Teresa Guraya

Automatic morphological categorisation of carbon black nano-aggregates

Nano-technology is the study of matter behaviour on atomic and molecular scale (i.e. nano-scale). In particular, carbon black is a nano-material generally used for the reinforcement of rubber compounds. Nevertheless, the exact reason behind its success in this concrete domain remains unknown. Characterisation of rubber nano-aggregates aims to answer this question. The morphology of the nano-aggregate takes an important part in the final result of the compound. Several approaches have been taken to classify them. In this paper we propose the first automatic machine-learning-based nano-aggregate morphology categorisation system. This method extracts several geometric features in order to train machine-learning classifiers, forming a constellation of expert knowledge that enables us to foresee the exact morphology of a nano-aggregate. Furthermore, we compare the obtained results and show that Decision Trees outperform the rest of the counterparts for morphology categorisation.

A methodology for finding the optimal iteration number of the SIRT algorithm for quantitative Electron Tomography

The SIRT (Simultaneous Iterative Reconstruction Technique) algorithm is commonly used in Electron Tomography to calculate the original volume of the sample from noisy images, but the results provided by this iterative procedure are strongly dependent on the specific implementation of the algorithm, as well as on the number of iterations employed for the reconstruction. In this work, a methodology for selecting the iteration number of the SIRT reconstruction that provides the most accurate segmentation is proposed. The methodology is based on the statistical analysis of the intensity profiles at the edge of the objects in the reconstructed volume. A phantom which resembles a a carbon black aggregate has been created to validate the methodology and the SIRT implementations of two free software packages (TOMOJ and TOMO3D) have been used.

A parameter for the assessment of the segmentation of TEM tomography reconstructed volumes based on mutual information

A method is proposed and verified for selecting the optimum segmentation of a TEM reconstruction among the results of several segmentation algorithms. The selection criterion is the accuracy of the segmentation. To do this selection, a parameter for the comparison of the accuracies of the different segmentations has been defined. It consists of the mutual information value between the acquired TEM images of the sample and the Radon projections of the segmented volumes. In this work, it has been proved that this new mutual information parameter and the Jaccard coefficient between the segmented volume and the ideal one are correlated. In addition, the results of the new parameter are compared to the results obtained from another validated method to select the optimum segmentation.

Methodology to classify the shape of reinforcement fillers: optimization, evaluation, comparison, and selection of models

Abstract This work applies statistical analysis, and classical and advanced machine learning algorithms to classify 7714 aggregates into four categories according to their shape. The aggregates under study are obtained from several grades of carbon black: Vulcan XC 605, Vulcan XC 72, CSX 691, Printex 25, N990, and N762. The classification of the shape is of great significance in order to explain and predict the end-use properties of the composite materials, like mechanical properties. The proposed approach combines transmission electron microscopy and automated image analysis to obtain the dataset of the morphological features that defines the shape of the aggregate, and statistical analysis and machine learning techniques to create the classification models using feature transformation and reduction, parameter tuning, and validation methods in order to achieve robust classification models. The best result is obtained from a classification tree based on evolutionary algorithms with a principal component analysis-based feature reduction that reports an acceptable accuracy, thereby validating both the final chosen model and the methodology.

On the perceptions of students and professors in the implementation of an inter-university engineering PBL experience

ABSTRACT The new educational paradigm has led to a change in the teaching methodologies toward those more focused on the student, among these, project-based learning (PBL) is postulated as one of the most promising. This work is focused on the description of the experience of using PBL methodology in Materials Science courses, conducted by four different Spanish universities on different engineering degrees. The other main objective is to analyse and evaluate how the PBL was perceived by the students and the lecturers that took part in the PBL process. This investigation was an embedded, sequential mixed-methods study, which began by administering a survey to one hundred and four students and then a focus group with six students and six lecturers in four different engineering degrees of four Universities. Results generally show a good degree of acceptance of this approach by all parties involved.

Multivariate Analysis of Composition Features to Perform Linear Predictions of Rubber Blends Tensile Strength

The goal in this work is to build a multivariate linear model to predict tensile strength since is one of the most significant mechanical properties of carbon-black reinforced rubber blends. This model is based in the relationship between the final mechanical properties and the material composition, with the advantage of using this model to improve the design of the composition of the blend. In order to predict this relevant physical attribute of rubber blends a linear regression is performed, but previously a multivariate analysis of the data is done to get a better accuracy in the validation of the model. The number of used instances and the values are determined by a Taguchi design of experiments, and the output values are obtained from the tensile strength test following the corresponding standard. After the performance of the multivariate analysis where the input variables are under a detail study, a selection of the best features help to improve the accuracy of the model, passing from a 24.80% to a 20.60% of error.