APLICAÇÃO DO MÉTODO BAYESIANO FILTRO DE PARTÍCULAS PARA ESTIMATIVA DE CURVAS DE RUPTURA DE ADSORÇÃO EM LEITO FIXO
APLICAÇÃO DO MÉTODO BAYESIANO FILTRO DE PARTÍCULAS PARA ESTIMATIVA DE CURVAS DE RUPTURA DE ADSORÇÃO EM LEITO FIXO
-
DOI: https://doi.org/10.22533/at.ed.1351625170310
-
Palavras-chave: Linear Driving Force; Sampling Importance Resamplig; Inferência Bayesiana
-
Keywords: Linear Driving Force; Sampling Importance Resampling; Bayesian Inference
-
Abstract: This study applied a Bayesian Sampling Importance Resampling (SIR) particle filter to model breakthrough curves in a fixed-bed for diclofenac, marbofloxacin, paracetamol, and CO₂, integrating the Linear Driving Force (LDF) kinetics and the Langmuir isotherm, with experimental validation based on literature data. The methodology combines discretization via the Method of Lines and parameter estimation through an objective function, revealing distinct behaviors among the adsorbates. CO₂ exhibited a balance between convection and diffusion (Pe ≈ 15) and a high adsorption capacity (Qₘₐₓ ≈ 40 mg/g), paracetamol showed dominance of the convective transport regime (Pe ≈ 82), while marbofloxacin demonstrated diffusive control (Pe ≈ 1). The results showed that a 5% model uncertainty optimized the balance between experimental adherence (RMSE < 0.041) and physical consistency, with residual analysis confirming the absence of systematic bias. The approach provides quantitative criteria for the design of contaminant removal systems.
- Miguel Fernando Saraiva Maia
- Matheus Bastos do Carmo
- Waldecléia Queiroz da Costa
- João Victor Tovany Soares da Silva
- Haianny Beatriz Saraiva Lima
- Ana Paula Souza de Sousa
- Diego Cardoso Estumano