Identificação de processos dinamicos não lineares : desenvolvimento de software e aplicação a uma coluna fracionadora de eteno

AUTOR(ES)
DATA DE PUBLICAÇÃO

2003

RESUMO

This paper describes the methodology used in the identification and development of petrochemical non-linear dynamic models. In the present study were used techniques of: data handling; process control; optimization and neural networks. A software tool was developed in order to automate of the procedures. On the other hand, the methodology was tested using an Ethylene Fractionator model identification with heavy non-linearity and pronounced response times. The presented methodology allowed successfully studies and the elaboration of non-linear dynamic models using an extensive data base of the industrial process. The models can be applied in diverse fields as: multivariable predictive control; virtual analyzers and control strategies studies

ASSUNTO(S)

etileno industria petroquimica redes neurais (computação)

Documentos Relacionados