Numerical and Parametric study of FSW process using a pneumaticsource

Abstract

This thesis explores the optimization of friction stir welding (FSW) processes, with a primary emphasis on harnessing pneumatic sources to introduce portability, minimize power consumption, enable workability in hazardous environments, and enhance efficiency. Structured into four chapters, the study is dedicated to identifying the pneumatic motor (or rotary hand tool) best suited for the pneumatic FSW process, leveraging the power of machine learning and metaheuristic algorithms.

The initial chapter provides a comprehensive exploration of FSW fundamentals and emphasizes the pivotal role of pneumatic sources in this welding technique. The discussion underscores the potential advantages offered by pneumatic sources, particularly in terms of portability and energy efficiency.

In the subsequent chapter, the methodology of the study is delineated, encompassing various artificial intelligence techniques such as artificial neural networks, random forest algorithms, and polynomial regression. Complementing these methods are the SHAPLEY additive explanation and Pelican optimization algorithm, tailored specifically to analyze and optimize FSW parameters with pneumatic sources.

The thesis then progresses to the development and validation of predictive models for FSW parameters in the third chapter. Through rigorous validation procedures, the performance of these models is assessed to reach an R² of 99.60% and a MAPE of 4.54%, providing insights into the optimization of parameters.

In the final chapter, a hybrid approach is proposed, synthesizing the developed models with the Pelican optimization algorithm. This novel approach aims to identify an optimum set of parameters, including tool geometry, welding speed, rotational speed, tilt angle, torque, and minimal power consumption, specifically tailored for pneumatic-driven FSW processes. Consequently, this facilitates the selection of suitable pneumatic motors or rotary hand tools, alongside appropriate tool geometries, to further streamline and enhance FSW operations.