Artificial neural network modeling for surface roughness prediction in cylindrical grinding of Al‐SiCp metal matrix composites and ANOVA analysis

Metal matrix composites (MMC) having aluminium (Al) in the matrix phase and silicon carbide particles (SiCp) in reinforcement phase, ie Al‐SiCp type MMC, have gained popularity in the re‐cent past. In this competitive age, manufacturing industries strive to produce superior quality products at reasonable price. This is possible by achieving higher productivity while performing machining at optimum combinations of process variables. The low weight and high strength MMC are found suitable for variety of components

Corrosion and tribocorrosion behaviour of Al-Si-Cu-Mg alloy and its composites reinforced with B4C particles in 0.05M NaCl solution

The corrosion behaviour of metal matrix composites (MMCs) is strictly linked with the presence of heterogeneities such as reinforcement phase, microcrevices, porosity, secondary phase precipitates, and interaction products. Most of the literature related to corrosion behaviour of aluminium matrix composites (AMCs) is focused on SiC reinforced AMCs. On the other hand, there is very limited information available in the literature related to the tribocorrosion behaviour of AMCs. Therefore, the present work aims to investigate corrosion and tribocorrosion behaviour of Al-Si-Cu-Mg alloy matrix composites reinforced with B4C particulates. Corrosion behaviour of 15 and 19% (vol) B4C reinforced Al-Si-Cu-Mg matrix composites and the base alloy was investigated in 0.05M NaCl solution by performing immersion tests and potentiodynamic polarisation tests. Tribocorrosion behaviour of Al-Si-Cu-Mg alloy and its composites were also investigated in 0.05M NaCl solution. The tests were carried out against alumina ball using a reciprocating ball-on-plate tribometer. Electrochemical measurements were performed before, during, and after the sliding tests together with the recording of the tangential force. Results suggest that particle addition did not affect significantly the tendency of corrosion of Al-Si-Cu-Mg alloy without mechanical interactions. During the tribocorrosion tests, the counter material was found to slide mainly on the B4C particles, which protected the matrix alloy from severe wear damage. Furthermore, the wear debris were accumulated on the worn surfaces and entrapped between the reinforcing particles. Therefore, the tendency of corrosion and the corrosion rate decreased in Al-Si-Cu-Mg matrix B4C reinforced composites during the sliding in 0.05M NaCl solution. © 2013 Elsevier B.V.

Efeitos de diferentes histórias experimentais sobre o comportamento de seguir regras em participantes classificados de flexíveis e inflexíveis

O presente estudo investigou se a manutenção, ou não, do comportamento de seguir regras discrepantes das contingências de reforço programadas em situação experimental depende mais da história experimental do ouvinte ou da sua história pré-experimental, inferida das respostas destes a um questionário sobre inflexibilidade. Dezesseis estudantes universitários selecionados previamente com base em suas respostas a um questionário sobre inflexibilidade, foram expostos a um procedimento de escolha segundo o modelo. Em cada tentativa, um estímulo modelo e três de comparação eram apresentados ao participante, que deveria apontar para os três de comparação, em uma determinada seqüência. Os participantes foram atribuídos a duas condições e cada condição continha quatro fases. As condições diferiram somente quanto ao esquema de reforço utilizado. Na Condição 1 o esquema de reforço era contínuo (CRF) e na Condição 2 era de razão fixa (FR4). Nas duas condições a Fase 1 era iniciada com a apresentação de instruções mínimas e uma seqüência de respostas era estabelecida por reforço diferencial; a Fase 2 era iniciada com a apresentação de uma regra discrepante; a Fase 3 era iniciada com a apresentação de uma regra correspondente e a Fase 4 com a reapresentação da regra discrepante. Oito participantes (quatro classificados de flexíveis e quatro classificados de inflexíveis) foram expostos à Condição 1 (CRF) e oito participantes (quatro classificados de flexíveis e quatro classificados de inflexíveis) foram expostos à Condição 2 (FR4). Os resultados mostraram que independente da classificação, os oito participantes da Condição 1 abandonaram o seguimento da regra discrepante das contingências, indicando que o controle exercido pela história experimental construída, impediu a observação dos efeitos de variáveis pré-experimentais sobre o comportamento de seguir regras discrepantes dos participantes. Já os resultados da Condição 2 mostraram que os quatro participantes classificados de flexíveis abandonaram o seguimento da regra discrepante e os quatro participantes classificados de inflexíveis mantiveram o seguimento da regra discrepante das contingências, indicando que sob estas condições o controle por diferentes histórias pré-experimentais, prevaleceu. Comparativamente os resultados das duas condições permitem concluir que a manutenção do comportamento de seguir regras discrepantes não depende somente da história experimental ou da história pré-experimental do ouvinte, mas sim da combinação de um número de condições favoráveis ou desfavoráveis a manutenção do comportamento de seguir regra discrepante.

