Model-based proposal of a control strategy for friction stir welding

Abstract

Friction stir welding is an alternative to traditional welding processes, that can be considered both environmentally friendly and a tool to face weldability problems, specially in aluminium alloys. This process uses a particularly design tool that rotates and travels along the join, using equipment similar to a milling machine. Control in friction stir welding helps to assure weld soundness if done correctly. Different control approaches have been developed using force, torque and temperature sensors as input variables. Depending on the application this control strategies are successful while using dedicated FSW equipment, looking into local implementation it is more viable and cost effective to employ an adapted milling machine for the job. This type of mechanism only allows setting parameters such as travel speed, rotational speed and depth of penetration, which can be insufficient to obtain sound welds. Based on the set outcome of local implementation model-based control is an alternative to classic control and it uses a set of equations that describe the process as the input. The proposed model consists of a set of algebraic balance equations describing how mass and energy, in terms of power consumption, transform along the process predicting the soundness of the joint in every step. This model serves as the base for a control system that can predict the soundness in the next step and react accordingly, modifying basic process parameters such as plunge depth as well as travel and rotational speed minimizing mass loss. Experimental data has been compared to the model output in order to validate it.