Fig. 15 — Experiment 3 results. A — Welding current and weld pool parameters; B — control input; C — frontside bead; D — backside bead. welding speed set at 1.4 mm/s. From 15 to 60 s, the desired pool parameters are set at 7, 8, and 0.16 mm. The controller is able to adjust the welding speed to 0.7 mm/s. It was noticed that the weld pool could not reach the exact set-point as specified. This was expected because the formation of the weld pool is a physical process, and thus cannot be specified arbitrarily. From 60 to 100 s, the desired pool parameters were set at 6, 6.5, and 0.13 mm. The welding speed was controlled at 0.9 mm/s. The weld pool parameters were controlled at 5.9, 6.8, and 0.14 mm. Automated Welding Experiments Automated control experiments were designed and conducted. Previously, different set points were applied as well as different welding current and speed disturbances, and the robustness of the controller was verified. Experiment 1: Tracking Varying Set Points A control experiment was conducted for tracking varying set points. An open-loop period was first applied to establish the weld pool. From 57 to 85 s, the set points were set at 4.5, 6.5, and 0.11 mm, and from 85 to 110 s, the set points were changed to 5, 6, and 0.1 mm. It was observed that the controller was able to control the weld pool characteristic parameters by increasing the welding speed from 0.7 to 0.75 mm/s. The backside bead width was maintained at about 3 mm — Fig. 13D. It was noticed, however, that small variations occurred and the weld pool characteristic parameters were not accurately controlled at their set points. This was expected because the pool parameters are governed by physical law and only controlling welding speed is not sufficient to control three pool parameters. Experiment 2: Varying Welding Currents In this experiment, the robustness of the control algorithm against different welding currents was tested. From 60 to 80 s, the welding current was 43 A. The controller was able to bring the pool parameters to about the set-point (5, 6, and 0.1 mm) by adjusting the welding speed to 0.72 mm/s. After 80 WELDING RESEARCH s, the welding current was set at 46 A. As a result, the weld pool width, length, and convexity increased. However, the controller was able to increase the welding speed to about 0.9 mm/s to compensate for this heat input increase — Fig. 14B. The backside bead width was maintained at about 3 mm — Fig. 14D. Experiment 3: Speed Disturbance In Experiment 3, an artificial error between the calculated optimal welding speed and actual applied welding speed was applied to evaluate the robustness of the proposed control algorithm against speed disturbances. During the first 28 s after the openloop period, no error existed between the calculated optimal speed and actual applied speed, and the controller was able to control the pool parameters to the set points 5, 5, and 0.11 mm. However, from 85 to 87 s, the welding speed was set to 0.5 mm/s, which is about 0.23 mm/s greater than the calculated optimal welding speed. As can be seen from Fig. 15A, the weld pool width decreased, and the length increased. However, by applying the controller to hold the desired pool pa- APRIL 2015 / WELDING JOURNAL 133-s A B C D
Welding Journal | April 2015
To see the actual publication please follow the link above