Discriminações condicionais após treino de pareamento consistente de estímulos complexos

O objetivo do presente estudo foi verificar a formação de discriminações condicionais através de um procedimento de treino por pareamento consistente de estímulos complexos. Na Fase 1, com quatro universitários, foi realizado o treino AF, BE e DC envolvendo estímulos modelos e de comparação simples e reforçamento diferencial explícito. Na Fase 2, houve o treino AB-E/F e AD-C/F, com modelos complexos e estímulos de comparação simples num formato de pareamento consistente. Após isso, os participantes foram expostos aos testes de transitividade e de equivalência. Na Fase 3, os treinos AB-E/F e AD-C/F foram desmembrados (simplificados) em AB-E; AB-F; AD-C; AD-F, e reaplicados os mesmos testes. Todos os participantes formaram as relações condicionais e dois deles formaram as relações emergentes, anterior e posterior ao desmembramento dos treinos. Os resultados sugerem que esse desmembramento foi um procedimento adequado para reverter o controle restrito de estímulo.

Nanofibrous-structured polymers and their engineering applications

Monolithic materials cannot always satisfy the demands of today’s advanced requirements. Only by combining several materials at different length-scales, as nature does, the requested performances can be met. Polymer nanocomposites are intended to overcome the common drawbacks of pristine polymers, with a multidisciplinary collaboration of material science with chemistry, engineering, and nanotechnology. These materials are an active combination of polymers and nanomaterials, where at least one phase lies in the nanometer range. By mimicking nature’s materials is possible to develop new nanocomposites for structural applications demanding combinations of strength and toughness. In this perspective, nanofibers obtained by electrospinning have been increasingly adopted in the last decade to improve the fracture toughness of Fiber Reinforced Plastic (FRP) laminates. Although nanofibers have already found applications in various fields, their widespread introduction in the industrial context is still a long way to go. This thesis aims to develop methodologies and models able to predict the behaviour of nanofibrous-reinforced polymers, paving the way for their practical engineering applications. It consists of two main parts. The first one investigates the mechanisms that act at the nanoscale, systematically evaluating the mechanical properties of both the nanofibrous reinforcement phase (Chapter 1) and hosting polymeric matrix (Chapter 2). The second part deals with the implementation of different types of nanofibers for novel pioneering applications, trying to combine the well-known fracture toughness enhancement in composite laminates with improving other mechanical properties or including novel functionalities. Chapter 3 reports the development of novel adhesive carriers made of nylon 6,6 nanofibrous mats to increase the fracture toughness of epoxy-bonded joints. In Chapter 4, recently developed rubbery nanofibers are used to enhance the damping properties of unidirectional carbon fiber laminates. Lastly, in Chapter 5, a novel self-sensing composite laminate capable of detecting impacts on its surface using PVDF-TrFE piezoelectric nanofibers is presented.

