Novas estratégias na modelagem multivariada de parâmetros de qualidade num processo químico industrial

• Autor(a) principal: Freitas, Leandro Valim de [UNESP]
• Data de Publicação: 2017
• Tipo de documento: Tese
• Idioma: por
• Título da fonte: Repositório Institucional da UNESP
• Texto Completo: http://hdl.handle.net/11449/151939
• topic: Multivariate analysis; Process monitoring; Taguchi; Análise multivariada; Monitoramento de processos

This work aims to develop and validate multivariate predictive mathematical models for the purpose of monitoring in a chemical manufacturing process. Derivatives have specifications based on physicochemical properties that can vary significantly with their modification of the cast of raw materials, while maintaining the same conditions of production control, which compromises the quality standards of their final products . This leads to the need to determine or predict them as often as possible over traditional laboratory analyzes. It was possible to mathematically model properties using the multivariate Vector Supporting Machine (SVM) technique from experimental data and compare it with the traditional multivariate approach by Partial Least Squares (PLS). As the PLS algorithm, widely used in the treatment of data of this nature does not have the regularization parameter as the machines, the PLS-Taguchi algorithm was proposed in an attempt to aid its configuration and optimize its parameters to overcome this absence.