Eine Plattform für die Wissenschaft: Bauingenieurwesen, Architektur und Urbanistik
Eine Plattform für die Wissenschaft: Bauingenieurwesen, Architektur und Urbanistik
Quality prediction of resistance spot welding joints of 304 austenitic stainless steel
AbstractThe quality level of a resistance spot welding (RSW) joint of 304 austenitic stainless steel (ASS) is estimated from its tensile shear load bearing capacity (TSLBC). The quality levels are set by ultrasonic nondestructive testing.The objective of the present work is to develop a tool capable of reliably predicting the TSLBC (and consequently the quality level) of RSW joints from three welding parameters: (1) welding time (WT); (2) welding current (WC); (3) electrode force (EF).Firstly, a linear regression model is attempted but the residuals analysis reveals nonlinear behaviour.An artificial neural network (ANN) is proposed because the ANNs are capable of mapping nonlinear systems. The inputs are 3-component vectors, a component for each of the aforementioned welding parameters. The training of the ANN uses supervised learning mechanism. Therefore each input must come with its respective desired output (target). This target is the TSLBC of the RSW joint obtained with the respective input. The number of neurons in the hidden layers is selected considering the overfitting phenomenon: the number of neurons in the hidden layers that minimizes the validation mean square error (MSE) is 4.With the selected ANN, 3–4–4–1, the aim of the present study is achieved because this ANN produces good results in prediction from inputs nonused in the training.
Quality prediction of resistance spot welding joints of 304 austenitic stainless steel
AbstractThe quality level of a resistance spot welding (RSW) joint of 304 austenitic stainless steel (ASS) is estimated from its tensile shear load bearing capacity (TSLBC). The quality levels are set by ultrasonic nondestructive testing.The objective of the present work is to develop a tool capable of reliably predicting the TSLBC (and consequently the quality level) of RSW joints from three welding parameters: (1) welding time (WT); (2) welding current (WC); (3) electrode force (EF).Firstly, a linear regression model is attempted but the residuals analysis reveals nonlinear behaviour.An artificial neural network (ANN) is proposed because the ANNs are capable of mapping nonlinear systems. The inputs are 3-component vectors, a component for each of the aforementioned welding parameters. The training of the ANN uses supervised learning mechanism. Therefore each input must come with its respective desired output (target). This target is the TSLBC of the RSW joint obtained with the respective input. The number of neurons in the hidden layers is selected considering the overfitting phenomenon: the number of neurons in the hidden layers that minimizes the validation mean square error (MSE) is 4.With the selected ANN, 3–4–4–1, the aim of the present study is achieved because this ANN produces good results in prediction from inputs nonused in the training.
Quality prediction of resistance spot welding joints of 304 austenitic stainless steel
Martín, Óscar (Autor:in) / Tiedra, Pilar De (Autor:in) / López, Manuel (Autor:in) / San-Juan, Manuel (Autor:in) / García, Cristina (Autor:in) / Martín, Fernando (Autor:in) / Blanco, Yolanda (Autor:in)
18.04.2008
10 pages
Aufsatz (Zeitschrift)
Elektronische Ressource
Englisch
Quality prediction of resistance spot welding joints of 304 austenitic stainless steel
| British Library Online Contents | 2009
Micro-Resistance Spot Welding of Nickel Free Austenitic Stainless Steel
| British Library Online Contents | 2007
| British Library Online Contents | 2010
Small-scale resistance spot welding of austenitic stainless steels
| British Library Online Contents | 2008
| British Library Online Contents | 2014