Damage Monitoring Of Bolted Connections To GFRP Plate Based On The Piezoelectric Impedance

In recent years,piezoelectric impedance technology has been widely used in the field of structural damage identification and health monitoring of civil engineering.However,the research on the damage detection of composite materials is relatively small.In this thesis,the feasibility of electromechanical impedance technology for glass fiber reinforced plastic plates was explored based on the existing research results of piezoelectric impedance technology.At the same time,the effect of temperature on the admittance signal of PZT-GFRP system and its temperature compensation were studied.What's more,BP neural network was used to compensate the effect of shear force.Combined the experiments with theoretical analysis,the results are as follows:The temperature sensitivity of PZT was analyzed based on the one-dimensional impedance model of single PZT and dual PZT,and the experimental study of temperature sensitivity was conducted.Its outer ring stimulating and inner ring receiving stimulation is used as dual PZT,the inner ring stimulating and inner ring receiving simulation is used as single PZT.The study found that dual PZT was less sensitive than single PZT.The effect of temperature on PZT-GFRP system was studied.It was found that the increase of temperature leads to the deviation of the admittance curve of PZT-GFRP system to the left,and the RMSD index gradually increased.In order to reduce the effect of temperature on it,this thesis compensates it by using the effective frequency offset method,which can effectively compensate the lateral and longitudinal deviation of the admittance curve caused by temperature.After the compensation,the RMSD has a significant declination.Experimental study on bolt loose damage detection of GFRP plate was conducted based on Piezoelectric Impedance Method.It is found that the increase loosing of the bolt and the increase of the external force will cause the admittance curve to move and increase the RMSD index.If the external force is not ruled out,it will lead to the misjudgment of the structural damage.In order to eliminate the influence of shear factors on the damage identification results,the BP neural network method was used to compensate the measured admittance signal.The BP neural network was established as a training sample with multiple sets of admittance signals in healthy state,and the RMSD index of different tension in healthy state was stabilized.Then the force and admittance under different working conditions was input network.The health state of the structure can be effectively judged by comparing the predictive admittance and the RMSD index of the measured admittance without the external force.