2013 年 53 巻 2 号 p. 240-244
With the rapid development of microelectromechanical systems, small scale resistance spot welding (SSRSW) is ever-increasing used in electronic and medical devices. Whereas there is limited research work dealing with quality control of SSRSW. This paper investigated a real-time and in-situ SSRSW quality monitoring method by means of taking the voltage as the monitoring signature. It was obtained through clipping two leads onto the electrodes during SSRSW. As the linear DC and the high frequency (HF) resistance welding power supplies were the common equipments in SSRSW and constant current mode was used in this study, the variation of voltage with time indicated the conditions of the welding process which issued in the final weld quality. Utilizing four factors extracted from the voltage curve an artificial intelligence algorithm to estimate the weld quality was proposed. The maximum average forecast error of the trained network is about 0.15 mm, showing that the voltage curve is a reliable quality monitoring signature of SSRSW. The most prominent advantage of this method is that weld quality can be perfectly estimated with only two sensor clips compared with other methods reported for normal scale or large scale resistance spot welding (LSRSW).
