William R. Longhurst

Heated Friction Stir Welding: An Experimental and Theoretical Investigation into How Preheating Influences Process Forces

As friction stir welding (FSW) has expanded to welding higher strength materials, large process forces and extreme tool wear have become issues. One possible solution is introducing an additional heating source in front of the FSW tool which softens the material and reduces the tool loads. We investigate the advantages of elevating temperature. Bead on plate welds were performed with a Trivex tool in aluminum alloy (AA 6061) heated to initial material temperatures up to 300°C. Macrograph cross-sections of the welds revealed a slight increase in material flow with increasing temperatures. More significant, the welding forces were analyzed to reveal up to a 43% reduction in the axial force with even moderate heating. An intriguing trend is observed that the process forces do not decrease steadily with increasing initial temperature, as might be expected, but exhibit a more complex polynomial shape, which actually increases for some heating intervals.

Investigation of force-controlled friction stir welding for manufacturing and automation

Abstract Friction stir welding (FSW) is a solid-state joining process for materials with low melting points. The process uses a rotating tool that consists of a shoulder and a pin. The tool plastically deforms the material with its pin and then forges together the parent materials underneath the shoulder. Past research has established that the axial force on the tool that creates the forging pressure is a function of plunge depth, traverse speed, and rotation speed. Historically, force control of FSW has been accomplished by varying the plunge depth of the tool. The research presented in this paper examines the force control of FSW by varying each of the process parameters separately. A force controller was implemented on a retrofitted milling machine. The closed-loop proportional—integral—derivative (PID) control architecture was tuned using the Ziegler—Nichols method. Welding experiments were conducted by butt welding ¼ inch (6.35 mm) × 1½ inch (38.1 mm) × 8 inch (203.2 mm) samples of aluminium 6061-T6511 with a ¼ inch (6.35 mm) FSW tool. The results indicate that force control via traverse speed is the most accurate and, as a by-product, heat distribution control along the weld seam occurs. Force control via plunge depth is the least accurate but it compensates for machine and robot deflection. Tensile test data show that greater strength can be obtained through force control via rotation speed. It is concluded that force is maintained by keeping the amount of tool surface area in contract with the workpiece constant throughout the welding process when plunge depth is used as the controlling variable. Force is maintained by varying the rate of heat generation when rotation speed is used as the controlling variable. Lastly, force is maintained by changing the amount of heat deposited per unit length along the weld seam when traverse speed is used as the controlling variable. Successful robotic FSW requires to be selected the appropriate controlling variable and the sensitivity of the interaction between the tool and the workpiece to be reduced.

The Identification of the Key Enablers for Force Control of Robotic Friction Stir Welding

Friction stir welding (FSW) is a solid state welding process that uses a rotating tool to plastically deform and then forge together materials. This process requires a large axial force to be maintained on the tool as the tool is plunged into the work piece and traversed along the weld seam. Force control is required if robots are to be used. Force control provides compensation for the compliant nature of robots. Without force control, welding flaws would continuously emerge as the robot repositioned its linkages to traverse the tool along the weld seam. Insufficient plunge depth would result and cause the welding flaws as the robots linkages yieldedfrom the resulting force in welding environment. As FSW continues to emerge in manufacturing, robotic applications will be desired to establish flexible automation. The research presented here identifies the key enablers for successful and stable force control of FSW. To this end, a FSW force controller was designed and implemented on a retrofitted Milwaukee Model K milling machine. The closed loop proportional, integral plus derivative control architecture was tuned using the Ziegler-Nichols method. Welding experiments were conducted by butt welding 0.25 in. (6.35 mm) x 1.50 in. (38.1 mm) x 8.0 in. (203.2 mm) samples of aluminum 6061 with a 0.25 in. (6.35 mm) threaded tool. The experimental force control system was able to regulate to a desired force with a standard deviation of 129.4 N. From the experiments, it was determined that tool geometry and position are important parameters influencing the performance of the force controller, and four key enablers were identified for stable force control of FSW. The most important enabler is the maintaining of the position of a portion of the tools shoulder above the work piece surface. When the shoulder is completely submerged below the surface, an unstable system occurs. The other key enablers are a smooth motion profile, an increased lead angle, and positional constraints for the tool. These last three enablers contribute to the stability of the system by making the tools interaction with the nonlinear welding environment less sensitive. It is concluded that successful implementation of force control in the robotic FSW systems can be obtained by establishing and adhering to these key enablers. In addition, force control via plunge depth adjustment reduces weld flash and improves the appearance of the weld.

Adaptive torque control of friction stir welding for the purpose of estimating tool wear

In this paper, an adaptive torque controller for friction stir welding (FSW) that can estimate parameters such as probe radius which may be changing throughout the welding process is presented. Implementing an adaptive controller with this capability would be of interest to industry sectors in which FSW is performed on high melting point alloys or metal matrix composites (MMC). Welding these materials has shown a greatly accelerated rate of tool wear. Simulations were conducted to examine how extreme tool wear would affect controller performance and how accurately the controller could estimate the probe radius. A simplified wear model consisting of a linear decrease in probe radius was used to verify controller performance. Next, a wear model consistent with wear patterns seen in the welding of highly abrasive materials was developed. Results indicate that torque is controlled effectively while a change in system dynamics is experienced, as would be expected with adaptive control, but also that the tool profile is accurately estimated after an initial identification period.

