Contactless monitoring of welding processes with computer processing of acoustic emission signals

Abstract

Signals processing and identification with informative characteristics of technological process, namely, the signals of acoustic emission, is a complex multi-level problem to solve it needs a lot of efforts. However, design of an integral, versatility system for analysis of these signals has not been announced yet. It is determined in this paper that there is a relationship between the parameters of laser welding technological process and registered acoustic emission signals. And that relationship is independent on investigation method as well. Universality should also apply to the employ of such approach to acoustic thermometry when welding live tissues where the welded tissue or laser beam is an exciter of acoustic oscillations. The signals of acoustic emission have been chosen in a feedback channel with laser welding technological process, since their transmission and analysis can be performed at minimum consumption of time and computational resources. The experimental part, by the example of laser welding of metal, demonstrates the relationship between the result of technological process (defect-less welded joint) and registered signals. A concept of a system for analyzing the accumulation and identification of signal informative characteristics was proposed and implemented based on the results of previous analysis of the signals by means of Spectrum Laboratory v.2.7, Sound Forge 10. The effect of parameters of time-frequency signal transform (band width, frequency range of the analysis, duration of time window and window function) on the registered signal was investigated. Model quickness under conditions of increased detail of the analysis indicates the necessity of optimizing the processing algorithms using time characteristics and performance of additional experiments with model registration and identification of the signals of laser welding technological process at different modes. The preliminary results were obtained to determine the dependence of the acoustic emission parameters on the temperature at the welding of living tissues. The perspective of using the model for its employ in acoustic thermometry has been shown.