam, V for the speed of the train, T represents the fundamental period of the bridge and d symbolises the distance between consecutive loads). Assuming a damping ratio equal to 1%, which is a usual value for prestressed high-speed bridges, a parametric analysis is conducted over realistic ranges of values of $ \alpha$ and L/d. The results can be extended to any simply supported bridge subjected to a train of equally spaced loads in virtue of the so-called Similarity Formulae. The validity of these formulae can be derived from the dimensionless formulation mentioned above. In the parametric analysis the maximum response of the bridge is obtained for one thousand values of speed that cover the range from the fourth resonance of the first mode to the first resonance of the second mode. The response at twenty-one different locations along the span of the beam is compared in order to decide if the maximum can be accurately computed with the sole contribution of the fundamental mode.

Estudio numérico y experimental del efecto de tamaño en modo mixto de los paneles de yeso laminado y lana de roca

Los paneles sándwich de yeso laminado y lana de roca presentan una abundante patología de fisuración debida a flechas excesivas de forjados. Existe, por tanto, la necesidad de avanzar en la simulación y predicción de comportamiento bajo solicitaciones de tracción y cortante de ese tipo de paneles, a pesar de que en las aplicaciones habituales no tienen responsabilidad estructural. El comportamiento de este material puede ser considerado cuasi-frágil, y en base a ello en este trabajo ha sido estudiado haciendo uso de modelos de fisura cohesiva, cuya aplicación a otros materiales cuasifrágiles, como el hormigón, ha aportado resultados muy satisfactorios. En esta comunicación se presenta el trabajo realizado para estudiar el efecto del tamaño del elemento de yeso laminado y lana de roca en su comportamiento mecánico-resistente. Para ello se diseñó una campaña de ensayos en modo mixto sobre probetas de diferente tamaño. Se han realizado ensayos de flexión en tres puntos en modo mixto de unas probetas entalladas, geométricamente similares y de diferente tamaño, obteniéndose las curvas carga-desplazamiento y cargaabertura de la boca de la entalla. Para simular numéricamente el comportamiento en fractura del panel en modo mixto se ha utilizado un modelo de elementos finitos con fisura embebida basado en la fisura cohesiva en el que se introducen como entrada los parámetros obtenidos a partir de la experimentación de trabajos anteriores, obteniéndose un buen ajuste. En función de estos resultados se analiza el efecto del tamaño en los paneles. Sandwich panels of laminated gypsum and rockwool have an abundant pathology of cracking due to excessive slabs deflection. Therefore, it is necessary to progress in the simulation and prediction of behaviour under tensile and shear load of such panels, although in typical applications have no structural responsability. The behaviour of this material may be considered quasi-brittle and, based on this idea, in this work has been studied using a cohesive crack model that has been applied to other quasi-brittle materials, such as concrete, and has provided very satisfactory results. This communication presents the work carried out to study the size effect of the specimen of plasterboard and rockwool in its mechanical and resistant behaviour. The authors designed an experimental campaign under mixed mode composed by testing specimens of different sizes. Assymetrical three-point bending tests have been performed on notched specimens, geometrically similar and of different size, to obtain load-displacement and load-crack moutn opening displacement curves. To numerically simulate the mixed-mode fracture behaviour of the panels we have used a finite element model with embedded crack, based on the cohesive crack model, using as input the experimental parameters obtained in previous work, obtaining a good adjustment. Based on these results we analyze the size effect of the panels.

A fracture model for pearlitic steel bars using a cohesive model

The fracture of ductile materials, such as metals, is usually explained with the theory of nucleation, growth and coalescence of microvoids. Based on this theory, many numerical models have been developed, with a special mention to Gurson-type models. These models simulate mathematically the physical growth of microvoids, leading to a progressive development of the internal damage that takes place during a tensile test. In these models, the damage starts to develop in very early stages of the test. Tests carried out by the authors suggest that, in the case of some eutectoid steels such as those used for manufacturing prestressing steel wires, the internal damage that takes place as a result of the growth of microvoids is only noticeable in very late stages of the tensile test. In the authors’ opinion, using a cohesive model as a failure criterion may be interesting in this case; a cohesive model only requires two parameters to be defined, with the fracture energy being one of them, which can be obtained experimentally. In addition to this, given that it is known that the stress triaxiality has a strong influence on the fracture of ductile materials, a cohesive model whose parameters are affected by the value of the stress triaxiality can be considered. This work presents a fracture model for steel specimens in a tensile test, based on a cohesive behaviour and taking into account the effect of stress triaxiality, which is different at each point of the fracture plane.