Self-reacting friction stir welding (SRFSW) is an advanced variant of
friction stir welding (FSW) and shows several superiorities with the double-sided tool configuration. Despite the considerable amount of experimental studies in this field, most of the process development efforts, including tool design and process parameter selections, are still empirical and resort to trial-and-error solutions. To address this issue, high fidelity process models are desired that can unravel the complex and cross-linked interactions of material flow, heat generation and temperature distribution as well as metallurgical evolutions during SRFSW. In this research, an integrated multi-physics SRFSW process model was developed using computational fluid dynamics (CFD) method, and the formation of potential weld defects were predicted based on the tool design and process parameters.