ABSTRACT The addition of nanomaterials to concrete allows structural modifications, improving their properties like mechanical strength and durability. Within these possible nanoscale materials is silica hydrosol (SH). There are many studies highlighting the influence of the surface area of SH particles on the pozzolanic reaction with cement, however little has been studied on the influence of percentage and particle size on the reaction. In this context, the objective of this work was to investigate the influence of the particle size distribution of silica nanoparticles on the pozzolanic reaction in cement pastes, from the comparison of 4 colloidal suspensions. The addition content of these particles, with different sizes, was defined as 3.00% (SH-A), 0.89% (SH-B), 0.77% (SH-C) and 0.32% (SH-D), to equalize the surface areas of all of them, so that the only variables were the percentage and particle size distribution of the silicas. The pozzolanic effect was evaluated by the direct method named Modified Chapelle test, and by indirect methods like Fourier transform infrared spectroscopy (FT IR), and mechanical strength tests. It was concluded that for the same surface area resulting from the addition of SH, the sample that presented the best performance was the CP SH-A, which presented the largest particle size and volume of addition. However, when analysing the relative performance, regarding the concentration of SH added to the paste, the CP SH-D sample obtained the highest mechanical resistance and the highest pozzolanic index. It showed that concentration and particle size, not just surface area, influenced the performance of the nanosilica pozzolanic reaction. The conclusions obtained in the present work allow a better understanding about the addition of SH in cementitious composites and how the content and particle size of such products impact their performance.

Resumo A Análise de Agrupamento inclui um número de diferentes algoritmos e métodos para agrupamento de objetos em categorias a partir de suas características similares. Recentemente, um esforço considerável tem sido realizado visando à melhoria de desempenho de tais algoritmos. Neste sentido, este artigo explora três diferentes meta-heurísticas bioinspiradas no problema de agrupamento, sendo elas: Algoritmos Genéticos (AG), Otimização por Colônia de Formigas (OCF) e Sistemas Imunológicos Artificiais (SAI). Este artigo propõe alguns refinamentos a serem aplicados a estas meta-heurísticas visando melhorar seu desempenho no problema de agrupamento de dados. O desempenho dos algoritmos propostos é comparado usando cinco diferentes bases de dados numéricas da UCI. Os resultados mostram que os algoritmos são capazes de formar grupos automática e eficientemente a partir de um número pré-definido de clusters.

Abstract Clustering analysis includes a number of different algorithms and methods for grouping objects by their similar characteristics into categories. In recent years, considerable effort has been made to improve such algorithms performance. In this sense, this paper explores three different bio-inspired metaheuristics in the clustering problem: Genetic Algorithms (GAs), Ant Colony Optimization (ACO), and Artificial Immune Systems (AIS). This paper proposes some refinements to be applied to these metaheuristics in order to improve their performance in the data clustering problem. The performance of the proposed algorithms is compared on five different numeric UCI databases. The results show that GA, ACO and AIS based algorithms are able to efficiently and automatically forming natural groups from a pre-defined number of clusters.