Intelligent Fusion Control Throughout Varying Thermal Regions

An intelligent monitoring and control system is presented that regulates the total heat input in order to maintain constant fusion zone geometry under varying thermal regions. The integrity and quality of the final fused joint rely heavily on the geometrical properties of the fusion zone, in particular the width-to-depth ratio. Assessment of the joint geometry is performed using Rayleigh’s suspended droplet analogy in which the mass, and hence the volume, of the droplet is related to the natural resonant frequency of the molten region. A non-intrusive, non-contact, top-side sensor collects the arc light reflected from the oscillations of the molten metal’s surface. An improved approach for inducing and monitoring the oscillations of a molten weld pool is presented. A software-based phase-locked loop (PLL) technique enables synchronized excitation of the molten pool. Improved locking and tracking characteristics in the presence of noise, signal distortions, and harmonic conditions are achieved through the use of intelligent signal monitoring algorithms which co-exist in parallel and cooperate with the PLL. Dynamic reconfiguration of the PLL’s digital filter coefficients allows superior tracking performance over a wide range of resonant frequency conditions. A fuzzy logic rule set, modeled after expert human welding knowledge, regulates the total heat input to the fusion process in order to demonstrate this unique synchronous excitation, monitoring, and control technique. A detailed discussion regarding each component’s contribution to the overall system is provided. Finally, an in-depth comparison between various openand closed-loop experiments of the joining process is discussed. Keywords—Industrial automation, arc welding, phase-locked loop, fuzzy logic, neural networks, control.