Uncertainty quantification and global sensitivity analysis of continuous distillation considering the interaction of parameter uncertainty with feed variability
[EN] In this work, uncertainty and sensitivity analyses were applied to study the joint effects of model parameter uncertainty and feed variability on the response of a computational code for methanol-water continuous distillation. First, model parameter uncertainty (liquid-vapour equilibrium (VLE),...
| Autores: | , , |
|---|---|
| Tipo de recurso: | artículo |
| Fecha de publicación: | 2021 |
| País: | España |
| Institución: | Universitat Politècnica de València (UPV) |
| Repositorio: | RiuNet. Repositorio Institucional de la Universitat Politécnica de Valéncia |
| Idioma: | inglés |
| OAI Identifier: | oai:riunet.upv.es:10251/183160 |
| Acceso en línea: | https://riunet.upv.es/handle/10251/183160 |
| Access Level: | acceso abierto |
| Palabra clave: | Distillation Uncertainty Sensitivity Morris analysis Sobol method ESTADISTICA E INVESTIGACION OPERATIVA INGENIERIA QUIMICA |
| Sumario: | [EN] In this work, uncertainty and sensitivity analyses were applied to study the joint effects of model parameter uncertainty and feed variability on the response of a computational code for methanol-water continuous distillation. First, model parameter uncertainty (liquid-vapour equilibrium (VLE), enthalpy and tray efficiency) was characterised using existing experimental data. Afterwards, three tower configurations working at two operational modes (fixed product composition and fixed operation conditions) were studied at three feed variability levels. Morris analysis revealed the high importance of the VLE and efficiency-related factors. Sobol sensitivity analysis determined with more precision the sensitivity of the response to the parameters and detected non-linear effects and interactions. The Monte Carlo propagation method allowed obtaining the uncertainty margins as a function of feed variability. The results showed high impact of the model parameter uncertainty and encourage the use of the methods shown to obtain robust designs and quantify simulation accuracy. |
|---|