Enabling Automation of Friction Stir Welding : The Modulation of Weld Seam Input Energy by Traverse Speed Force Control

Friction stir welding (FSW) joins materials by plunging a rotating tool into the work piece. The tool consists of a shoulder and a pin that plastically deforms the parent materials and then forges them together under the applied pressure. To create the pressure needed for forging, a rather large axial force must be maintained on the tool. Maintaining this axial force is challenging for robots due to their limited load capacity and compliant nature. To address this problem, force control has been used, and historically, the force has been controlled by adjusting the plunge depth of the tool into the work piece. This paper develops the use of tool traverse speed as the controlling variable instead of plunge depth. To perform this investigation, a FSW force controller was designed and implemented on a retrofitted Milwaukee Model K milling machine. The closed loop proportional, integral plus derivative (PID) control architecture was tuned using the Ziegler‐Nichols method. Results show that the control of axial force via traverse speed is feasible and predictable. The resulting system is more robust and stable when compared with a force controller that uses plunge depth as the controlling variable. A standard deviation of 41.5 N was obtained. This variation is much less when compared with a standard deviation of 129.4 N obtained when using plunge depth. Using various combinations of PID control, the system’s response to step inputs was analyzed. From this analysis, a feed forward transfer function was modeled that describes the machinery and welding environment. From these results, a technique is presented regarding weld seam input energy modulation as a by product of force control via traverse speed. A relative indication of thermal energy in the welding environment is obtained with the feedback of axial force. It is hypothesized that, while under force control, the controller modulates weld seam input energy according to the control signal. The result is constant thermomechanical conditions in the welding environment. It is concluded that the key enablers for force control are the unidirectional behavior and load dynamics of the traverse motor. Larger bandwidths and more stable weld conditions emerge when using traverse speed instead of plunge depth to control the force. Force control of FSW via traverse speed has importance in creating efficient automatic manufacturing operations. The intelligence of the controller naturally selects the most efficient traverse speed. DOI: 10.1115/1.4001795

Evaluation of torque as a means of in-process sensing of tool wear in friction stir welding of metal matrix composites

Purpose The purpose of this paper is to evaluate the tool experiences using torque during welding as a means of in-process sensing for tool wear. Metal matrix composites (MMCs) are materials with immense potential for aerospace structural applications. The major barrier to implementation of these materials is manufacturability, specifically joining MMCs to themselves or other materials using fusion welding. Friction stir welding (FSW) is an excellent candidate process for joining MMCs, as it occurs below the melting point of the material, thus precluding the formation of degradative intermetallics’ phases present in fusion welded joints. The limiting factor for use of FSW in this application is wear of the tool. The abrasive particles which give MMCs their enhanced properties progressively erode the tool features that facilitate vertical mixing and consolidation of material during welding, resulting in joints with porosity. While wear can be mitigated by careful selection of process parameters and/or the use of harder tool materials, these approaches have significant complexities and limitations. Design/methodology/approach This study evaluates using the torque the tool experiences during welding as a means of in-process sensing for tool wear. Process signals were collected during linear FSW of Al 359/SiC/20p and correlated with wear of the tool probe. The results of these experiments demonstrate that there is a correlation between torque and wear, and the torque process signal can potentially be exploited to monitor and control tool wear during welding. Findings Radial deterioration of the probe during joining of MMCs by FSW corresponds to a decrease in the torque experienced by the tool. Experimentally observed relationship between torque and wear opens the door to the development of in-process sensing, as the decay in the torque signal can be correlated to the amount of volume lost by the probe. The decay function for tool wear in FSW of a particular MMC can be determined experimentally using the methodology presented here. The decay of the torque signal as the tool loses volume presents a potential method for control of the wear process. Originality/value This work has near-term commercial applications, as a means of monitoring and controlling wear in process could serve to grow commercial use of MMCs and expand the design space for these materials beyond net or near-net-shape parts.

Applied torque control of friction stir welding using motor current as feedback

When controlling friction stir welding, effort must be given to maintaining proper tool shoulder contact with the workpiece in order to achieve consolidation of the parent materials. Axial force control has been used prior with some success. The research presented in this article examines the controlling of welding torque as an alternative to force control. A mathematical model of welding torque was enhanced for the design and study of convex shoulder profiles. Focus was placed on linearizing the response between plunge depth and torque. The model predicted that a spherical profiled shoulder was preferred for a more linear response. In conjunction with the spherical shoulder, a closed-loop torque controller was implemented and its performance evaluated. Welding torque for feedback control was sensed indirectly through the spindle motor current using a commercially available clamp-on current meter. The system produced 1/4 in (6 mm) bead-on-plate welds that were 10 ft (3 m) in length. Over the course of the welds, the torque controller responded to workpiece elevation changes of 1/8 in (3 mm) and 1/4 in (6 mm). Results show that the tool maintained a near constant plunge depth into the workpiece as the tool traversed along the workpiece. It was concluded that the presented method of torque control is a reliable and less complex alternative to axial force control of friction stir welding.

Process monitoring of friction stir welding via the frequency of the spindle motor current

Friction stir welding is a solid-state process that is gaining preference for the joining of metals with low melting points. Despite the clear advantages of friction stir welding over traditional fusion welding, voids within the weld seam arise when improper conditions are present. The work presented in this article examines the development of an automated process monitoring system for friction stir welding. The system indirectly monitors the welding torque through the supplied current to the spindle motor. To measure the current, a clamp-on current meter was used. Our results have shown that using a simple and inexpensive clamp-on current meter provides good insight into the welding torque. Examination focused on the frequency spectrum of the current. A Fourier transform decomposed the signal into various frequencies present. The results consistently showed that when no void was present, there was a component of the current’s frequency at 14 Hz. However, when the tool encountered a void, the frequency spectrum changed. The component at 14 Hz went away while content in the range of 1–4 Hz increased.