Susana Marrero Iglesias

Prof. Dr. Susana Marrero Iglesias

orcid

Research Focus & Metrics

Academic Career Length

19Years (2008-Present)

(Calculated once Scilit publication data is synced: current year − year of first publication + 1)

Representative Publications

View All Publications
13 May 2015 in IEEE Latin America Transactions
Verified
Journal Article
Closed access
JDJoel Sanchez Dominguez
Joel Sanchez Dominguez
AJAdailton J. Cerqueira Junior
Adailton J. Cerqueira Junior
DDDany Sanchez Dominguez
Dany Sanchez Dominguez
DFDiego Frias
Diego Frias
SISusana Marrero Iglesias
Susana Marrero Iglesias
In this work we describe the Service Quality Multi-Agent System (SQ-MAS), a software developed using a multi-agent system platform to register failures in electric power distribution networks allowing the real-time calculation of several service quality indexes supervised by the Brazilian National Electric Agency (ANEEL). We introduce some basic concepts of generic MAS in order to illustrate the complexity of the problem related with the propagation of faults across the distribution network. The main features of SQ-MAS are listed and the calculation of the monitored indexes is addressed. A case study consisting of an Operation Unit of Petrobras, at Bahia State, Brazil, is used to illustrate how the system works.

1 January 2018 in International Journal of Applied Pattern Recognition
Verified
Journal Article
Closed access
SISusana Marrero Iglesias
Susana Marrero Iglesias
RMRodrigo Silva Mendonça
Rodrigo Silva Mendonça
EOEsbel Tomás Valero Orellana
Esbel Tomás Valero Orellana
DDDany Sanchez Dominguez
Dany Sanchez Dominguez
Civil construction is an alternative for the reuse of industrial discarded materials. The ethylene-vinyl acetate's residue as an aggregate, generates light materials with interesting thermal and acoustic properties. However, adding EVA reduces the material strength. To soften this effect, natural fibres like piassava can be added. This work proposes a methodology of three dimensional tomographic image analysis for the characterisation of lightweight mortars reinforced with piassava fibres. For this, several image features were calculated using different algorithms approaches focusing on the memory efficiency, the main handicap of three-dimensional image analysis. The methodology identified the EVA grains, the fibres and the pores, being insufficient collected information for the cracks identification. Also, it was possible to verify the fibre action as reinforcement. Regarding the algorithm performance analysis, the two-passage approaches were better, being the memory focused approach, the only one that can work with very big 3D images.

18 October 2018 in Matéria (Rio de Janeiro)
Verified
Journal Article
Open access
PDF file
MSMarlesson Rodrigues Oliveira de Santana
Marlesson Rodrigues Oliveira de Santana
DDDany Sanchez Dominguez
Dany Sanchez Dominguez
SISusana Marrero Iglesias
Susana Marrero Iglesias
JPJosé Renato de Castro Pessôa
José Renato de Castro Pessôa
LDLeonardo Alves Dias
Leonardo Alves Dias
A indústria da construção civil tem um grande potencial para a reciclagem de resíduos mediante a incorporação desses materiais como agregados. O resíduo de EVA (Etileno-Acetato de Vinila) vem sendo utilizado como uma alterativa na produção de compósitos cimentícios leves em conjunto com fibras vegetais de Attalea funifera Martius (piaçava), como reforço. Neste trabalho, procurou-se modelar as propriedades mecânicas de resistência à compressão, resistência à tração e módulo de deformação de argamassas contendo agregados não convencionais como EVA e fibras de piaçava. Foi utilizado um modelo de regressão polinomial com três variáveis explicativas para modelar cada propriedade mecânica, utilizando como variáveis independentes as porcentagens de EVA e fibras presentes na mistura, e a relação água/cimento do material. A qualidade de ajuste dos modelos foi avaliada por meio de testes hipótese, e por critérios estatísticos, dentre deles Shapiro-Wilk, Durbin-Watson, Breusch-Pagan, coeficientes de determinação R2 e critério PRESS (Prediction Error Sum of Squares). Os resultados indicam que os modelos construídos reproduzem satisfatoriamente o comportamento do material. Por meio de simulações realizadas com os modelos propostos, constatou-se que a fibra de piaçava pode ser utilizada como reforço para melhorar as propriedades de resistência à tração e módulo de deformação para a argamassa, dessa forma, foi possível identificar uma proporção ótima entre os agregados.

