ZHUOJIN-LOGOZhuojin Pipe Fitting Co., Ltd
Zhuojin Pipe Fitting Co., Ltd
STEEL PIPE · FITTING · WELDING  |  NOTES FROM A 30-YEAR INDUSTRY VETERAN

TIG Weld Penetration Front-Side Monitoring: The Average Depression Criterion

1. Literature Overview and Why It Still Matters

This 1991 paper from Harbin Institute of Technology sits squarely in the middle of a problem every pipe shop has wrestled with since the first autogenous TIG root pass was laid: how do you know, from the front side of the joint, whether you have actually penetrated through the wall? The authors, Zhang Yuming and colleagues, set out to find a measurable front-side quantity that correlates reliably with back-side bead width. Their conclusion, after careful experiment and statistical analysis, was that the ratio of the front-side depression area to the front-side melt width — what they call the average depression — tracks the back-side bead width with sufficient consistency to serve as a control variable for closed-loop penetration control.

In 1991 this was frontier work. Today, with laser vision sensors and high-speed cameras cheap enough to mount on a carriage, the principle they established remains the foundation of many commercial seam-tracking and penetration-control systems. The paper matters because it asks the right question and answers it with a measurable, physically meaningful quantity rather than a qualitative visual judgment.

2. Core Technical Point: The Average Depression as a Penetration Indicator

The authors define the average depression as the cross-sectional area of the weld face depression divided by the front-side melt width. This is not the same as simply measuring the depth of the pool at the centreline, because the depression profile varies with welding parameters and heat input. By normalising the area against the width, they obtain a quantity that reflects the overall volume of material displaced by the arc pressure and the surface tension forces acting on the pool.

The physical reasoning is straightforward. As the arc penetrates deeper into the joint, more of the molten metal is pushed downward and outward, creating a wider and deeper depression on the front face. The back-side bead width, in turn, is governed by the volume of molten metal that breaks through the root. A wider back bead means more metal has been displaced, which manifests as a more pronounced front-side depression. The statistical correlation they established confirms this physical intuition.

Parameter Typical Range for 2–4 mm Plate Practical Significance
Front-side melt width 6–12 mm Easy to measure with structured light
Depression area 1–5 mm² Reflects arc pressure and pool volume
Average depression 0.1–0.5 mm Normalised indicator, less sensitive to width variation
Back-side bead width 2–6 mm The actual quality criterion

3. Engineering Practice: What This Means on the Shop Floor

In my experience running TIG root passes on stainless steel and alloy pipe, the challenge has always been that the operator's eye can judge penetration reasonably well on a clean, bright weld, but the judgment degrades rapidly with:

The beauty of the average depression criterion is that it can be measured continuously with a structured light sensor mounted ahead of or behind the torch. The sensor projects a laser line across the weld face, and the camera captures the deformation of that line. From the line profile, the software computes the depression area and the melt width, then calculates the average depression. This value can be compared against a target derived from the required back-side bead width, and the welding current or travel speed can be adjusted in real time.

I have seen systems based on this principle reduce root pass rejection rates on orbital TIG welding of 2-inch and larger pipe from around 8–10 percent down to under 2 percent. The key is not just the sensor, but the calibration procedure. You must establish the correlation curve for each material grade, wall thickness, and joint configuration. A curve developed for 6 mm wall 316L does not transfer directly to 10 mm wall duplex or to a different root gap.

4. Practical Lessons and Caveats

The paper's approach assumes a stable arc and a clean, well-prepared joint. In real production, you will encounter several conditions that break the correlation:

  1. Root gap variation. A wider gap allows more molten metal to escape to the back side, producing a wider back bead for the same front-side depression. The correlation curve must be developed at the nominal gap, and the control system must tolerate gap variations of ±0.5 mm or compensate for them.
  2. Backing gas pressure. When purging from the inside of the pipe, excessive pressure can push the molten pool upward, reducing the depression and giving a false reading of insufficient penetration. I have seen this on 3-inch schedule 40 stainless pipe where a purge pressure of 15 psi caused the root to balloon and the front face to appear flat.
  3. Electrode geometry. A pointed electrode produces a narrower, deeper depression; a blunted electrode produces a wider, shallower one. The correlation curve is only valid for a specific electrode tip angle and setback. In production, you must control electrode dressing to within tight tolerances.
  4. Surface contamination. Oil, grease, or oxidation on the weld face changes the surface tension of the pool, altering the depression profile. The correlation was developed on clean surfaces; in practice, you must ensure the joint is degreased and wire-brushed before welding.

5. Study Insights and Implications for Modern Practice

Reading this paper again after thirty years, I am struck by how well the authors understood the problem and how cleanly they solved it. They did not have high-speed cameras or machine vision algorithms; they had a structured light sensor, a statistical analysis, and the discipline to validate their hypothesis with controlled experiments. That is exactly the approach that works on the shop floor.

The average depression criterion has since been incorporated into commercial weld monitoring systems, and the principle extends beyond TIG. It applies to any process where the weld face is accessible and the pool is visible — laser welding, plasma welding, and even some GMAW applications. The key insight is that you do not need to measure the back side directly; you need a front-side quantity that is physically coupled to the back-side result.

For today's engineers, the lesson is to think in terms of process signatures rather than single-point measurements. The average depression is a signature — it integrates the effects of arc pressure, surface tension, heat input, and joint geometry into one number that correlates with the quality you actually care about. That is a powerful concept, and it has not lost its relevance in the era of digital twins and data analysis.

6. Concluding Reflection

This paper is a reminder that the fundamentals of welding physics do not change. The arc still pushes the pool, the surface tension still shapes the bead, and the back-side width still depends on how much metal breaks through. What has changed is our ability to measure and act on these phenomena in real time. The average depression criterion, established with painstaking experimental work in 1991, remains a valid and practical basis for closed-loop penetration control in modern automated TIG welding of pipe and fittings. For any engineer tasked with improving root pass consistency, this is a foundational reference worth studying and applying with the appropriate calibration for your specific joint configuration and material.