Physical-driven or simulation-driven experiments and data mining algorithms are combined for the identification of input–output relationships in complex manufacturing processes. To overcome time and cost consuming procedures, mathematical models have been applied finding and evaluating the factors that mostly affect the examined responses. In this context, a novel metamodeling technique was developed. This is able to use historical information on similar problems minimizing the amount of data necessary to the design of reengineered processes. The procedure was validated by applying it to the porthole extrusion optimizing the die geometries used for processing profiles characterized by various cross sections.
Metamodeling technique for designing reengineered processes by historical data
Gagliardi, Francesco;Ambrogio, Giuseppina;Ciancio, Claudio;Filice, Luigino
2017-01-01
Abstract
Physical-driven or simulation-driven experiments and data mining algorithms are combined for the identification of input–output relationships in complex manufacturing processes. To overcome time and cost consuming procedures, mathematical models have been applied finding and evaluating the factors that mostly affect the examined responses. In this context, a novel metamodeling technique was developed. This is able to use historical information on similar problems minimizing the amount of data necessary to the design of reengineered processes. The procedure was validated by applying it to the porthole extrusion optimizing the die geometries used for processing profiles characterized by various cross sections.I documenti in IRIS sono protetti da copyright e tutti i diritti sono riservati, salvo diversa indicazione.