| H-shaped steel structure has many advantages such as strong bending ability,light relative weight,low cost,easy construction,energy saving and environmental protection.H-beam steel is mostly used as beams,columns and other components,once the failure may lead to the entire structure of instability failure,which will cause major casualties and losses,but the existing monitoring technology has not yet been developed.It is of great significance to further research and develop the online real-time monitoring technology and non-destructive testing technology of H-beam.Based on the theory of Model Acoustic Emission(MAE),this paper first simulates typical in-plane excitation(In-Plane,IP)and out-of-Plane excitation(Out-of-Plane,OP)sound sources by means of lead-breaking excitation on the sides and upper surfaces of rectangular flat specimens,H-shaped steel webs,and wing edges,and analyzes the modal characteristic analysis and research of the obtained lead-breaking excitation signal,and analyzes its sound source propagation law.The visual recognition method of the modal components of a typical sound source signal is introduced,and the modal separation of the S0 and A0 modes in the acoustic emission signal is realized by using the Empirical Mode Decomposition(EMD)method.Secondly,the rectangular flat panel specimen with less boundary interference is used to simulate the plane of the H-shaped steel web,and the linear attenuation experiment of the acoustic emission signal is first performed to obtain the relevant information of the modal acoustic emission waveform on the rectangular flat panel specimen,and the simulation experiment is carried out on this basis,and the simulation experiment results and the physical experiment results are confirmed with each other.After that,the EMD decomposition algorithm and the Multi-resolution analysis method(MRA)were combined to initially establish a method that can reflect the size of the defect.In this process,at the same time,the waveform of the acoustic emission analog signal on the rectangular flat panel specimen was analyzed for sound source propagation.Finally,the linear attenuation experiment of the acoustic emission signal is carried out on the web of the H-shaped steel specimen,and the numerical simulation experiment is carried out based on the information of the collected modal acoustic emission waveform.The method established in the rectangular flat specimen that can reflect the size of the defect is introduced into the H-shaped steel structure experiment.The experimental results and the simulation experiment results confirm each other,proving that this method is also applicable to H-shaped steel.The lead-breaking excitation signals with different defects are modal decomposed using the EMD method,and the separated modes are used as eigenvalues to train the Artificial Neural Network(ANN),and the trained artificial neural network model is used to identify and classify different defect signals. |