| Summary: | 碩士 === 國立屏東科技大學 === 車輛工程系碩士班 === 94 === Micro-resistance spot welding (MRSW) is a group of micro-joining processes in which micro-joints are formed between two sheets by resistance heating caused by the passage of electric current. These processes are commonly applied to the weld of auto electrical components, and micro-electrical components, and medical packing. Because of its many advantages such as high manufacturing speed and low cost, the MRSW is very suitable for industrial mass-production applications.
Surprisingly, there is no satisfactory non-destructive on line monitoring systems to assure the quality of the welding process. This thesis is aimed to develop a MRSW equipment and its associated monitoring system that can predict the welding quality during the welding process.
The thesis was preceded by designing an experimental test rig for the MRSW. Then a series of experiments were carried out to find the key parameters that can feature the welding quality. Experiments have shown that the values of the maximum electrode displacement and minimum dynamic resistance of a joint relate its welding quality quite well. Using these two parameters, a neural network based on-line welding quality monitor was developed.
The developed quality monitor system has been successfully applied to the welds of 0.1mm stainless steel and 0.1594mm KOVAR sheets, respectively. In the experiments, the achieved success rates in the two-class welding quality classification for the KOVAR sheets was 100% whilst the three-class classification for the stainless steel sheets was 93%.
The results show that the proposed MRSW monitor system processes an excellent accuracy in predicting the quality of a welding joint in terms of the tensile strength.
The developed monitoring system has the advantage of easy implementation in the field.
|
|---|
