EDUARDO LOBO LUSTOSA CABRAL

EDUARDO LOBO LUSTOSA CABRAL

Página do item completo

Resultados de Busca

Agora exibindo 1 - 10 de 51

Epithermal lead-cooled micro-reactor using fuel-moderator assemblies
2024 - LEE, S.M.; MATTAR, M.; CABRAL, E.L.L.
Assessing shielding thickness in Am-241 nuclear battery
2024 - ANTUNES, P.C.G.; SOUZA, C.D. de; SHORTO, J.M.B.; BELCHIOR JUNIOR, A.; JUNQUEIRA, F.C.; ZEITUNI, C.A.; CABRAL, E.L.L.; RIBEIRO, M.A.M.
On the use of AI formetamodeling
2024 - DRIEMEIER, LARISSA; CABRAL, EDUARDO L.L.; RODRIGUES, GABRIEL L.; TSUZUKI, MARCOS; ALVES, MARCILIO; COSTA, LUCAS P. da; MOURA, RAFAEL T.
In scenarios where complex analyses are routinely conducted on similar structures, such as in a redesign process to meet performance requirements or when input parameters require frequent adjustments within a specified domain, a practical approach involves the use of metamodels calibrated using machine learning methodologies. In our investigation, we introduce a metamodel that utilizes an artificial neural network to analyze 3D nonlinear structures undergoing plastic deformations and large strains. Snap-through and snapback behaviors are addressed through network training, which is based on 10,000 Force vs Displacement curves (target outputs) obtained from nonlinear finite element analyses. This interplay between finite element analysis and machine learning, as demonstrated here, exhibits promising potential as an effective technique. The results indicate that the proposed deep neural network can learn from the simulations of finite elements. The discussion explores scenarios where the utilization of AI in the analysis of nonlinear structures is justified.
Radiation shielding for a nuclear fusion device with inertial electrostatic confinement
2022 - LEE, S.M.; YORIYAZ, H.; CABRAL, E.L.L.
In an inertial electrostatic confinement nuclear fusion device, IECF, thermal neutron population is created near the neutron shielding that is proportional to the fast neutrons generation rate; nevertheless, this proportionality varies with the experimental arrangement. Thus, to properly measure the fast neutron generation rate by the IECF device it is necessary to previously elaborate a suitable neutron transport model between the IECF device and the radiation shield, where the neutron detector will be located. This model is elaborated using the Monte Carlo N-Particle Code and the same is used to design the required radiation shield for the safe operation of the device.
Análise de viabilidade do emprego de Reatores Modulares Pequenos (Small Modular Reactors – SMR) no Brasil
2022 - LIMA, ANA C. de S.; CABRAL, EDUARDO L.L.; SABUNDJIAN, GAIANE; TERREMOTO, LUIS A.A.; ROCHA, MARCELO da S.
Este trabalho apresenta uma análise SWOT sobre os Reatores Modulares Pequenos (Small Modular Reactor- SMR), a fim de avaliar a viabilidade de implantação desses reatores nucleares no Brasil. A análie PESTLA foi utilizada como coadjuvante da análise SWOT servindo para auxiliar na categorização dos fatores considerados de maior relevância no sentido de possibilitar um melhor entendimento das condições de contorno relativas à implantação dos SMRs no Brasil. A análise PESTLA, envolve um estudo dos aspectos Tecnológico, Ambiental, Político, Social (Recursos Humanos/Infraestrutura), Econômico e Legal. As análises SWOT e PESTLA consideraram diversos aspectos no âmbito da instituição governamental responsável pela orientação e planejamento do programa nuclear brasileiro, a CNEN, que através de suas unidades desenvolve atividades de pesquisa e formação especializada na área nuclear. A metodologia adotada neste estudo selecionou os pontos positivos e negativos tanto da instituição quanto dos SMRs. Os reaotres modulares descritos neste trabalho são do tipo Pressurized Water Reactor (PWR) e que se encontram em estágio avançado de desenvolvimento, são eles: CAREM, KLT-40S, SMART e NuScale. O estudo realizado neste documento possibilitará a tomada de decisão sobre a utilização de SMRs no Brasil.
Reinforcement learning control of robot manipulator
2021 - COTRIM, LUCAS P.; JOSE, MARCOS M.; CABRAL, EDUARDO L.L.
