Parameter identification of hardening laws for bulk metal forming using experimental and numerical approach

The simulative prediction of material behaviour in forming processes necessitates a precise determination of the material parameters. The present work focusses on the modelling of the isostatic part of the flow stress using a flow curve with an analytical suppression of the influence of friction and an adequate analytical law. The experimental data are obtained from isothermal upsetting tests with various upsetting ratios. The different ratios are based on a variation of the height of the sample, remaining the diameter constant. For the proposed flow stress law five parameters are identified. In order to decrease the number of function evaluations, a new reduction model method based on both analytical and sequential quadratic programming (SQP) algorithms is developed and applied to identify flow stress law parameters. A comparison with traditional SQP algorithm is also done. A 3D finite element model is built in order to simulate a side pressing test and an experimental validation is done. As numerical results fit very well experimental data, the proposed model achieves a precise prediction of the flow behaviour. The identification of the other parts of the model (i.e. dependencies on strain-rate and temperature) are conducted in further works.     

Cross-modal interactions between taste and smell: Odour-induced saltiness enhancement depends on salt level

In recent years, health concerns related to salt have led to extensive research on low-salt foods. One major remaining issue is to reduce sodium content while still providing the salty fix, to maintain food appreciation, acceptance and choice. Using well-selected odours has been proposed to compensate for sodium chloride reduction in food, due to the cross-modal interactions between odour and taste. However, interactions between taste and odour, with regards to saltiness, remain poorly investigated. Moreover, although a few studies have assessed the influences of odour and sweet supra-threshold intensity on cross-modal interactions, none focused directly on how they influence saltiness. Therefore, the aim of the present study is to determine if the level of saltiness, i.e., the concentration of sodium chloride in aqueous solutions, influences odour-induced saltiness enhancement (OISE). A panel of 64 untrained panellists evaluated the saltiness of water solutions including salt and a tasteless odorant. Following a full factorial design, three concentrations of salt were used in conjunction with three aroma conditions. The results confirmed that OISE depends on odour-taste congruency but also clearly indicated that OISE depends on salt concentration (salty intensity): the saltiness of a low- or medium-salt-content solution increased significantly when subjects perceived simultaneously the congruent sardine aroma, but OISE was no more significant with high-saltiness solutions. This effect has to be taken into account when using OISE as a compensation strategy in low-salt food.     

Electrostatic painting of insulating surfaces

The purpose of this work was to develop methods for painting insulating surfaces by the conventional electrostatic painting technique used to coat conducting samples. This technique is not applicable to insulators since the charges carried by the paint accumulate on insulating targets, thereby preventing a proper coating. A detailed investigation of the decay of charges deposited on insulators by a corona discharge simulating the charged paint has confirmed that a grounded counterelectrode (CE) on the back face of the sample speeds up the decay and that the better the contact, the faster the decay. Because a material CE with proper contacts is inconvenient for painting in production configuration, it was suggested that ionized air might behave as a satisfactory CE. Several ionizers have been tested, and good coatings have been obtained, provided that the flows of paint and ionized air are well separated. The possibility of self-creation of ionized air by the voltage of the early paint deposit has also been successfully tested     

Variational Bayesian inference for infinite generalized inverted Dirichlet mixtures with feature selection and its application to clustering

We developed a variational Bayesian learning framework for the infinite generalized Dirichlet mixture model (i.e. a weighted mixture of Dirichlet process priors based on the generalized inverted Dirichlet distribution) that has proven its capability to model complex multidimensional data. We also integrate a “feature selection” approach to highlight the features that are most informative in order to construct an appropriate model in terms of clustering accuracy. Experiments on synthetic data as well as real data generated from visual scenes and handwritten digits datasets illustrate and validate the proposed approach.     

On Bayesian analysis of a finite generalized Dirichlet mixture via a Metropolis-within-Gibbs sampling

In this paper, we present a fully Bayesian approach for generalized Dirichlet mixtures estimation and selection. The estimation of the parameters is based on the Monte Carlo simulation technique of Gibbs sampling mixed with a Metropolis-Hastings step. Also, we obtain a posterior distribution which is conjugate to a generalized Dirichlet likelihood. For the selection of the number of clusters, we used the integrated likelihood. The performance of our Bayesian algorithm is tested and compared with the maximum likelihood approach by the classification of several synthetic and real data sets. The generalized Dirichlet mixture is also applied to the problems of IR eye modeling and introduced as a probabilistic kernel for Support Vector Machines.

High-dimensional unsupervised selection and estimation of a finite generalized Dirichlet mixture model based on minimum message length

We consider the problem of determining the structure of high-dimensional data, without prior knowledge of the number of clusters. Data are represented by a finite mixture model based on the generalized Dirichlet distribution. The generalized Dirichlet distribution has a more general covariance structure than the Dirichlet distribution and offers high flexibility and ease of use for the approximation of both symmetric and asymmetric distributions. This makes the generalized Dirichlet distribution more practical and useful. An important problem in mixture modeling is the determination of the number of clusters. Indeed, a mixture with too many or too few components may not be appropriate to approximate the true model. Here, we consider the application of the minimum message length (MML) principle to determine the number of clusters. The MML is derived so as to choose the number of clusters in the mixture model which best describes the data. A comparison with other selection criteria is performed. The validation involves synthetic data, real data clustering, and two interesting real applications: classification of web pages, and texture database summarization for efficient retrieval.