Ceramics Predictor System (CPS): sistema de análise e previsão de propriedades físicas de materiais cerâmicos

Abstract

Ceramic materials are popularly known for being used in glass, crockery, refractories and cement. In addition to the most common applications, there is a range of ceramic materials given in more technological solutions, such as thermal, electro-electronic, optical, alloys and bio-medical. The study and analysis regarding this class of materials are of paramount importance for technological advances. In this regard, several complementary researches are being carried out in order to find new materials with different properties, opening up innovative possibilities for applications in different segments. In this work we explore an existing database and focus on two properties (Dielectric Constant and Quality Factor), electromagnetic layers. For this purpose, first we present an interactive tool for analysis and exploration of ceramics, showing information on the structure and their physical properties. Then, applying machine learning algorithms, we developed a model to predict the physical properties of interest for new materials. The results show good predictability of the Dielectric Constant and Quality Factor of the compound, the accuracy of predictive models is 86% and 75% respectively. All modules developed here were integrated into an interactive web tool, allowing users to analyze and explore in a simple way the proposition of new ceramic materials. We believe that this work is a fundamental step to accelerate the efficient discovery of new ceramic materials, focused on specific applications.