| Sensor technology has been developed for real-time monitoring and adaptive control of laser beam welding in this work, which include acoustic and optical monitoring of penetration condition of welds, optical measurement of gap width, weld fault detection, as well as adaptive control of beam focal position, beam position on the seam and welding speed. Acoustic sensor is an ordinary condensed microphone. Optical sensor is a unique system of many capabilities that has been developed in this work. With this sensor, a neural fuzzy control system has been designed to improve the control action for these parameter controls.; Acoustic monitoring is based on the variation in frequency distribution of the air-borne acoustic signals. Interpretation of the acoustic signals to extract information on weld quality is not straightforward. Three approaches have been proposed to identify the weld using acoustic signals: statistical deviation, statistical pattern recognition and neural network. Each method has shown a certain degree of accuracy of prediction.; Optical monitoring is to use the as-developed multi-spot and multi-color detection techique to measure gap width, to detect weld defeats and to identify penetration condition. With optical monitoring, high sampling rate is reached. In weld fault detection, an inverse-filtering algorithm has been developed to successfully resolve information on the weld faults.; Closed loop controls of beam focal position, seam tracking and welding speed have been successfully achieved during CO2 laser beam welding experiment. A neural fuzzy control system has been designed for adaptive control of welding speed and seam tracking using optical feedback signals. Simulation results show improvement in control action in comparison to conventional PID controller. |
