ON-LINE WELD QUALITY DIAGNOSTICS IN FRICTION STIR WELDING D. Jandric, J.H Ouyang, M. Valant and R. Kovacevic ABSTRACT The temperature distribution on the plate surface in front of the shoulder and around the pin very close to the workpiece surface is analyzed in this paper. Consideration is given to the influence of the welding parameters, the traverse speed and the rotational speed on the temperature distribution and their effect on weld quality. Infrared sensing is applied in order to collect relevant data for this analysis. KEYWORDS Infrared Sensing, Surface Temperature Distribution, Welding Parameters and Weld Quality INTRODUCTION Friction Stir Welding (FSW), invented and patented by The Welding Institute, Cambridge, England, in 1991, has recently captured the attention of the welding community for fabricating high-quality joints of aluminum alloys. This is an energy efficient, environment friendly and versatile process. Friction Stir Welding is a very attractive technology because of its simplicity and applicability to a wide range of alloys. It offers a variety of new design and production welding opportunities and should be of particular interest to the shipbuilding, aerospace, and automotive industries. The basic principle of FSW is relatively simple. A rotating pin is inserted in the material to be joined and traversed through the joint line while the shoulder is in contact with the top surface of Southern Methodist university, 1500 International Pkwy. Suite # 100, Richardson, TX 75081 temperature around the pin was measured by thermocouples embedded into the workpiece in order to get the temperature history during the welding process. Simultaneous sensing is performed by an infrared camera and by force/torque sensors. Sensing the surface temperature in front of the tool using an infrared camera will provide conditions for controlling the welding parameters during the friction stir welding process. Very little has been done in FSW for process monitoring with an infrared camera, but based on research results from other welding techniques, infrared sensing can be successfully applied to FSW. This process monitoring technique is used in the analysis of the surface temperature distribution during the laser beam welding process, (Ref. 2). In Gas Tungsten Arc Welding, infrared process monitoring represents the basis for weld quality control during the welding. Variations in the welding process parameters such as bead width, penetration depth and torch position produce an infrared radiation response, (Ref. 3). In general, the infrared thermal imaging system is a valuable tool that may be used to pinpoint problem areas and to avoid costly delays. Monitoring the pin tool depth during the FSW process is one of the most difficult tasks, but it is possible using the force/torque sensors, (Ref. 4). The relationship between the axial load and the pin tool depth during the welding was observed to be constant for all welds made with the same tool pin geometry and the pin tool depth. By controlling the axial load, the pin tool depth is also controlled. NASA has introduced (Ref. 5) a new tool known as Retractable Pin-Tool. This tool will assist in developing the closed-looped force control system capable for performing defect free welding. The microstructure of the joint made by the FSW technique has been the subject of several investigations (Ref. 6). The extrusion and forging operation, coupled with the tool rotation, are collectively responsible for the characteristic microstructure of friction stir welded joints. Microstructural analyses gave information about the weld quality and also provided proof that FSW is a solid state process. The present paper focuses on the effects of the welding parameters, rotational speed and traverse speed on weld quality. It is shown that infrared sensing is feasible technique for nondestructive monitoring of the FSW process. records the infrared radiation in front of the shoulder on the plate surface, Fig.2. Additional software is developed for storing and processing recorded data. Experimental procedure A number of FSW experiments are carried out to obtain the optimum parameters by adjusting the rotational and traverse speeds. Experiments are performed on 6061-T6 aluminum plates, 280x200x12.7 mm in size. Temperatures during the Friction Stir Welding process were measured by K type NiAI/NiCr thermocouples placed in a series of small holes equidistant from the specimen's surface. It was found that the existence of these holes does not significantly affect the temperature field during the welding process, (Ref. 7). Peak temperatures were measured for five different rotational tool speeds (151 rpm, 344 rpm, 416 rpm, 637 rpm and 914 rpm) and five different welding speeds (linear workpiece speeds, 95 mm/min, 140 mm/min, 190 mm/min, 229 mm/min and 330 mm/min). The temperatures were recorded digitally using the embedded thermocouples and the data acquisition system. In the effort to understand the process parameter/joint quality relationship, surface temperatures near the edge of shoulder, at the centerline of the nugget zone, are recorded by an infrared camera. The recorded infrared data is processed to obtain the temperature distribution on the top side of the workpiece in front of the tool shoulder. A frame grabber captures images, and image processing is performed using LabView and Matlab programming. The stored images are digitized into 640x480 matrices which are composed of gray level values from the infrared image. The cross-sections of the weld beads are etched by a modified Keller's etchant. Microstructural analyses are performed using a high-resolution microscope. As a solid state process, FSW operates below the with the time when thermocouples registered the temperature peak, the thermocouples' positions are found on the isothermal images. A typical result is shown in Fig.5. The measured temperatures are not the real surface temperatures but they can be treated as such because the thermocouples are placed very close to the plate surface. The temperature of 410 °C, which is about 75% of the peak temperature, was measured 25mm from the center of the pin. This corresponds to a gray level of 220. The surface temperature in front of the shoulder, being 70-80% of the peak temperature, as measured by thermocouples very close to the nugget, could be a good indicator of the process, measured by an infrared camera placed in front of the tool. Optimization of the welding parameters relative to the gray level intensity in the thermal images gives a basis for on-line control of the FSW process. Figure 5. Temperature distribution in front of the tool shoulder 720 Influence of welding speed on temperature distribution Welding speed, one of the most important parameters in the FSW process, affects temperature distribution on the plate surface in front of the tool shoulder. The temperature as a function of the distance from the shoulder is shown in Fig. 6. This temperature is obtained by processing the original images recorded by an infrared camera. It follows that the temperature in front of the shoulder will decrease if the welding speed increases for a constant value of heat input. In a real welding situation this eventually leads to pin fracture and the formation of weld defects as the Temperature in C 910 95 mm/min material becomes less formable, (Ref. 10). On the other hand, the temperature on the shoulder edge is measured by thermocouples as well. It can be shown that for welding speed of 95 mm/min temperature around the shoulder is 405 °C and for welding speed of 190 mm/min that temperature is 382 °C. When comparing the temperature results from thermocouples with results obtained from thermal images on the shoulder edge and by extrapolating the curve, see Fig. 6, small difference in temperature on the shoulder edge can be expected Figure 6. Temperature in function of because of the 0.5 mm depth from the top surface where 190 mm/min 18 distance from the shoulder the thermocouples are placed. Influence of rotational speed on the temperature distribution |