Study On Damage Identification Method Of Steel Structure Based On Weak Signal Identification

Steel structure is widely applied in transportation,civil engineering,road and bridge,mining and oil fields.The steel structures used in these large-scale engineering fields are more complex,and are in a long period of environmental corrosion and bear huge external forces,which make the steel structure easy to produce fatigue damage,crack damage and so on,and easily cause safety accidents.Therefore,ensuring the safe operation of steel structure has become the focus of attention of many scholars and related technicians.The background noise is unavoidable in the actual vibration response signal of the steel structure,and the strong background noise increases the difficulty of the structural damage based on the vibration signal recognition.Therefore,this paper studies the damage identification method of steel structure based on the weak signal recognition.Firstly,the basic theory of stochastic resonance is studied,and it is difficult to detect the weak signal under strong background noise.The weak signal of arbitrary frequency is successfully detected by the normalized variable scale stochastic resonance.The asymmetric stochastic resonance model driven by associated noise is proposed.The two state model theory and the fastest descent method are used to derive the expression of the signal to noise ratio and the average first crossing time of the asymmetric model driven by the correlated noise,and the two evaluation indexes of the signal to noise ratio and the average first crossing time are studied,and the correlated noise and the coefficients of the asymmetric terms and the correlation noise are also studied.It is found that adjusting the coefficient of asymmetric term can change the intensity of the best matching noise.It provides a theoretical basis for identifying weak signals in strong noise background based on stochastic resonance.The normalized scale transformation stochastic resonance model and cosine fitting technique are used to deal with the acceleration response of the beam under the simple harmonic excitation,and the response residual of the calculation response and measurement response and response sensitivity matrix are combined to identify the structural parameters of the beam and to locate the structural damage of the beam,thus realizes structural damage identification based on the response signal under strong noise background.The basic theory of Hilbert-Huang Transform(HHT)is studied.The end effects of Empirical Mode Decomposion(EMD)method is described.The cause of the end effects is analyzed.Then the extremum symmetric extension EMD is used to improve the end effect.The instantaneous frequency and instantaneous amplitude of the eigenmode component after decomposition are more accurate.It lays the foundation for the damage characteristics of the steel structure of the derrick.A structural damage identification method based on stochastic resonance and improved EMD is proposed for damage identification of derrick steel structure under strong noise background.Firstly,the natural frequency of the derrick under the strong noise background is identified by stochastic resonance,and the influence of the weak damage on the natural frequency of the steel structure is detected.Then the filter bandwidth is determined by the natural frequency and the improved EMD method is used to analyze the acceleration response of the finite element model and the experimental model of the derrick steel structure to bear the impact load,and the damage is identified.The damage location is identified,and the IMF natural frequency amplitude curvature is used to identify the damage effectiveness.