This study presents an automated induction brazing system for thin-walled aluminum waveguide assemblies, integrating dual-loop control with real-time positional feedback. Two proportional control strategies were evaluated: a baseline single-loop heating-rate controller and an enhanced dual-loop configuration incorporating active inductor realignment based on thermal asymmetry. The single-loop system, operating at 20 Hz, achieved an average absolute temperature error of 3.4 °C with a 92% success rate, but exhibited sensitivity to workpiece-inductor gap variations, leading to inter-element temperature deviations exceeding 15 °C under moderate misalignments. To address this, a secondary feedback loop was introduced to dynamically correct flange-to-tube temperature differentials, improving thermal uniformity across the brazing zone. Full-scale validation on 120 assemblies across four geometric classes demonstrated that the dual-loop system reduced average temperature error to 2.1 °C, limited peak temperature spread to 8 °C, and increased yield to 97%. The control architecture features distinct bandwidth separation for power and positional loops, minimizing cross-coupling. A modular C++ software environment synchronizes multi-source data streams-including pyrometry, manipulator kinematics, power metrics, and video analytics-with real-time KPI tracking and long-term logging to an SQL-based historian. These results highlight the potential of feedback-driven control architectures in precision thermal joining, and suggest pathways for future enhancements using adaptive or fractional-order control algorithms.
Building similarity graph...
Analyzing shared references across papers
Loading...
Sergei Kurashkin
Bauman Moscow State Technical University
Dmitry Martysyuk
Bauman Moscow State Technical University
Sergei Kurashkin
Bauman Moscow State Technical University
Scientific Reports
Bauman Moscow State Technical University
Building similarity graph...
Analyzing shared references across papers
Loading...
Kurashkin et al. (Thu,) studied this question.
synapsesocial.com/papers/6988278b0fc35cd7a884664e — DOI: https://doi.org/10.1038/s41598-026-37593-w