Research On Model Updating Of Overhead Type Cranes And Damage Identification Of Main Girder Structure

The large cranes often work in bad conditions.In addition,because of the high cost of the crane,its service life is usually extended on the basis of theoretical design life.In order to effectively check and deal with the potential safety hazards caused by the bad working environment and extended service,as well as promote the establishment and development of the damage assessment system of crane structure,this paper takes simplified gantry crane prototypes as the research object,and studies various identification methods for different degrees of damage:Firstly,the modal test of the prototype is carried out.Based on the test data,the finite element model of the prototype is updated by the approximate model updating technology.The maximum error between the simulated and the experimental value of the target natural frequency after updating is less than 4%,and the comparison of frequency responses also verifies the updating effect well.Then the updated finite element model is used as the reference model.Taking the natural frequency as the input and the damage index as the output,the simulation damage sample set is established by optimizing Latin hypercube method.Based on the radial basis function neural network algorithm,the global damage assessment prediction model is established to evaluate the main girder structure of the severely damaged experimental prototype quickly and accurately,which provides an effective reference for judging whether the crane is scrapped or not.Next,the damage location method based on the change of flexibility matrix is introduced into the study of local multiple corrosion damage identification of crane girder structure.With the help of the updated model and the prefabricated local damage prototype,the identification effect of the method is verified by simulation and experiment.The results show that the method can identify the serious local corrosion damage by moving the excitation position under the condition of limited sensors,but it is difficult to identify the small crack damage.At last,a method based on dynamic strain in time domain is proposed to identify the tiny damage which is difficult to locate.Through experiments,the identification effects of different excitation positions or different signal decomposition methods are compared and analyzed.The results show that the damage index based on the total empirical mode decomposition can identify the crack obviously,and the index size can reflect the crack depth to a certain extent.