Simulated annealing for maximum a posteriori parameter estimationof hidden Markov models

Hidden Markov models are mixture models in which the populations from one observation to the next are selected according to an unobserved finite state-space Markov chain. Given a realization of the observation process, our aim is to estimate both the parameters of the Markov chain and of the mixture model in a Bayesian framework. We present an original simulated annealing algorithm which, in the same way as the EM (expectation-maximization) algorithm, relies on data augmentation, and is based on stochastic simulation of the hidden Markov chain. This algorithm is shown to converge toward the set of maximum a posteriori (MAP) parameters under suitable regularity conditions     

Extraction of Finite Element Basis Functions From the Cellular Topology of Meshes

This paper is devoted to the presentation of a simple but powerful topological approach for the computation of basis functions involved in the expression of finite element interpolants. It deals with the so-called nodal and edge elements but similar considerations can be made for the construction of facet ones. All standard cells are treated including pyramids.     

Co/SiO2 catalysts prepared from Co2(CO)8 for CO hydrogenation into alcohols and hydrocarbons: characterization by magnetic methods and temperature-programmed hydrogenation

The hydrogenation of carbon monoxide over Co/SiO₂ catalysts obtained by reduction of precursors prepared by the reaction of dicobaltoctacarbonyl with silica gives rise to a mixture of alcohols and hydrocarbons. The chain growth probability for hydrocarbon synthesis is similar to that observed over conventional catalysts prepared by the ammonia method (=0.74) with a comparable metal dispersion (metal particle size 4 nm). Alcohol formation yields a smaller chain growth probability (=0.42). Magnetic measurements have shown that in the latter catalyst, small cobalt clusters (1 nm) are likely to be present together with the cobalt particles (4 nm), this observation reconciles apparently conflicting results from the literature. It has been speculated that alcohol formation might be related to the presence of these clusters according to a concerted mechanism. Furthermore, the observation of large amounts of acetates by temperature-programmed hydrogenation, not detected in hydrocarbon selective catalysts, suggests that they may play a role in ethanol synthesis.     

New temperature dependent thermal conductivity data for water-based nanofluids

This paper presents effective thermal conductivity measurements of alumina/water and copper oxide/water nanofluids. The effects of particle volume fraction, temperature and particle size were investigated. Readings at ambient temperature as well as over a relatively large temperature range were made for various particle volume fractions up to 9%. Results clearly show the predicted overall effect of an increase in the effective thermal conductivity with an increase in particle volume fraction and with a decrease in particle size. Furthermore, the relative increase in thermal conductivity was found to be more important at higher temperatures. Obtained results compare favorably with certain data sets and theoretical models found in current literature.     

Optimisation of extraction of phenolics and antioxidants from black currant leaves and buds and of stability during storage

Health benefits of a diet rich in fruits and vegetables are attributed in part to their contents of phenolics and other antioxidant compounds. In this research, the extraction of phenolics and antioxidant compounds from black currant was optimised for different plant organs. The extraction solvent affected yield: aqueous acetone was better than methanol and acetate or glycine buffer. In aqueous buffer, maximum yields of total phenolics and antioxidant activities were obtained at pH 3. Extraction from lyophilised materials yielded extracts with higher phenolic contents and antioxidant activities.     

Experimental characterization of the chaotic dynamics of cohesionless particles: application to a V-blender

So far, most of the investigations of the dynamics of granular material in blenders have been done in 2D tumblers due to the current lack of accurate measurement methods for the investigation of complex 3D flows. However, recent advances in the field of non-intrusive methods have paved the way to the characterization of the chaotic dynamics in 3D blenders. This work aims to present such an analysis in the case of a V-blender using radioactive particle tracking (RPT). Special attention is given to the chaotic properties of mixing by performing a time-series analysis of the position of a single tracer in motion in this blender. An original investigation of the mixing properties is also presented. More generally, this work shows that the mixing properties of a blender can be characterized by using tools of dynamical systems theory through a time-series analysis of data obtained from non-intrusive measurements.     

Ultrasonic welding of thermoplastic composites: a numerical analysis at the mesoscopic scale relating processing parameters,flow of polymer and quality of adhesion

Ultrasonic continuous welding of thermoplastic composite plates is a very promising process of particular interest for the assembly of aeronautics large parts. Its modeling and simulation however suffers from the difficulty of accounting for the very different time scales that rule the thermo-mechanical phenomena at the level of the adhesion zone. This problem was addressed in our previous works and led to an original simulation tool presented in Levy et al. (Eur J Mech A, Solids 30(4):501–509, 2011a). In this paper, the adopted time-homogenized multiphysical modeling of the flow at the mesoscopic scale of the energy directors is first presented. Then, using the numerical software in a 2D approach, an extensive numerical parametric study of the process is presented. The phenomena allowing welding are confirmed to be an initial strain concentration in the energy director, and the formation of a flowing fold. The influence of the following process parameters are finally investigated: amplitude of vibrations, holding force of the sonotrode, thickness of the plates, radius of curvature at the tip of the director, angle of the director. Process efficiency and weld quality is evaluated through simple indicators such as the equivalent stiffness analysis, the healing degree and the risk of porosity entrapment. The present study, carried at the mesoscopic scale, provides a better understanding of the complex interactions between physical and process parameters and enables to draw important technological conclusions for the design of energy directors.