Policy gradient based Reinforcement Learning for real autonomous underwater cable tracking

This paper proposes a field application of a high-level reinforcement learning (RL) control system for solving the action selection problem of an autonomous robot in cable tracking task. The learning system is characterized by using a direct policy search method for learning the internal state/action mapping. Policy only algorithms may suffer from long convergence times when dealing with real robotics. In order to speed up the process, the learning phase has been carried out in a simulated environment and, in a second step, the policy has been transferred and tested successfully on a real robot. Future steps plan to continue the learning process on-line while on the real robot while performing the mentioned task. We demonstrate its feasibility with real experiments on the underwater robot ICTINEU AUV

Evaluation of Composite Pavement Unbonded Overlays: Phase III, August 2006

In recent years, thin whitetopping has evolved as a viable rehabilitation technique for deteriorated asphalt cement concrete (ACC) pavements. Numerous projects have been constructed and tested, allowing researchers to identify the important elements contributing to the projects’ successes. These elements include surface preparation, overlay thickness, synthetic fiber reinforcement usage, joint spacing, and joint sealing. Although the main factors affecting thin whitetopping performance have been identified by previous research, questions still existed as to the optimum design incorporating these variables. The objective of this research is to investigate the interaction between these variables over time. Laboratory testing and field testing were conducted to achieve the research objectives. Laboratory testing involved shear testing of the bond between the portland cement concrete (PCC) overlay and the ACC surface. Field testing involved falling weight deflectometer deflection responses, measurement of joint faulting and joint opening, and visual distress surveys on the 9.6-mile project. The project was located on Iowa Highway 13 extending north from the city of Manchester, Iowa, to Iowa Highway 3 in Delaware County. Variables investigated include ACC surface preparation, PCC thickness, slab size, synthetic fiber reinforcement usage, and joint spacing. This report documents the planning, construction, and performance of each variable in the time period from summer 2002 through spring 2006. The project has performed well with only minor distress identification since its construction.

Policy gradient based Reinforcement Learning for real autonomous underwater cable tracking

This paper proposes a field application of a high-level reinforcement learning (RL) control system for solving the action selection problem of an autonomous robot in cable tracking task. The learning system is characterized by using a direct policy search method for learning the internal state/action mapping. Policy only algorithms may suffer from long convergence times when dealing with real robotics. In order to speed up the process, the learning phase has been carried out in a simulated environment and, in a second step, the policy has been transferred and tested successfully on a real robot. Future steps plan to continue the learning process on-line while on the real robot while performing the mentioned task. We demonstrate its feasibility with real experiments on the underwater robot ICTINEU AUV

Development of alpha-SiAlON-SiC ceramic composites by liquid phase sintering

In this work, in situ alpha-SiAlON-SiC ceramic composites were obtained,by, liquid phase sintering, using SiC as reinforcement. Different beta-SiC powder contents (0-20 wt.%), were added to Si3N4-AlN-RE2O3. powder mixtures, and compacted by cold isostatic pressing. The samples were sintered at 1950 degrees C for 1 h, in N-2 atmosphere. Sintered: samples were characterized by relative density, weight loss, X-ray diffraction and scanning electron microscopy. Furthermore, mechanical properties such as hardness and fracture toughness were determined by Vickers indentation method. Lattice parameters of the alpha' phase did not considerably change with increase of SiC content. However, morphology, average grain size and aspect ratio of the alpha' phase were considerably changed with increase of the SiC content. These behavior influences significantly the mechanical properties of this hard ceramic composite. (C) 2006 Elsevier Ltd. All rights reserved.

