Study on the Application of Sensors Array Technique for Monitor Spot Welding Quality

The valid method study for welding quality monitor has become a serious problem. It is difficult to use single sensor to identify the quality of actual nugget because of the affected factors in the welding process. A spot welding quality monitor system based on the sensors array technique is developed. A modified signal processing algorithm based on Fast ICA is proposed to extract the objected signals. The purpose is first to choose a perfectible way for the annular sensors array, then use the modified Fast ICA algorithm to extract the objective signals. The result shows that the best arranging way for the sensors array is the un-balanced model. And the magnetic signal source can be well separated with the modified Fast ICA method during the welding process.