1,721,071 research outputs found
Internet Topologies and Their Influence on End-to-End Communications. Tesi di Dottorato in Ingegneria dell’Informazione, tutor: prof. M. Longo, D.I.I.I.E., Università degli Studi di Salerno
No full textSome advancements on degradation empirical models for reliability evaluation of solid oxide fuel cells
No full textSolid oxide fuel cells (SOFCs) are electrochemical devices working at high temperature and producing electricity with high efficiency. An importantopen issue of this promising technology is the characterization of the degradation process. Being a direct observation of degradation phenomena difficult to implement, indirect performance indicators based on voltage measurements are frequently adopted, that are influenced also by the temperature of the furnace containing the cells. The voltage measurements collected during degradation tests on SOFC stacks can be effectively modelled by empirical random-effects regression models. They allow us to describe the variability components present in the measurements, such as the slow decay of voltage over time of each single cell of the stack, the variability of voltage decay among cells, and the fluctuations of voltage due to experimental noise and lack of fit, depending also on the temperature fluctuations. Some advancements are introduced on the available degradation empirical models in order to cope with correlation structure induced by temperature. Point and interval estimates are also derived for some performance measures of interest, for instance the prediction of cell voltage and the reliability function. Finally, the proposed methodology is applied to a real degradation test of a SOFC prototype
End-to-End Loss Probabilities in Different Internet-like Networks with a Given Average Hop Count
No full textRealistic networks generators are necessary for simulation and performance evaluation of data communication systems. Such an aspect has driven the collection and the analysis of data on the Internet large-scale structure. The evidence of a power-law behavior of real networks has stimulated the introduction of new procedures to generate Internet-like topologies. Assuming a simple loss model for the links, this paper analyses how the prediction of the loss probabilities during communications obtained by simulation can be influenced by the adoption of a specific topological model for the Internet graph (here, the Waxman or the Barabasi-Albert model), given a average node distance in terms of hops
Connectivity of Ad Hoc Networks with Link Asymmetries Induced by Shadowing
No full textThe aim of this letter is to determine the minimum node density to achieve a connected large-scale ad hoc network, where every node has the same transmitting and receiving capabilities. Due to the log-normal shadowing, links are unidirectional in general. Contrary to the prevailing opinion, we argue that such asymmetries result into a “reduced” connectivity graph, which, from the point of view of MAC and routing protocols, is to be considered the true or effective connectivity graph. Accordingly, we derive a new formula for the connection probability between two nodes in order to compute global connectivity. Finally, theoretical findings, borrowed from random graphs theory, are compared to numerical simulation results in synthetic wireless network scenarios
A time-discrete extended Gamma process for time-dependent degradation phenomena
No full textThe non-stationary Gamma process is a widely used mathematical model to describe degradation phenomena whose growth rate at time t depends only on the current age of the item and not on the accumulated damage up to t. Nevertheless, the Gamma process is not a proper choice when there is empirical evidence that the variance-to-mean ratio of the process varies with time, because the Gamma process implies a constant variance-to-mean ratio. This paper proposes a generalization of the non-stationary Gamma process, which can be viewed as a time discretization of the extended Gamma process and allows one to describe time-dependent degradation phenomena whose variance varies with time t, not necessarily in proportion to the mean. A way to approximate the exact distribution of the degradation growth over a given time interval is given and a test for assessing whether the assumption of the Gamma process can be rejected or not is discussed. Finally, the proposed model is applied to a real dataset consisting of the sliding wear data of four metal alloy specimens
A random-effects model for long-term degradation analysis of solid oxide fuel cells
Full text linkSolid oxide fuel cells (SOFCs) are electrochemical devices converting the chemical energy into electricity with high efficiency and low pollutant emissions. Tough very promising, this technology is still in a developing phase, and degradation at the cell/stack level with operating time is still an issue of major concern. Methods to directly observe degradation modes and to measure their evolution over time are difficult to implement, and indirect performance indicators are adopted, typically related to voltage measurements in long-term tests. In order to describe long-term degradation tests, three components of the voltage measurements should be modelled: the smooth decay of voltage over time for each single unit; the variability of voltage decay among units; and the high-frequency small fluctuations of voltage due to experimental noise and lack of fit. In this paper, we propose an empirical random-effects regression model of polynomial type enabling to evaluate separately these three types of variability. Point and interval estimates are also derived for some performance measures, such as the mean voltage, the prediction of cell voltage, the reliability function and the cell-to-cell variability in SOFC stacks. Finally, the proposed methodology is applied to two real case-studies of long-term degradation tests of SOFC stacks
A Bayesian Estimation Procedure for the Non-Homogeneous Gamma Process
No full textWhen the degradation growth of a unit depends only on its age, a widely accepted model to describe the degradation phenomenon is the non-homogeneous gamma process, which proved to be suitable to model such degradation phenomena as wear, fatigue, corrosion, crack growth, erosion, and consumption. In this paper, a Bayesian inferential procedure using the Markov Chain Monte Carlo technique is proposed for the non-homogeneous gamma process with power-law shape function. All the process parameters are left to be assessed, and prior information is formalized on some quantities having a “physical” meaning. Both vague and informative priors are provided. Point and interval estimation of the process parameters and of some functions thereof are developed, as well prediction on some observable quantities that are useful in defining the maintenance strategy is proposed. Finally, the proposed procedure is applied to a real dataset consisting of the sliding wear data of four metal alloy specimens
An empirical mixed-effect regression model for comparative analysis of long-term SOFC degradation tests
No full textA New Estimation Algorithm for Interval Censored Data from Repairable Systems
Full text linkFor a minimally repaired system, whose failure process is described by a non-homogeneous Poisson process (NHPP), the classical maximum likelihood estimation procedures cannot be used when the system failures are hidden and detected only at inspection epochs. By assuming that the failure process follows a NHPP with power law intensity function, the Expectation-Maximization (EM) algorithm was recently proposed to estimate the model parameters and a procedure to test the presence of trend in the real failure data of some components of identical medical infusion pumps was discussed. However, the EM algorithm suffers in this application from some limitations due to its complexity and the large computational time required for convergence. This paper proposes a new estimation algorithm which is still iterative but, unlike the EM algorithm, is not based on the expectation of the log-likelihood function with respect to the conditional distribution of the unobserved data, but rather on the expectation of the conditioning variables, that is, of the unknown age of the system at the previous failure. This approach allows one to specify a simpler and much faster estimation procedure. A comparison between the proposed and the EM algorithms shows that the former performs as well as the latter, while requiring a drastically reduced computational burden. In addition, the proposed scheme can be applied to other intensity functions, such as the log-linear and the 2-parameter logarithmic functions. As a result, the test hypothesis of no trend in one of the analyzed datasets, which can not be rejected under the power law intensity function, is instead rejected under the alternative hypothesis of a log-linear intensity function
- …
