Publicação

Artificial intelligence in single screw polymer extrusion: Learning from computational data

Ver documento

Detalhes bibliográficos
Resumo:Single screw polymer extrusion can be seen as a multi-objective optimization problem where a set of design variables must be defined as a function of objectives and constraints that are to be satisfied simultaneously. The development of powerful modelling routines based on the use of numerical methods allows linking those objectives with the decision variables. In reality, only a single solution can be used in the problem under consideration. However, the computation times become prohibitive when effective optimization algorithms dealing with multi-objectives and decision-making are to be used, such as those based on populations of solutions. It is proposed here the use of Artificial Intelligence techniques to determine the interrelation between the design variables and the objectives. For that, a data analysis technique, named DAMICORE, was used to define these interrelations. Examples, involving the design of a screw extruder, a barrel grooves section, and a rotational barrel segment, were investigated using the proposed AI techniques. The results obtained show a good correspondence with the expected thermomechanical behaviour of the process. This constitutes an initial step in the application of AI techniques in different fields of engineering in the way of accomplishing, in the future, optimization based on the use of available data.
Autores principais:Gaspar-Cunha, A.
Outros Autores:Monaco, Francisco; Sikora, Janusz; Delbem, Alexandre
Assunto:Polymer processing Optimization Artificial intelligence Polymer extrusion Single screw Multi-objective optimization Data-mining
Ano:2022
País:Portugal
Tipo de documento:artigo
Tipo de acesso:acesso aberto
Instituição associada:Universidade do Minho
Idioma:inglês
Origem:RepositóriUM - Universidade do Minho
Descrição
Resumo:Single screw polymer extrusion can be seen as a multi-objective optimization problem where a set of design variables must be defined as a function of objectives and constraints that are to be satisfied simultaneously. The development of powerful modelling routines based on the use of numerical methods allows linking those objectives with the decision variables. In reality, only a single solution can be used in the problem under consideration. However, the computation times become prohibitive when effective optimization algorithms dealing with multi-objectives and decision-making are to be used, such as those based on populations of solutions. It is proposed here the use of Artificial Intelligence techniques to determine the interrelation between the design variables and the objectives. For that, a data analysis technique, named DAMICORE, was used to define these interrelations. Examples, involving the design of a screw extruder, a barrel grooves section, and a rotational barrel segment, were investigated using the proposed AI techniques. The results obtained show a good correspondence with the expected thermomechanical behaviour of the process. This constitutes an initial step in the application of AI techniques in different fields of engineering in the way of accomplishing, in the future, optimization based on the use of available data.