Development of an acoustic emission monitoring system for crack detection during arc welding

Meng, Xing (2019). Development of an acoustic emission monitoring system for crack detection during arc welding. University of Birmingham. Ph.D.

Preview

Meng2019PhD.pdf
Text - Accepted Version
Available under License All rights reserved.
Download (42MB) | Preview

Abstract

Condition monitoring techniques are employed to monitor the structural integrity of a structure or the performance of a process. They are used to evaluate the structural integrity including damage initiation and propagation in engineering components. Early damage detection, maintenance and repairs can prevent structural failures, reduce maintenance and replacement costs, and guarantee that the structure runs securely during its service life. Acoustic emission (AE) technology is one of the condition monitoring methods widely employed in the industry. AE is an attractive option for condition monitoring purposes, the number of industrial applications where is used is rising. AE signals are elastic stress waves created by the fast release of energy from local sources occurring inside of materials, e.g. crack initiating and propagating. The AE technique includes recording this phenomenon with piezoelectric sensors, which is mounted on the surface of a structure. The signals are subsequently analysed in order to extract useful information about the nature of the AE source. AE has a high sensitivity to crack propagation and able to locate AE activity sources. It is a passive approach. It listens to the elastic stress waves releasing from material and able to operate in real-time monitoring to detect both cracks initiating and propagating. In this study, the use of AE technology to detect and monitor the possible occurrence of cracking during the arc welding process has been investigated. Real-time monitoring of the automated welding operation can help increase productivity and reliability while reducing cost.

Monitoring of welding processes using AE technology remains a challenge, especially in the field of real-time data analysis, since a large amount of data is generated during monitoring. Also, during the welding process, many interferences can occur, causing difficulties in the identifications of the signals related to cracking events. A significant issue in the practical use of the AE technique is the existence of independent sources of a signal other than those related to cracking. These spurious AE signals make the discovering of the signals from the cracking activity difficult. Therefore, it is essential to discriminate the signal to identify the signal source. The need for practical data analysis is related to the three main objectives of monitoring, which is where this study has focused on. Firstly, the assessment of the noise levels and the characteristics of the signal from different materials and processes, secondly, the identification of signals arising from cracking and thirdly, the study of the feasibility of online monitoring using the AE features acquired in the initial study.

Experimental work was carried out under controlled laboratory conditions for the acquisition of AE signals during arc welding processing. AE signals have been used for the assessment of noise levels as well as to identify the characteristics of the signals arising from different materials and processes. The features of the AE signals arising from cracking and other possible signal sources from the welding process and environment have also collected under laboratory conditions and analysed. In addition to the above mentioned aspects of the study, two novel signal processing methods based on signal correlation have been developed for efficiently evaluating data acquired from AE sensors.
The major contributions of this research can be summarised as follows. The study of noise levels and filtering of different arc welding processes and materials is one of the areas where the original contribution is identified with respect to current knowledge. Another key contribution of the present study is the developing of a model for achieving source discrimination. The crack-related signals and other signals arising from the background are compared with each other. Two methods that have the potential to be used in a real-time monitoring system have been considered based on cross-correlation and pattern recognition. The present thesis has contributed to the improvement of the effectiveness of the AE technique for the detection of the possible occurrence of cracking during arc welding.

Type of Work:

Thesis (Doctorates > Ph.D.)

Award Type:

Doctorates > Ph.D.

Supervisor(s):