Since the establishment of robotics in industrial applications, industrial robot programming involves the repetitive and time-consuming process of manually specifying a fixed trajectory, resulting in machine idle time in production and the necessity of completely reprogramming the robot for different tasks. The increasing number of robotics applications in unstructured environments requires not only intelligent but also reactive controllers due to the unpredictability of the environment and safety measures, respectively. This paper presents a comparative analysis of two classes of Reinforcement Learning algorithms, value iteration (Q-Learning/DQN) and policy iteration (REINFORCE), applied to the discretized task of positioning a robotic manipulator in an obstacle-filled simulated environment, with no previous knowledge of the obstacles’ positions or of the robot arm dynamics. The agent’s performance and algorithm convergence are analyzed under different reward functions and on four increasingly complex test projects: 1-Degree of Freedom (DOF) robot, 2-DOF robot, Kuka KR16 Industrial robot, Kuka KR16 Industrial robot with random setpoint/obstacle placement. The DQN algorithm presented significantly better performance and reduced training time across all test projects, and the third reward function generated better agents for both algorithms.
An end-to-end approach to autonomous vehicle control using deep learning
2021 - NOVELLO, GUSTAVO A.M.; YAMAMOTO, HENRIQUE Y.; CABRAL, EDUARDO L.L.
The objective of this work is to develop an autonomous vehicle controller inside Grand Theft Auto V game, used as a simulation environment. It is used an end-to-end approach, in which the model maps directly the inputs from the image of a car hood camera and a sequence of speed values to three driving commands: steering wheel angle, accelerator pedal pressure and brake pedal pressure. The developedmodel is composed of a convolutional neural network and a recurring neural network. The convolutional network processes the images and the recurrent network processes the speed data. Themodel learns fromdata generated by a human driver´s commands. Two interfaces are developed: one for collecting in-game training data and another to verify the performance of themodel for the autonomous vehicle control. The results show that themodel after training is capable to drive the vehicle as well as a human driver. This proves that a combination of a convolutional network with a recurrent network, using an end-to-end approach, is capable of obtaining a good driving performance even using only images and speed velocity as sensory data.
Radiation shielding for a nuclear fusion device with inertial electrostatic confinement
2021 - LEE, S.M.; YORIYAZ, H.; CABRAL, E.L.L.
Pixel-position-based lossless image compression algorithm
2019 - CABRAL, EDUARDO L.L.; SABUNDJIAN, GAIANE; CONTI, THADEU das N.
In this paper we present a novel lossless image compression method that is very simple and fast. The method uses linear prediction followed by arithmetic coding. Different prediction functions are used to estimate the intensity of image pixels. Two variants of the prediction algorithm are presented. One variant uses two different prediction functions and the other uses three different prediction functions. The position of the pixel in the image determines which prediction function is used. The method can be applied for images of any size and of high bit-depths. Standard images available in the literature are used to test the method. The compression ratios obtained with the proposed method are compared with the compression ratios obtained with the JPEG-LS and JPEG2000 methods and the results are satisfactory.
Development of neutron shielding for an inertial electrostatic confinement nuclear fusion device
2019 - LEE, SEUNG M.; YORIYAZ, HELIO; CABRAL, EDUARDO L.L.
This work aims to develop a suitable neutron shielding for an Inertial Electrostatic Confinement Nuclear Fusion device (IECF). Neutrons are generated in the IECF device as results of nuclear fusion reactions and their detection is fundamental for the development of the IECF device, because experimental data is needed to perform efficiency analysis and model validation. Nevertheless, it is essential to moderate the neutrons down to the thermal state to make it possible to detect those using conventional detectors. Therefore, to properly measure the fast neutron generation rate by the IECF device it is necessary to previously elaborate a detailed neutron transport model between the IECF device and the radiation shielding, where the neutron detector will be located. In this work, a model is elaborated using the Monte Carlo N-Particle Code and is used to design the required radiation shielding for the device. Later, the same model will be used to determine the proportionality factor between the fast neutron generation in the IECF device and the thermal neutron population in the shielding.