Análise cinética da decomposição térmica de sólidos através do uso inovador de redes neurais

Abstract

Artificial neural networks (ANNs) emerge as robust tools to tackle complex challenges, simulating the learning, self-organization, and parallelism of the human brain. The developed algorithm, incorporating the Hopfield network and two other ANNs, stands out in the kinetic analysis of solids, covering polymorphic conversion, intermolecular interaction, and thermal decomposition. Applied to various solids such as pharmaceuticals and hybrid materials, the method shows a significant reduction in residual error, demonstrating precision and reliability in interpreting thermal data. The substantial computational time savings, exceeding 500 times in certain analyses, are crucial for scientific research, providing agile insights into the kinetics of complex thermal processes. The project, pre-accelerated in the Catalisa ICT program, stands out for the online platform based on thermal analyses, offering precise kinetic data. The strategic approach includes categorized risk analyses, providing an in-depth view of challenges and the development of mitigation strategies. Interaction with experts and entrepreneurs contributed to adjustments and refinements. The consolidation of experiences positions the project to advance, exemplifying the acceleration of innovative businesses in Catalisa ICT.