Papers by Author: Juana Coello

Abstract: The aim of this investigation is to analyze the performance of several supervised machine learning algorithms for solving the automatic classification problem of steel image microstructures. We conducted an experiment using a public-domain dataset of Ultra High Carbon Steel Micrographs (UHCSM). This image database consists of a collection of scanning electron micrographs (SEM) taken from samples of a commercial roll-mill casting with a nominal carbon of 2%. Heat treatments such as annealing, water quenching, air and furnace cooling were performed on steel samples so primary microconstituents could be found in micrographs. Each of these microconstituents defines each of the categories of classification to be accomplished by machine learning algorithms. The heat treatments brought about 4 usable classes (sets of images) of primary microconstituents: pearlite, spheroidite, proeutectoid cementite network, pearlite containing spheroidite. All labeled images are prepared to improve models' accuracy in a preprocessing stage so that the image dataset is ready for feature extraction. In order to develop classification models, we put to the test distinct machine learning approaches by working with Matlab's classification learner application where we perform automated training to search for the best classification model type, including Decision Trees, Support Vector Machines (SVM), Discriminant Analysis, Nearest Neighbors, Naive Bayes, Ensemble k-NN, and Neural Network classification. For obtaining the features of the images (feature extraction) we choose the method of Bag-of-features with 400 words for the first experiment, and 327 words by removing less important features for a second experiment. The experimented models reached very different accuracy values on training, with SVM as the best classifier which gets 91.6% accuracy. We can conclude that classic machine learning algorithms solve the classification, but an accuracy improvement can be reached by investigating deep learning techniques.

Abstract: Titanium alloys have been reported as potential materials for aeronautical and automotive applications due to their interesting mechanical properties, combined with their low density. The manufacturing processes developed for these alloys require finishing machining operations to improve the surface quality of the parts and to meet the desired geometrical tolerances. Nevertheless, titanium aluminides exhibit extremely low machinability in comparison to traditional titanium alloys. The combination of the low thermal diffusivity of these materials and the high chemical affinity and friction coefficient with the cutting tools accelerate tool wear phenomena and lead to a deterioration of the part surface quality. Moreover, the mechanical properties of titanium aluminides contribute to increase the cutting forces which generates tool repulsion resulting in undesirable vibration or chatter phenomena. In this paper, the machining suitability of the turning process of Ti48Al2Cr2Nb titanium aluminide has been evaluated based on the analysis of chatter phenomena and the inspection of the surface roughness and roundness tolerance of the machined part. Experimental turning tests have been carried out by varying the main parameters of the process, cutting speed, feed rate and tool geometry, with the objective of determining the best cutting combination. For this purpose, a harmonic analysis methodology of the roundness profile based on the application of the discrete Fourier transform (DFT) has been employed. This technique has made it possible to isolate the vibration-induced machining effects from the lower frequency defects generated by part bending and to relate them to the surface quality and geometrical accuracy of the machined part.

Abstract: The present work develops a learning methododology based on experiments related to the cutting temperature concept in turning processes. This proposal allows students to measure the temperature actually reached during a typical turning operation with a semi-automatic lathe. Temperature data are collected by a thermographic camera, which implies acquiring competences in this technique. The different tests involved in the practical experiment are defined for various cutting speeds and feed rates, and for a constant depth of cut. Two different materials are considered to point out the influence of turning parameters on cutting temperature.

Abstract: This work develops a methodology that allowed students to associate the effect of some cutting variables with chip type. For that, turning processes were carried out on two different materials since their physical and mechanical properties have a relevant influence on the chip formation. Cutting tests were run by varying the feed rate, while cutting speed and depth of cut had fixed values. These cutting conditions led to different chip geometries being obtained. Thus, it was possible to establish a chip type classification in turning operations. The methodology herein presented is based on new technologies, which raised great expectations among students, who positively considered this experience. The effect of other variables, apart from feed rate, must be theoretically analyzed in order to improve the comprenhension of the process herein involved.

Abstract: TRIP steels, or Transformed Induced Plasticity steels, have excellent mechanical properties if compared with conventional steels. The highlighted characteristic of these steels is that they modify the microstructure with the deformation process as part of the austenite transforms to martensite, with the consequent change of the material properties. One of the main problems of TRIP steels is strong elastic recovery, or springback, after forming. In this work, the springback phenomenon is evaluated by bending tests and the influence of the variables involved in it is determined. Experimental bending forces do not agree with theoretical predictions that are proposed in the literature. In spite of the bending radius having been considered an influence factor in the process, this work demonstrates that the aforementioned factor has a minor influence, at least for TRIP 800. The factor found to affect material recovery the most was the bending angle.

Abstract: As it is well-known, TRIP 800 steels modify their structure with the deformation grade. So, part of the retained austenite turns into martensite by plastic deformation. The usual techniques tried out to evaluate this transformation whether do not lead to obtain good results or the experimentation with them is very complex. In this work, a magnetic induction method is experimented and developed in order to determine the evolution of a TRIP 800 steel microstructure with the strain grade. The variables that can have influence on this kind of analysis methodology have been studied and their effects evaluated. This method has been applied to determine the induced martensite by deformation under conditions of pure shear deformation. Results point that this method allows to state that the microstructure evolution taking place in TRIP 800 steels is not proportional to the strain applied.

Abstract: A traditional educational strategy for teaching sheet stamping processes consists in theoretical planning. Having experimental media available to conduct stamping processes in university teaching laboratories is not that usual. In this work Problem Based Learning is carried out. Students must evaluate some data from deep-drawing experiences that have been previously done by the teaching team. The innovation proposal starts with deep-drawing experimental demonstrations performed by an experimental device that has been designed in the laboratory for researching and teaching aims. Processing the information that students acquire helps them to analyse the main deep-drawing technological fundamentals.