摘要 |
<p>The invention relates to a method which includes at least the steps of studying values of manufacturing process variables (10a, 10b, 10c), comparing said values with previously defined quality values, obtaining a set of models (50, 51, ... 5n) using techniques based on artificial intelligence, thereby obtaining a prediction of real-time quality value estimates from the values of the manufacturing process variables (10a, 10b, 10c) at every instant, and identifying lengths of the sections that fulfil or fall short of established quality criteria. The values in the study phase can include variables (20) associated with the actual elastomers. An additional step (180) of marking lengths can be performed in order for an operator to know whether said lengths are valid or should be rejected when cut. An additional step of registering (170) the type of defect detected and the intensity thereof can also be performed.</p> |
申请人 |
UNIVERSIDAD DE LA RIOJA;UNIVERSIDAD POLITECNICA DE MADRID;ALBA ELIAS, FERNANDO;BELLO GARCIA, ANTONIO;CASTEJON LIMAS, MANUEL;GONZALEZ MARCOS, ANA;MARTINEZ DE PISON ASCACIBAR, FRANCISCO JAVIER;PERNIA ESPINOZA, ALPHA VERONICA;ORDIERES MERE, JOAQUIN;COBO BENITA, JOSE RAMON |
发明人 |
ALBA ELIAS, FERNANDO;BELLO GARCIA, ANTONIO;CASTEJON LIMAS, MANUEL;GONZALEZ MARCOS, ANA;MARTINEZ DE PISON ASCACIBAR, FRANCISCO JAVIER;PERNIA ESPINOZA, ALPHA VERONICA;ORDIERES MERE, JOAQUIN;COBO BENITA, JOSE RAMON |