Modelagem matemática da cinética do processo de biodigestão anaeróbia: implementação, investigação e análise em Python

• Main Author: Ramires, Matheus Belchior
• Publication Date: 2023
• Other Authors: Carvalho, Yuri Claro Costa
• Format: Bachelor thesis
• Language: por
• Source: Repositório Institucional da UTFPR (da Universidade Tecnológica Federal do Paraná (RIUT))
• Download full: http://repositorio.utfpr.edu.br/jspui/handle/1/37100
• topic: Digestão anaeróbia; Modelos matemáticos; Otimização matemática; Simulação (Computadores); Anaerobic digestion; Mathematical models; Mathematical optimization; Computer simulation; CNPQ::ENGENHARIAS::ENGENHARIA QUIMICA

Due to the imminent climate crisis, the production of new sustainable energy sources is one of the main challenges of our time. Within this scope, anaerobic digestion (AD) is a technology extensively studied for its dual mitigating bias, on one hand serving as a method of effluent treatment and on the other, as a way of obtaining biogas, that is, bioenergy. In this context, considering that AD is a complex bioprocess sensitive to operational conditions, mathematical modeling presents itself as a useful tool to assist both in the analysis and description of the process, as well as in the estimation and monitoring of kinetic parameters. Thus, in the present work, it was proposed to develop the modeling of five models applied to anaerobic digestion, using the implementation in Python language. For this, a two and three-phase model with 1st order kinetics described in the literature was implemented and structured in three different ways in order to validate the computational implementation in the chosen language and investigate the influence of mathematical structuring and computational methods of integration and optimization in the model fit. Two single-phase models based on Monod kinetics were also proposed in order to verify the response and fit of these simplified models to experimental data, especially with regard to biomass growth. According to the results obtained, it was possible to demonstrate the effectiveness of the application of the Python language as a tool for the proposed modeling processes, and the response of the two and three-phase model with 1st order kinetics was consistent with what was reported in the literature, obtaining determination coefficients higher than 0.9. For the single-phase model, the ability to adequately describe the overall behavior of the system was observed, with determination coefficients ranging between 0.96 and 0.99. In addition, the interpretation of the calibration of the implemented models allowed inference about the limiting step of the process, functioning appropriately as a simplified empirical model. Finally, the evaluation of the models based on the values of the determination coefficients and Akaike Information Criteria calculated allowed to identify Structure 2 as the most statistically adequate model.