Development and evaluation of a video-based modeling strategy for the dissemination phase of a school health tobacco -use prevention program

Even the best school health education programs will be unsuccessful if they are not disseminated effectively in a manner that encourages classroom adoption and implementation. This study involved two components: (1) the development of a videotape intervention to be used in the dissemination phase of a 4-year, NCI-funded diffusion study and (2) the evaluation of that videotape intervention strategy in comparison with a print (information transfer) strategy. Conceptualization has been guided by Social Learning Theory, Diffusion Theory, and communication theory. Additionally, the PRECEDE Framework has been used. Seventh and 8th grade classroom teachers from Spring Branch Independent School District in west Houston participated in the evaluation of the videotape and print interventions using a 57-item preadoption survey instrument developed by the UT Center for Health Promotion Research and Development. Two-way ANOVA was used to study individual score differences for five outcome variables: Total Scale Score (comprised of 57 predisposing, enabling, and reinforcing items), Adoption Characteristics Subscale, Attitude Toward Innovation Subscale, Receptivity Toward Innovation, and Reinforcement Subscale. The aim of the study is to compare the effect upon score differences of video and print interventions alone and in combination. Seventy-three 7th and 8th grade classroom teachers completed the study providing baseline and post-intervention measures on factors related to the adoption and implementation of tobacco-use prevention programs. Two-way ANOVA, in relation to the study questions, found significant scoring differences for those exposed to the videotape intervention alone for both the Attitude Toward Innovation Subscale and the Receptivity to Adopt Subscale. No significant results were found to suggest that print alone influences favorable scoring differences between baseline and post-intervention testing. One interaction effect was found suggesting video and print combined are more effective for influencing favorable scoring differences for the Reinforcement for the Adoption Subscale.^ This research is unique in that it represents a newly emerging field in health promotion communications research with implications for Social Learning Theory, Diffusion Theory, and communication science that are applicable to the development of improved school health interventions. ^

Evaluating the reinforcement of inorganic fullerene-like nanoparticles in thermoplastic matrices by depth-sensing indentation

The reinforcing effect of inorganic fullerene-like tungsten disulfide (IF-WS2) nanoparticles in two different polymer matrices, isotactic polypropylene (iPP) and polyphenylene sulfide (PPS), has been investigated by means of dynamic depth-sensing indentation. The hardness and elastic modulus enhancement upon filler addition is analyzed in terms of two main contributions: changes in the polymer matrix nanostructure and intrinsic properties of the filler including matrix-particle load transfer. It is found that the latter mainly determines the overall mechanical improvement, whereas the nanostructural changes induced in the polymer matrix only contribute to a minor extent. Important differences are suggested between the mechanisms of deformation in the two nanocomposites, resulting in a moderate mechanical enhancement in case of iPP (20% for a filler loading of 1%), and a remarkable hardness increase in case of PPS (60% for the same filler content). The nature of the polymer amorphous phase, whether in the glassy or rubbery state, seems to play here an important role. Finally, nanoindentation and dynamic mechanical analysis measurements are compared and discussed in terms of the different directionality of the stresses applied.

Directed Selection using Reinforcement Learning for the Examination Timetabling Problem

Traditional heuristic approaches to the Examination Timetabling Problem normally utilize a stochastic method during Optimization for the selection of the next examination to be considered for timetabling within the neighbourhood search process. This paper presents a technique whereby the stochastic method has been augmented with information from a weighted list gathered during the initial adaptive construction phase, with the purpose of intelligently directing examination selection. In addition, a Reinforcement Learning technique has been adapted to identify the most effective portions of the weighted list in terms of facilitating the greatest potential for overall solution improvement. The technique is tested against the 2007 International Timetabling Competition datasets with solutions generated within a time frame specified by the competition organizers. The results generated are better than those of the competition winner in seven of the twelve examinations, while being competitive for the remaining five examinations. This paper also shows experimentally how using reinforcement learning has improved upon our previous technique.

