This study analyses the influence of the key process parameters during friction assisted joining process on the frictional power. Experimental joining tests were performed by varying the plunging force, the tool rotation speed, and the tool diameter. During the tests, the main processing loads were acquired through an instrumented equipment. The physical correlation between the process parameters and the frictional power was determined and modeled through machine learning. Different model configurations were tested varying the number of neurons in the hidden layer. The results proved the possibility to predict the frictional power through machine learning with good reliability (R2 = 0.90) and generalization capability. This was pursued using a limited number of experimental tests through a signal breakdown approach.
This study analyses the influence of the key process parameters during friction assisted joining process on the frictional power. Experimental joining tests were performed by varying the plunging force, the tool rotation speed, and the tool diameter. During the tests, the main processing loads were acquired through an instrumented equipment. The physical correlation between the process parameters and the frictional power was determined and modeled through machine learning. Different model configurations were tested varying the number of neurons in the hidden layer. The results proved the possibility to predict the frictional power through machine learning with good reliability (R2 = 0.90) and generalization capability. This was pursued using a limited number of experimental tests through a signal breakdown approach.
Prediction of the power supplied in friction-based joining process of metal-polymer hybrids through machine learning
Lambiase F.
;Scipioni S. I.;Paoletti A.
2021-01-01
Abstract
This study analyses the influence of the key process parameters during friction assisted joining process on the frictional power. Experimental joining tests were performed by varying the plunging force, the tool rotation speed, and the tool diameter. During the tests, the main processing loads were acquired through an instrumented equipment. The physical correlation between the process parameters and the frictional power was determined and modeled through machine learning. Different model configurations were tested varying the number of neurons in the hidden layer. The results proved the possibility to predict the frictional power through machine learning with good reliability (R2 = 0.90) and generalization capability. This was pursued using a limited number of experimental tests through a signal breakdown approach.File | Dimensione | Formato | |
---|---|---|---|
prediction of the power supplied.pdf
Open Access dal 24/06/2023
Tipologia:
Documento in Versione Editoriale
Licenza:
Dominio pubblico
Dimensione
6.11 MB
Formato
Adobe PDF
|
6.11 MB | Adobe PDF | Visualizza/Apri |
Pubblicazioni consigliate
I documenti in IRIS sono protetti da copyright e tutti i diritti sono riservati, salvo diversa indicazione.