Ahanpanjeh et al 2025a - Revision history

Maryam Ahanpanjeh at 23:26, 18 November 2025

2025-11-18T23:26:44Z

Maryam Ahanpanjeh at 23:25, 18 November 2025

2025-11-18T23:25:22Z

Maryam Ahanpanjeh at 23:22, 18 November 2025

2025-11-18T23:22:59Z

Maryam Ahanpanjeh at 23:22, 18 November 2025

2025-11-18T23:22:10Z

Maryam Ahanpanjeh at 23:19, 18 November 2025

2025-11-18T23:19:55Z

Maryam Ahanpanjeh at 23:17, 18 November 2025

2025-11-18T23:17:54Z

Maryam Ahanpanjeh at 23:12, 18 November 2025

2025-11-18T23:12:20Z

Maryam Ahanpanjeh at 23:12, 18 November 2025

2025-11-18T23:12:19Z

Maryam Ahanpanjeh at 23:07, 18 November 2025

2025-11-18T23:07:48Z

Maryam Ahanpanjeh at 23:06, 18 November 2025

2025-11-18T23:06:04Z

@@ Line 161: / Line 161: @@
 <div class="center" style="width: auto; margin-left: auto; margin-right: auto;">
 <span style="text-align: center; font-size: 75%;">''Fig. 11: Velocity of the end effector during welding measured via RSI and T-mac/laser tracker mounted on the Axis-6 of the robot.''</span></div>
 These findings offer important insights into the sources of resistance within the overall system, which directly impact speed stability during the welding process and highlight the need for component redesign in the next development phase.

@@ Line 104: / Line 104: @@
 [[Image:Review_888306976947-picture-x0000_t202.svg|center|288px]]
 These measurements were obtained using a laser tracker from Hexagon in combination with a T-Mac, sensor capable of 6D measurements, mounted on the end effector, as shown in Fig. 2. The perpendicular alignment of the sonotrode to the top adherend was ensured via the measurements conducted using reflectors and a laser tracker before welding. As shown in Fig. 4, the rotational deviations observed during welding were minimal, suggesting that the force distribution across the contact surface was relatively uniform. However, it is important to note that these measurements reflect the orientation of the end effector, not the sonotrode itself. In practice, sonotrode misalignments may still occur, particularly at points along the weld path where frictional forces are high. Since there is currently no method available to directly measure the orientation or angular deviations of the sonotrode during welding, such misalignments can only be inferred indirectly, posing a limitation for precise in-situ evaluation of sonotrode alignment.
@@ Line 123: / Line 124: @@
 [[Image:Review_888306976947-picture- 1.svg|center|281px]]
 Despite the robot being programmed for a constant welding speed, noticeable acceleration and deceleration phases were observed during the welding process. These variations are attributed to the high contact forces generated as the sonotrode and consolidator slide over the surface of the top adherend. The resulting frictional forces trigger the stick-slip effect, a phenomenon where motion alternates between sticking and slipping due to the difference between static and kinetic friction. The stick-slip behavior arises because static friction, which resists initial motion, is greater than kinetic friction, which governs motion once sliding begins. Initially, static friction resists motion until the applied force exceeds a threshold, triggering kinetic friction and a sudden acceleration beyond the set speed. As resistance builds, the system slows, static friction returns, and velocity drops below the target. These cycles repeat, especially when system stiffness is insufficient [10]. The stiffness of the combined robot and end effector structure plays a crucial role in this behavior. Lower stiffness leads to greater compliance, resulting in elastic deformation and amplification of stick-slip dynamics. This inconsistent velocity can lead to inconsistent heat generation and consolidation along the weld line, leading to inhomogeneous weld quality with potential defects.
@@ Line 135: / Line 137: @@
 As mentioned in the previous section, the robot’s absolute inaccuracies and the forces involved in the process cause deflections in both the robot and the end effector, leading to deviations from the desired path and misalignments. Therefore, it is essential to develop a compensation strategy to correct any deviations in the sonotrode’s position along the Y axis and its orientation relative to the top workpiece. To meet the demands of the highly dynamic robotic ultrasonic welding process, a fast-response sensor system is essential. For this purpose, a 2D laser scanner from Keyence was integrated into the end effector, as shown in Fig. 2. The scanner captures edge position data every 12 ms and transmits it to the RSI. At each sampling interval, the position and orientation of a target point on the weld path are measured. The deviation between the actual and desired pose of the sonotrode is calculated to generate an error signal. This signal is then fed into the robot controller, which adjusts the sonotrode’s pose accordingly. The correction is updated every 12 ms to maintain accurate tracking of the weld path. The implemented compensation strategy consisted of two RSI Programs. In the first programme (RSI-I), the laser scanner identified the initial weld point along the edge and determined the orientation of the top adherend. Based on this information, the sonotrode was positioned and aligned perpendicularly to the adherend’s surface. In the second programme (RSI-II), after the application of the necessary contact forces, the sonotrode followed the weld path perpendicular to the top adherend while compensating for real-time deviations using a proportional-integral (PI) controller. The results are illustrated in the figures 8 and 9. The only information that was given to the robot was the length of the weld line in the X direction.
 [[Image:Review_888306976947-picture- 2.svg|center|281px]]

