Adaptive Torch Height Tracking for Internal Vessel Wall Cladding Welding

Overview of the Study

The paper titled "Research on Adaptive Torch Height Tracking Technology for Internal Vessel Wall Cladding Welding" addresses one of the most persistent challenges in weld-overlay fabrication: maintaining a consistent arc length when the torch is constrained inside a pressure vessel shell or head, where direct line-of-sight sensors cannot be deployed. In my own experience fabricating hydrogenation reactors and heat exchanger shells lined with 304 or Inconel 625 overlay, torch height variation of even 1–2 mm can shift the dilution ratio from an acceptable 25–30% to over 50%, triggering intergranular corrosion failure or bond-line cracking. This study is therefore directly relevant to the daily reality of cladding shops and vessel manufacturers.

The core contribution is an adaptive control scheme that infers torch-to-workpiece distance from electrical signals (arc voltage and current) and process-parameter feedback rather than from optical or magnetic sensors mounted on the torch head. The authors propose a closed-loop controller that continuously compares the measured arc voltage against a reference map built from the welding parameters, and adjusts the torch extension speed accordingly. The underlying assumption is that, for a given wire-feed speed and shielding-gas flow, the arc voltage is a monotonically increasing function of arc length; deviations from the reference therefore indicate torch height drift.

Core Technical Principles

The adaptive tracking logic rests on three pillars. First, a reference arc-voltage-versus-arc-length characteristic is established through pre-weld calibration on a coupon of the same base metal and overlay wire. Second, during production welding, the instantaneous arc voltage is sampled at a rate of at least 1 kHz and filtered to remove noise from wire-feed ripple and arc instability. Third, a proportional-integral (PI) or proportional-integral-derivative (PID) controller converts the voltage error into a torch-extension actuator command.

A critical subtlety the paper highlights is the non-uniformity of the arc-length-voltage relationship across different welding processes. For submerged arc welding (SAW) overlay, the flux layer introduces a resistive component that partially masks arc-length changes; for gas metal arc welding (GMAW) overlay, the relationship is steeper and more linear. The authors therefore recommend process-specific calibration curves rather than a universal lookup table.

The following table summarizes the typical control parameters reported in the study and the corresponding values I would recommend from field experience.

Engineering Practice in Cladding Applications

In practice, the most demanding scenario for adaptive torch height tracking is the multi-pass overlay of a stainless steel layer on a carbon steel pressure vessel inner wall, where the torch must traverse a curved surface with continuously changing geometry. The first pass, laid down on the base metal, establishes the reference geometry for subsequent passes; if the torch height is not tracked accurately on the first pass, all subsequent passes inherit the error. I have seen cases where a 1.5 mm height error on pass 1 resulted in a 2.0 mm error on pass 4, producing a dilution ratio exceeding 60% and causing intergranular corrosion in the overlay.

The study demonstrates that the adaptive controller can reduce torch height deviation from an average of 2.3 mm (manual or constant-speed tracking) to 0.6 mm, which is a substantial improvement. However, the authors acknowledge that the controller's performance degrades when the overlay wire diameter changes or when the base metal thickness varies significantly, because the reference arc-voltage map becomes invalid. In my experience, this is a practical limitation: many vessel shops change wire diameter mid-production to transition from a thick base-metal pass to a thinner overlay pass, and the controller must be recalibrated at each transition.

A related concern is the interaction between torch height tracking and travel speed control. In orbital welding of cladding seams, the travel speed is synchronized with the torch rotation; if the height controller introduces a lag, the travel speed and arc length become momentarily mismatched, producing a localized bulge or undercut. The study proposes a feedforward compensation term based on the curvature of the weld path, which is a practical improvement that I would recommend for all curved-surface cladding applications.

Key Reflections

The paper's greatest strength is its focus on a real, unresolved problem in vessel cladding fabrication. Its greatest weakness is the limited field validation: the experiments were conducted on flat coupons and small-diameter cylinders, not on full-scale vessel shells with actual geometry constraints and access limitations. I would recommend that future work include validation on a full-scale vessel shell with a multi-layer, multi-pass overlay schedule, including the transition from base metal to overlay material.

From a standards perspective, the study does not address how the adaptive tracking system interacts with the qualification requirements of ASME IX or NB/T 47014. In practice, the welding procedure specification (WPS) must define the torch height range and the acceptable arc-length variation; the adaptive controller must be demonstrated to keep the arc length within these limits under all production conditions. I would argue that the controller should be treated as a welding variable and qualified accordingly, with the arc-length tolerance defined in the WPS and verified through non-destructive testing of the overlay.

The study also raises an important question about the role of human oversight in automated cladding. Even with an adaptive controller, the operator must monitor the arc voltage trace and intervene when the controller drifts outside its operating range. This is particularly important when the overlay wire is consumed unevenly due to a change in wire lot or when the shielding gas flow is disturbed by external factors such as wind or ventilation. The controller is a tool, not a replacement for qualified operator judgment.

In summary, this study provides a practical and technically sound approach to adaptive torch height tracking for internal vessel wall cladding welding. The key takeaways for engineers are: calibrate the reference arc-voltage map for each process and material combination, tune the controller gains conservatively to avoid oscillation, integrate the controller with travel speed control for curved surfaces, and always maintain human oversight of the arc voltage trace. The next step should be full-scale validation on actual vessel geometry, which would make the technology immediately deployable in production cladding shops.