Predicting Characteristics of Dissimilar Laser Welded Polymeric Joints Using a Multi-Layer Perceptrons Model Coupled...

Area of Science:

• Materials Science
• Polymer Engineering
• Computational Modeling

Background:

• Laser transmission welding (LTW) is a key technique for joining polymers.
• Optimizing LTW parameters is crucial for achieving high-quality dissimilar joints.
• Predictive modeling can enhance process control and joint performance.

Purpose of the Study:

• To develop and evaluate a multi-layer perceptrons (MLP) model coupled with the Archimedes optimizer (AO) for predicting dissimilar polymer lap joint characteristics.
• To compare the predictive accuracy of the AO-MLP model against conventional gradient descent and particle swarm optimization (PSO) coupled MLP models.
• To investigate the influence of LTW parameters on seam width and shear strength in polymethyl methacrylate (PMMA) and polycarbonate (PC) joints.

Main Methods:

• Utilized a coupled multi-layer perceptrons (MLP) model integrated with the Archimedes optimizer (AO).
• Input parameters included laser power, welding speed, pulse frequency, wobble frequency, and wobble width.
• Outputs predicted were seam width and shear strength of dissimilar PMMA/PC lap joints welded via LTW with beam wobbling.
• Compared AO-MLP with MLP, PSO-MLP using root mean square error (RMSE) for performance evaluation.

Main Results:

• The AO-MLP model demonstrated superior prediction accuracy for both shear strength and seam width compared to MLP and PSO-MLP.
• Achieved significantly lower RMSE values: 2.283 for shear strength and 0.0321 for seam width with AO-MLP.
• MLP and PSO-MLP models yielded RMSEs of 39.798 and 19.909 for shear strength, and 0.153 and 0.084 for seam width, respectively.

Conclusions:

• The AO-MLP model offers a highly accurate and efficient approach for predicting the performance of dissimilar polymer lap joints produced by LTW.
• The Archimedes optimizer effectively enhances the parameter optimization of MLP models for complex material joining processes.
• This predictive capability can aid in optimizing LTW process parameters for improved joint quality and reliability.