GB/T 38683-2020 in English

This standard specifies a complete technical system for detecting inclusions ≥0.4mm in bearing steel bars using the liquid immersion pulse reflection method. Its innovation is reflected in:

• For the first time, the definition of ultrasonic reflection equivalent of large inclusions is clarified
• A graded sensitivity evaluation system (level 1-5) is established
• A two-stage control of process detection (C) and final detection (L) is proposed

Ultrasonic probe must meet the following requirements:

• Whip diameter: 6.3±0.3mm
• Nominal frequency: 10MHz (bandwidth 6±1MHz)
• Water layer thickness: 40-45mm (temperature 15-30℃)

Sensitivity calibration uses 1.0mm flat-bottom hole artificial reflector, requirements:

1. At least 4 reflectors of different depths
2. Hole spacing ≥ 2 times of chip diameter
3. Bottom wave and defect wave can be distinguished

Scanning method selection

• Cylindrical parts: diameter ≤50mm, use 180° fold scan; >50mm, use 360° spiral scan
• Cuboid: thickness ≤50mm, detect 2 vertical surfaces; >50mm, detect 4 surfaces

Sensitivity Setting

Benchmark sensitivity (1.0mm flat-bottom hole 80% wave height) superimposed graded gain:

Note: It is forbidden to use AVG curve to convert defect size. It must be calibrated by actual measurement.

Detection process of GCr15 steel billet (φ80mm) in a bearing steel plant:

1. Prepare the sample in C state, and turn the surface to Ra1.8μm
2. Use 10MHz probe, 42mm water layer, temperature 25℃
3. Use 360° spiral scanning, grid spacing 0.2×1.0mm
4. 3 inclusions ≥0.4mm were found, and the error was <0.05mm after metallographic verification

Compared with the old version, this standard has the following major breakthroughs:

• Newly added digital C-scan imaging requirements
• Specified signal-to-noise ratio ≥ 6dB (10dB recommended)
• Introduced three-dimensional adjacent criteria assessment rules
• Specified an 8-hour equipment calibration cycle