Laser Weld Quality Inspection Using Plasma Anomaly Scoring
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing laser welding quality inspection methods face challenges in providing consistent defect detection due to sensitive plasma sensing values that fluctuate with process environment changes, requiring manual adjustment and leading to increased man-hours and reduced detection reliability, especially in mass production settings.
Innovation Solution
A machine learning-based system that processes radiation plasma data using a training model to determine anomaly scores, allowing for consistent quality inspection by classifying data into normal and defective categories, minimizing the need for manual adjustments and reducing the influence of process environment trends.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If plasma sensing values are used for quality inspection, then defect detection capability is improved, but sensing values fluctuate with process environment changes requiring manual adjustment
Solution Approach 1:
The system automatically adapts to process environment changes by using the welding position detection unit to identify the current bead position and automatically selecting the corresponding upper and lower limit values from multiple position-specific sets, eliminating the need for manual adjustment when environment changes occur
Solution Approach 2:
The system dynamically adjusts the quality determination criteria by switching between multiple sets of upper and lower limit values based on the detected welding position, making the inspection system adaptive rather than static
2Measurement precision
If multiple quality determination standards are set for different welding positions, then detection accuracy is improved, but system complexity and man-hours for maintenance increase
Solution Approach 1:
The welding position detection unit serves multiple functions: it detects the current bead position, selects the appropriate set of upper and lower limit values, and enables the single inspection system to handle multiple welding positions with position-specific criteria, reducing the need for separate inspection systems for each position
3Ease of manufacture
If traditional upper and lower limit value methods are used, then implementation simplicity is maintained, but detection reliability decreases due to minute defects being misclassified
Solution Approach 1:
The system applies different quality determination criteria (different sets of upper and lower limit values) to different welding positions based on the detected bead position, allowing each position to be evaluated with locally optimized parameters that account for position-specific variations in plasma characteristics
Applied Scientific Principles
This section explains which scientific principles are used to turn an abstract innovation direction into a practical engineering solution.
Function Achieved in This Case
The system provides reliable and efficient quality inspection by continuously updating the training model, reducing the need for manual threshold adjustments and improving detection accuracy by distinguishing between normal and defective welds.
Implementation Method 1
by considering the fact that processing waves are emitted from a welded portion during laser welding, when the welding quality is good, the radiated processing wave (plasma) is converted into an optical signal (near infrared, near ultraviolet)
Data Source
Figure 1
Figure 2
AI summary
The present disclosure relates to a system and method for checking quality of laser welding. The system and method for checking quality of laser welding of the present disclosure relates to a technology capable of applying a consistent defect determination criterion for laser welding quality by applying an anomaly score that is inferred based on machine learning.FIG. 1