| Acoustic emission testing technology is an effective detecting method for structural crack in welding process. Due to acoustic emission signals(AE signals)characteristics as multi-modal and nonlinear, caused by formation and state variation of cracks in welding process, together with friction between specimen and clap,interference from arc, AE signals present to have multiple AE sources co-existence phenomenon and is weak. Therefore, realizing separation and extraction of multisource excitation AE signals, providing theories and methods are the keys for welding cracks AE signals detection, which also possessing significant theoretical study and engineering application values.In this thesis, through manufacturing cracks in welding and simulating AE testing platform, combining AE signals time domain acquisition system, a simulative AE testing platform for welding cracks is established to do according AE signals collection experiments and characteristic analysis works. What’s more, this thesis also proposes using separation method for multisource AE signals based on harmonic analysis to separate and extract AE signals. The main contents of the thesis are as follows:(1) A simulative welding cracks AE testing platform is established through combination of adjustable experimental devices for transverse restraint stress and AE signals online acquisition system. This platform can simulate various physical processes like malformation, crack formation by using different welding procedures like manual welding, MIG welding, etc. Meanwhile, gathering and testing for multisource AE signals can be performed.(2) The proposal of multisource AE signal separation method based on harmonic analysis, in which signal segments are fitted into AR matrix, analysis of reconstruction matrix under SVD is used to determine the active signal harmonic layers. Afterwards, active harmonics are separated by FFT and thenreconstructed for effective signals. Signal harmonic time and frequency domain characteristics can be achieved simultaneously. In general, the former multi-source AE signal separationmethod is proved to be effective through strict analysis on simulation and emulation signal.(3) Respective simulation of friction excitation source, welding arc impact excitation source, welding crack excitation source for AE testing experiments. After efficient extraction and frequency analysis on the collected various excitation source AE signals, frequency domain characteristics for single excitation source AE signal are acquired accordingly.(4) Manual welding experiments are done and mixed excitation source AE signals are collected simultaneously during the heating and cooling welding process.Depending on frequency domain characteristics of various excitation AE signals,various excitation source AE signals separation method based on harmonic analysis is used to separate and extract AE signals during heating and cooling welding process severally, through which AE signals extraction for welding structural cracks are obtained significantly. |
PDF Full Text Request