@@ Line 83: / Line 83: @@
 <div class="center" style="width: auto; margin-left: auto; margin-right: auto;">
- [[Image:Review_888306976947-image1.png|378px]] </div>
+[[Image:Review_888306976947-image1.png|378px]] </div>
 <div class="center" style="width: auto; margin-left: auto; margin-right: auto;">

@@ Line 98: / Line 98: @@
 <div class="center" style="width: auto; margin-left: auto; margin-right: auto;">
- [[Image:Review_888306976947-image2.png|366px]] </div>
+[[Image:Review_888306976947-image2.png|366px]] </div>
 <div class="center" style="width: auto; margin-left: auto; margin-right: auto;">
 <span style="text-align: center; font-size: 75%;">''Fig.2: Welding end effector with sensors.''</span></div>
 [[Image:Review_888306976947-picture-x0000_t202.svg|center|288px]]
@@ Line 143: / Line 142: @@
 <div class="center" style="width: auto; margin-left: auto; margin-right: auto;">
- [[Image:Review_888306976947-image9.jpg|366px]] </div>
+[[Image:Review_888306976947-image9.jpg|366px]] </div>
 <div class="center" style="width: auto; margin-left: auto; margin-right: auto;">
@@ Line 155: / Line 154: @@
 <div class="center" style="width: auto; margin-left: auto; margin-right: auto;">
- [[Image:Review_888306976947-image10.jpg|366px]] </div>
+[[Image:Review_888306976947-image10.jpg|366px]] </div>
 <div class="center" style="width: auto; margin-left: auto; margin-right: auto;">

@@ Line 105: / Line 105: @@
 [[Image:Review_888306976947-picture-x0000_t202.svg|center|288px]]
 These measurements were obtained using a laser tracker from Hexagon in combination with a T-Mac, sensor capable of 6D measurements, mounted on the end effector, as shown in Fig. 2. The perpendicular alignment of the sonotrode to the top adherend was ensured via the measurements conducted using reflectors and a laser tracker before welding. As shown in Fig. 4, the rotational deviations observed during welding were minimal, suggesting that the force distribution across the contact surface was relatively uniform. However, it is important to note that these measurements reflect the orientation of the end effector, not the sonotrode itself. In practice, sonotrode misalignments may still occur, particularly at points along the weld path where frictional forces are high. Since there is currently no method available to directly measure the orientation or angular deviations of the sonotrode during welding, such misalignments can only be inferred indirectly, posing a limitation for precise in-situ evaluation of sonotrode alignment.
@@ Line 115: / Line 114: @@
 <div class="center" style="width: auto; margin-left: auto; margin-right: auto;">
- [[Image:Review_888306976947-image5.jpg|366px]] </div>
+[[Image:Review_888306976947-image5.jpg|366px]] </div>
 <div class="center" style="width: auto; margin-left: auto; margin-right: auto;">
@@ Line 123: / Line 122: @@
 <span id='_heading=h.3sb2ue8c07sr'></span>A correct and stable speed of the components at the end effector ensures consistent heating under the sonotrode and cooling under the consolidator, leading to uniform welds. To monitor this, the robot’s speed was tracked using a Robot Sensor Interface (RSI) program integrated into the robot’s welding program. Within the RSI program, velocity values were computed every 4 ms based on changes in the X and Y positions of the Tool Center Point (TCP), which in this case corresponds to the sonotrode.
 [[Image:Review_888306976947-picture- 1.svg|center|281px]]
@@ Line 137: / Line 135: @@
 :''''''3.1.''' In-line Pose Correction and Path-Planning'''
 As mentioned in the previous section, the robot’s absolute inaccuracies and the forces involved in the process cause deflections in both the robot and the end effector, leading to deviations from the desired path and misalignments. Therefore, it is essential to develop a compensation strategy to correct any deviations in the sonotrode’s position along the Y axis and its orientation relative to the top workpiece. To meet the demands of the highly dynamic robotic ultrasonic welding process, a fast-response sensor system is essential. For this purpose, a 2D laser scanner from Keyence was integrated into the end effector, as shown in Fig. 2. The scanner captures edge position data every 12 ms and transmits it to the RSI. At each sampling interval, the position and orientation of a target point on the weld path are measured. The deviation between the actual and desired pose of the sonotrode is calculated to generate an error signal. This signal is then fed into the robot controller, which adjusts the sonotrode’s pose accordingly. The correction is updated every 12 ms to maintain accurate tracking of the weld path. The implemented compensation strategy consisted of two RSI Programs. In the first programme (RSI-I), the laser scanner identified the initial weld point along the edge and determined the orientation of the top adherend. Based on this information, the sonotrode was positioned and aligned perpendicularly to the adherend’s surface. In the second programme (RSI-II), after the application of the necessary contact forces, the sonotrode followed the weld path perpendicular to the top adherend while compensating for real-time deviations using a proportional-integral (PI) controller. The results are illustrated in the figures 8 and 9. The only information that was given to the robot was the length of the weld line in the X direction.
 :''''''3.2.''' In-Situ Monitoring and Control of Forces'''

