• Journal of Welding and Joining
• /
• 제32권6호
• /
• pp.22-28
• /
• 2014

Most publications on friction stir welding describe phenomena or results with given process parameters like feed rate, rotation speed, angle and depth of penetration. But without a complete documentation of tool design, the results under the same process parameters are completely different. For this purpose, the Institute of Cutting and Joining Manufacturing Processes (tff), University of Kassel investigated the influence of tool roughness on the friction stir welding process. Therefore a defined surface finish was produced by turning and die sinking. As basis of comparison the constant parameters were rotation speed, feed rate, tilt angle and a heel plunge depth. Sound butt-welds were produced in aluminium alloy 6082 (AlMgSi1) with 1.5 mm sheet thickness with a turned reference tool with a surface of $Ra=0.575{\mu}m$ in position controlled mode. The surfaces are manufactured from a very fine to a very rough structure, classified by the VDI-classes with differences in the arithmetical mean roughness. It can be demonstrated with the help of temperature measures, that less heat is generated at the surfaces of the shoulder and the pin by the higher roughness due to lower active friction contact surface. This can also be seen in the resulting wormhole defects.

• 대한조선학회논문집
• /
• 제59권3호
• /
• pp.173-182
• /
• 2022

The present study suggests a procedure of establishing a ship dynamics modeling by regression of free-running model test results. The hydrodynamic force and moment of the whole model ship is derived from the low-pass filtered acceleration in the turning circle and zigzag maneuver tests. Force and moment of the propeller and rudder are separated from that of the whole ship to acquire the hull force and moment terms, based on the principles of the component model. The low-pass filter frequency is verified in prior to dynamics modeling, to find the threshold frequency of 2.5 Hz. The dynamics modeling of the hull is compared with the component modeling by captive model tests. Because of strong correlation between sway velocity, yaw angular velocity, and heel angle, each maneuvering coefficient is not able to be validated, but the whole modeling shows good agreement with the captive model tests.