Online Detection of Welding Quality Based on ZYNQ and Data Mining

摘要

With the rapid development of manufacturing industry, traditional quality detection methods can no longer meet the demand. As an important part of intelligent manufacturing, the research of online quality monitoring technology of arc welding is imminent. The welding electrical signal reflects various changes of arc composition in the welding process and contains abundant information about the welding quality. Therefore, an online quality monitoring method based on Apriori algorithm is proposed. The ZYNQ board is used to sample welding electric signals under three shielding gas flow rates levels. Apriori algorithm is used to mine the potential distinguishing rules under three levels of shielding gas flow rates, and Verilog hardware language is used to design a suitable rule to monitor the shielding gas flow rate automatically. Finally, ZYNQ board is used to control the shielding gas flow rate of welding machine. Abundant online experiments have demonstrated that the classification results obtained by Apriori algorithm can distinguish the current signals of three shielding gas flow rates.