@@ Line 137: / Line 137: @@
 :''''''3.1.''' In-line Pose Correction and Path-Planning'''
-<div class="center" style="width: auto; margin-left: auto; margin-right: auto;">              [[Image:Review_888306976947-picture- 2.svg|center|283px]] </span></div>
+<div class="center" style="width: auto; margin-left: auto; margin-right: auto;">              [[Image:Review_888306976947-picture- 2.svg|center|283px]] </span></div></div>
 As mentioned in the previous section, the robot’s absolute inaccuracies and the forces involved in the process cause deflections in both the robot and the end effector, leading to deviations from the desired path and misalignments. Therefore, it is essential to develop a compensation strategy to correct any deviations in the sonotrode’s position along the Y axis and its orientation relative to the top workpiece. To meet the demands of the highly dynamic robotic ultrasonic welding process, a fast-response sensor system is essential. For this purpose, a 2D laser scanner from Keyence was integrated into the end effector, as shown in Fig. 2. The scanner captures edge position data every 12 ms and transmits it to the RSI. At each sampling interval, the position and orientation of a target point on the weld path are measured. The deviation between the actual and desired pose of the sonotrode is calculated to generate an error signal. This signal is then fed into the robot controller, which adjusts the sonotrode’s pose accordingly. The correction is updated every 12 ms to maintain accurate tracking of the weld path. The implemented compensation strategy consisted of two RSI Programs. In the first programme (RSI-I), the laser scanner identified the initial weld point along the edge and determined the orientation of the top adherend. Based on this information, the sonotrode was positioned and aligned perpendicularly to the adherend’s surface. In the second programme (RSI-II), after the application of the necessary contact forces, the sonotrode followed the weld path perpendicular to the top adherend while compensating for real-time deviations using a proportional-integral (PI) controller. The results are illustrated in the figures 8 and 9. The only information that was given to the robot was the length of the weld line in the X direction.

@@ Line 138: / Line 138: @@
 <div class="center" style="width: auto; margin-left: auto; margin-right: auto;">
-[[Image:Review_888306976947-picture- 2.svg|center|283px]]</div>
+[[Image:Review_888306976947-picture- 2.svg|center|283px]] </div>
 As mentioned in the previous section, the robot’s absolute inaccuracies and the forces involved in the process cause deflections in both the robot and the end effector, leading to deviations from the desired path and misalignments. Therefore, it is essential to develop a compensation strategy to correct any deviations in the sonotrode’s position along the Y axis and its orientation relative to the top workpiece. To meet the demands of the highly dynamic robotic ultrasonic welding process, a fast-response sensor system is essential. For this purpose, a 2D laser scanner from Keyence was integrated into the end effector, as shown in Fig. 2. The scanner captures edge position data every 12 ms and transmits it to the RSI. At each sampling interval, the position and orientation of a target point on the weld path are measured. The deviation between the actual and desired pose of the sonotrode is calculated to generate an error signal. This signal is then fed into the robot controller, which adjusts the sonotrode’s pose accordingly. The correction is updated every 12 ms to maintain accurate tracking of the weld path. The implemented compensation strategy consisted of two RSI Programs. In the first programme (RSI-I), the laser scanner identified the initial weld point along the edge and determined the orientation of the top adherend. Based on this information, the sonotrode was positioned and aligned perpendicularly to the adherend’s surface. In the second programme (RSI-II), after the application of the necessary contact forces, the sonotrode followed the weld path perpendicular to the top adherend while compensating for real-time deviations using a proportional-integral (PI) controller. The results are illustrated in the figures 8 and 9. The only information that was given to the robot was the length of the weld line in the X direction.