Techniques for batch reinforcement learning in robotics

This thesis addresses the Batch Reinforcement Learning methods in Robotics. This sub-class of Reinforcement Learning has shown promising results and has been the focus of recent research. Three contributions are proposed that aim to extend the state-of-art methods allowing for a faster and more stable learning process, such as required for learning in Robotics. The Q-learning update-rule is widely applied, since it allows to learn without the presence of a model of the environment. However, this update-rule is transition-based and does not take advantage of the underlying episodic structure of collected batch of interactions. The Q-Batch update-rule is proposed in this thesis, to process experiencies along the trajectories collected in the interaction phase. This allows a faster propagation of obtained rewards and penalties, resulting in faster and more robust learning. Non-parametric function approximations are explored, such as Gaussian Processes. This type of approximators allows to encode prior knowledge about the latent function, in the form of kernels, providing a higher level of exibility and accuracy. The application of Gaussian Processes in Batch Reinforcement Learning presented a higher performance in learning tasks than other function approximations used in the literature. Lastly, in order to extract more information from the experiences collected by the agent, model-learning techniques are incorporated to learn the system dynamics. In this way, it is possible to augment the set of collected experiences with experiences generated through planning using the learned models. Experiments were carried out mainly in simulation, with some tests carried out in a physical robotic platform. The obtained results show that the proposed approaches are able to outperform the classical Fitted Q Iteration.

Unsupervised reinforcement learning via state entropy maximization

Reinforcement Learning (RL) provides a powerful framework to address sequential decision-making problems in which the transition dynamics is unknown or too complex to be represented. The RL approach is based on speculating what is the best decision to make given sample estimates obtained from previous interactions, a recipe that led to several breakthroughs in various domains, ranging from game playing to robotics. Despite their success, current RL methods hardly generalize from one task to another, and achieving the kind of generalization obtained through unsupervised pre-training in non-sequential problems seems unthinkable. Unsupervised RL has recently emerged as a way to improve generalization of RL methods. Just as its non-sequential counterpart, the unsupervised RL framework comprises two phases: An unsupervised pre-training phase, in which the agent interacts with the environment without external feedback, and a supervised fine-tuning phase, in which the agent aims to efficiently solve a task in the same environment by exploiting the knowledge acquired during pre-training. In this thesis, we study unsupervised RL via state entropy maximization, in which the agent makes use of the unsupervised interactions to pre-train a policy that maximizes the entropy of its induced state distribution. First, we provide a theoretical characterization of the learning problem by considering a convex RL formulation that subsumes state entropy maximization. Our analysis shows that maximizing the state entropy in finite trials is inherently harder than RL. Then, we study the state entropy maximization problem from an optimization perspective. Especially, we show that the primal formulation of the corresponding optimization problem can be (approximately) addressed through tractable linear programs. Finally, we provide the first practical methodologies for state entropy maximization in complex domains, both when the pre-training takes place in a single environment as well as multiple environments.

Predictive Value Of Phase I Trials For Safety In Later Trials And Final Approved Dose: Analysis Of 61 Approved Cancer Drugs.

Phase I trials use a small number of patients to define a maximum tolerated dose (MTD) and the safety of new agents. We compared data from phase I and registration trials to determine whether early trials predicted later safety and final dose. We searched the U.S. Food and Drug Administration (FDA) website for drugs approved in nonpediatric cancers (January 1990-October 2012). The recommended phase II dose (R2PD) and toxicities from phase I were compared with doses and safety in later trials. In 62 of 85 (73%) matched trials, the dose from the later trial was within 20% of the RP2D. In a multivariable analysis, phase I trials of targeted agents were less predictive of the final approved dose (OR, 0.2 for adopting ± 20% of the RP2D for targeted vs. other classes; P = 0.025). Of the 530 clinically relevant toxicities in later trials, 70% (n = 374) were described in phase I. A significant relationship (P = 0.0032) between increasing the number of patients in phase I (up to 60) and the ability to describe future clinically relevant toxicities was observed. Among 28,505 patients in later trials, the death rate that was related to drug was 1.41%. In conclusion, dosing based on phase I trials was associated with a low toxicity-related death rate in later trials. The ability to predict relevant toxicities correlates with the number of patients on the initial phase I trial. The final dose approved was within 20% of the RP2D in 73% of assessed trials.