Controle preditivo não linear baseado em técnica orientada a dados de decomposição de modo dinâmico polinomial

Abstract

The increasing complexity in business, industrial, and scientific processes, driven by current technological innovations, demands the ability to model operations, predict outcomes, categorize indicators, estimate gains, and control systems swiftly and efficiently, given the vast amount of data generated and collected daily. In this context, data-driven studies, work, and developments play a pivotal role in addressing contemporary technological challenges. Therefore, this work aims to propose advancements in data-driven techniques, focusing on the identification and control of complex nonlinear dynamic systems, whether they rational or not, based on a data-driven approach. Consequently, we suggest the exploration of the Dynamic Mode Decomposition with Polynomials method applied to Nonlinear Model Predictive Control. This theoretical evolution seeks practical application in a nonlinear and rational electric-hydraulic system, with the purpose of identifying polynomial prototypes that can be used as predictive models in liquid level control within the system.