11 July 2008
Verified
Conference Paper
Closed access
PDF file
JAJoaquim Teixeira Assis
Joaquim Teixeira Assis
VMVladimir I. Monin
Vladimir I. Monin
SISusana Marrero Iglesias
Susana Marrero Iglesias
Study of stress gradients is one of the important problems of the X-ray tensometry, especially, in the case of analysis of residual stress state arising after different modern treatments like material modifications by ion beam technologies or surface treatments by laser. These problems related directly with incompleteness of theoretical and experimental measurement techniques with using of X-ray diffraction for study of stress gradients and they are connected with nonlinear character of dependency of diffraction angle θφ,ψ versus sin2ψ and with broadening of diffraction line caused by surface stress gradient. Computer simulation gives possibility to resolve the problem of determination of stress state characterized by strong gradient. The objective of presented paper is to develop a methodology of determination of stress gradient parameters. The methodology is based on computer simulation of diffraction line and permits to make corrections of stress measurements made by diffraction sin2ψ-method.

30 May 2019 in Revista Mundi Engenharia, Tecnologia e Gestão (ISSN: 2525-4782)
Verified
Journal Article
Open access
PDF file
DSDarcio R Da Silva
Darcio R Da Silva
DDDany Sanchez Dominguez
Dany Sanchez Dominguez
PAPaulo Ambrosio
Paulo Ambrosio
SISusana Marrero Iglesias
Susana Marrero Iglesias
Técnicas de modelagem computacional têm sido extensivamente aplicadas na análise e previsão de series temporais de dados socioeconômicos. Estas aplicações adquirem maior relevância no contexto de governança eletrônica, isto é, quando aplicadas na otimização de políticas públicas. Os métodos estatísticos e os modelos baseados em redes neurais oferecem resultados satisfatórios na previsão de series temporais. Entretanto, diversos autores destacam que o uso de modelos híbridos, que combinem ambas as abordagens, deve gerar previsões de maior qualidade. Neste trabalho, apresentamos um modelo híbrido que combina o método Autoregressivo Integrado e de Médias Móveis (ARIMA) com Redes Neurais Artificiais (RNA) na previsão de séries temporais do custo da cesta básica de alimentos na cidade de Ilhéus, região Sul da Bahia. A construção do método híbrido envolve selecionar o modelo ARIMA apropriado e a estimação de seus parâmetros através da analise das funções de correlação. Em relação a rede neural devemos selecionar a topologia da rede, e calcular os pesos sinápticos de cada neurônio mediante um algoritmo de treinamento. O modelo híbrido resultante é utilizado para gerar previsões em curto e médio prazo, avaliando os resultados através de medidas estatísticas de acurácia. Ademais, os resultados do modelo híbrido são comparados com os gerados pelos modelos de abordagem individual RNA e ARIMA, destacando as vantagens do uso de modelos híbridos.

4 September 2016 in TEMA (São Carlos)
Verified
Journal Article
Open access
PDF file
RRRogério Vinícius Matos Rocha
Rogério Vinícius Matos Rocha
DDDany Sanchez Dominguez
Dany Sanchez Dominguez
SISusana Marrero Iglesias
Susana Marrero Iglesias
RBRicardo Carvalho de Barros
Ricardo Carvalho de Barros
The physical phenomenon of neutrons transport associated with eigenvalue problems appears in the criticality calculations of nuclear reactors and can be treated as a diffusion process. This paper presents a new method to solve eigenvalue problems of neutron diffusion in slab geometry and one energy group. This formulation combines the Finite Element Method, considered an intermediate mesh method, with the Spectral Green's Function Method, which is free of truncation errors, and it is considered a coarse mesh method. The novelty of this formulation is to approach the spatial moments of the neutron flux distribution by the first-order polynomials obtained from the spectral analysis of diffusion equation. The approximations provided by the new formulation allow obtaining accurate results in coarse mesh calculations. To validate the method, we compare the results obtained with the methods described in the literature, specifically the Diamond Difference method. The accuracy and the computational performance of the proposed formulation were characterized by solving benchmarks problems with a high degree of heterogeneity.