Ultrasonic phased array inspection technology, by flexibly controlling the sound beam with multiple-element probes, can achieve high-precision three-dimensional imaging of the penetration depth of stud welds. The core method for evaluating the uniformity of the penetration depth is as follows:
I. Technical Implementation Principle
Beam Deflection and Focusing
Electronically delayed control of 128/256-element probe arrays creates multi-angle (0°-70°) focused sound beams, covering the entire depth of the weld seam.3D Data Acquisition
Employing a combined mechanical and electronic scanning method:
AxialThe probe advances along the axis of the screw bolt (step distance ≤ 0.5mm).
Zhou XiangElectronic fan扫 covers 360° weld seams
DepthPulse Echo Method for Measuring Melting Depth (Resolution up to 0.1mm)
Two: Uniformity of Melting Depth Evaluation Method
Feature Parameter Extraction
Melting Depth AverageAll testing points' arithmetic mean of melt depth
Melting Depth Standard DeviationReflects the degree of penetration depth fluctuation (optimal at σ < 0.3mm)
Melt Depth - Small/Large ValuesIdentifying localized overmelting or incomplete fusion defects
Imaging Analysis Technology
C-Scan ImagingGenerate a two-dimensional grayscale image of the weld cross-section, providing a direct visualization of the penetration distribution.
S-scan ProjectionExpand the 3D data to generate a distribution curve of the melt depth along the circumferential direction of the screw column.
Dynamic Threshold Setting
Based on material sound velocity (e.g., steel: 5900m/s, aluminum: 6300m/s), automatically calculate the theoretical melting depth and set a ±10% deviation threshold.
III. Key Parameters for Test System Configuration
| Parameters | Recommended Value | Clarification |
|---|---|---|
| Central Frequency | 5MHz | Balancing Penetration and Resolution |
| Array pitch | 0.5mm | Minimize grating lobes effect |
| Sampling Rate | 100MHz | Ensure the integrity of the echo signal |
| Focus on Depth | Mid-weld seam | Dynamic Focus on Technology Optimization |
| Scanning Speed | ≤50mm/s | Ensure Data Collection Density |
Typical Defect Identification
UnmeltedThe melting depth curve exhibits localized minor values (less than 80% of the design value).
OvermeltingMelt depth exceeds the thickness of the base material, accompanied by backside burn-through characteristics.
PorousA scanning signal exhibits "grass-like echo" and the melt depth measurement value is abnormal.
V. Data Post-Processing Workflow
Signal FilteringWavelet packet denoising applied to retain the 2-8MHz frequency band
Edge DetectionBased on Canny algorithm for fusion line extraction
3D ReconstructionGenerated melt depth distribution model via Voronoi algorithm
Statistical ReportOutput histogram of melt depth distribution and defect location coordinates
Section VI: Engineering Application Optimization
Temperature CompensationWhen the weld seam temperature exceeds 80℃. the sound velocity (Δv ≈ 3m/s/℃) must be adjusted.
Coupling OptimizationUtilizing high-temperature coupling agents (such as glycerol-based gels stable above 50°C)
Comparison Test BlocksCustomized screw studs with artificial defects (such as φ0.5mm flat-bottom holes) for calibration
By integrating AI image recognition technology (such as the U-Net neural network), automatic grading of uniformity in weld penetration can be achieved, boosting detection efficiency by over 50%. In practical applications, it is recommended to combine it with X-ray inspection for a comprehensive assessment of internal weld quality.
