The Method Research To Model Updating Of Finite Element And Structural Damage Identifying With Less Parameter

With the rapid development of China’s economic construction, all aspects of the country’s infrastructure is also developing rapidly, many large and representative civil engineering structures has been built gradually. These large buildings such as large span structure, super high rise building,Sea-Crossing Bridge and so on plays an important role in the development of the society, But because of its large scale the structure have become increasingly complex. The incidence of various damage also increases, So the importance of civil engineering structural health monitoring and damage identification is more and more prominent. Make damage analysis of structure, on the one hand it can prevent some potential damage occurred in the whole structure, save the cost of maintenance; On the other hand, can determine the damaged structure’s condition of damage, Make the repair scheme, To avoid the possibility of suddenly accidents of structure. Damage identification is the key part of the health monitoring system of structure, The accuracy of results of damage identification directly affects the health monitoring system’s function.At present the mainly method of damage identification is damage identification that based on finite element model updating, but the finite element model updating and damage identification both are the inverse analysis problem, Because of the widely application foreground in the actual project it has became the popular research topic. Currently the methods used more frequently is to through nonlinear optimization or regression analysis method to search the suitable model parameter vector. Using the element finite model’s experimental modal fitting the modal of theory, make the error norm of this two modals minimum. In order to get the finite element method that can reflect the actual situation of structure model. But the problem’s singularities is very strong, on the one hand model updating and damage identification’s results are very sensitive to the parameters set that need to be selected to be updated; on the other hand, the optimization process of nonlinear have great uncertainty, the iterative process may converge to a local minimum rather than the global minimum, so it don’t have practical engineering significance. For more complex structures or damage mode, need to establish complex finite element model so that can correctly simulate the structure, the element needed is numerous, this problem need to use large nonlinear optimization algorithm to solve, this will lead to lower stability calculation results, can hardly get meaningful results. Aiming at these problems, this article has did the following works:1.From the angele of multiscale analysis, using the method of the second generation wavelet transform, making the stiffness model updating as example, expressing the the updated stiffness curve of structure by using the format of multi-scale, realizing the updating of parameters of finite element model and damage identification of structure, reducing the uncertainty in the model updating process. Put forward the low scale expression of finite element model’s updating stiffness curve. Using the limited experimental information, through genetic algorithm to obtain the updating stiffness curve’s scaling function coefficients and wavelet coefficients which is in the condition of the curve is expressed in low scale, obtain the updating stiffness curve.2. Using multiscale analysis proposed, realizing geometric dimension reduction of the complex finite element model, obtain the simple finite element model similar to the fine model in limited range of parameters, and realizing model updating and damage identification on this simple finite element model.3. Using a variable cross-section box beam as an example, through calculating the different damage conditions designed and comparing the calculation result to demonstrate the effectiveness of the proposed method.