Data Clustering Utilization Technologies Using Medians of Current Values for Improving Arc Sensing in Unstructured...

Area of Science:

• Robotics and Automation
• Materials Science and Engineering
• Manufacturing Processes

Background:

• Welding automation in shipbuilding frequently uses arc-sensing due to space constraints.
• Short-circuit transitions in arc sensing reduce feedback current reliability with curved or gapped workpieces, causing seam tracking failures and damage.

Purpose of the Study:

• To develop a robust seam tracking algorithm for robotic welding in challenging shipbuilding environments.
• To enhance the accuracy and responsiveness of welding robots dealing with workpiece curvature and gaps.

Main Methods:

• Proposed a novel algorithm: median-based spatial clustering (MBSC).
• MBSC is based on the DBSCAN (density-based spatial clustering of applications with noise) clustering algorithm.
• The algorithm clusters data based on median values within weaving areas and analyzes feedback current data characteristics, using outliers to improve tracking.

Main Results:

• The MBSC algorithm demonstrated enhanced seam tracking accuracy.
• Improved responsiveness in unstructured and challenging welding scenarios was observed.
• Effectiveness was validated through real-world welding experiments in a shipyard.

Conclusions:

• The MBSC algorithm offers a significant improvement over traditional methods for robotic welding seam tracking.
• The technique is particularly effective in overcoming challenges posed by workpiece curvature and gaps.
• This advancement contributes to more reliable and damage-free automated welding in shipbuilding.