The Wavelet Time Element Approcah For Dynamic Load Identifcation And The Research On Moving Load Identification Technique

In the practical engineering problems, the acquisition of varibration source or the determination of the load applied on the structure, is more and more meaningful for many situations. In most cases, the exiciting loads change with time, namely the dynamic loads. An exact dynamic load has an important influence on the reliability and safety for the design of engineering structures. Based on the structural responses, this paper studies the theory of the dynamic load identification and moving load identificatoin. In the process of dynamic load identification, both the wavelet analysis and the minimum of the residual value are used to establish the forward problem model between the dynamic load and the structural response of the measured points. The forward problem model of the dynamic load identification is usually ill-posed, which means that a small noise contained in the measurement response will lead to a large error in the identification process, thus regularization method is used to solve this problem. The distributed parametric beam model and the influence with different noisy levels are researched for the moving load identification process. What’s more, the uncertain of the parameters is considered in the moving load identification. A series of researches have been done as follow:(1) A new dynamic load identification method, which is based on wavelet multi-resolution analysis and the weighted residual least square time element method, is proposed. The dynamic load is approximated by wavelet shape functions and the approximate load is brought into the original convolution equation of structure response. Define a residual function which is the difference of the measured response and the response calculated by the convolution equation, and make the sum of squares of the residual to be minimum, so as to establish the forward model of dynamic load identification. Due to the ill-posedness of the kernel function matrix and the unavoidable noise in the measured response, the results of load identification would be unaccepted. The regularization method is adopted to realize the stable identification.(2) Based on the beam model of distributed parameter, a new method of moving load identification is studied in this paper. This method adopts the model coordinate transform to calculate the model respo ne from the measured response. And then the forward model of each model and model load is established using the Green kernel method. Because the farword model is also ill-posed, regularization method is used to identify the model load. At last, the model load is transformed to the actual moving load via the inverse modal coordinate transform. Beacause t he regularization method has the effect to filter the noise, therefore it has the good ability to reduce the influence of the noise.(3) Considering uncertainty of structural parameters, this paper reasearchs the moving load identification method based on the interval analysis. The beam parametric uncertainty is discripted as interval, and then the moving load is expansed to Taylor expansion at the midpoint of parametic interval. The regularization method is firstly used to identify the moving load at the midpoint of the interval. Then, the central difference method is used for the uncertain parameters sensitivity analysis to get the sensitivity curves. Finally, the upper and lower bounds of moving load are calculated through combining the moving load at midpoint and the sensitivity curves. In order to verify the correctness and validity of the method, the upper and lower bounds of the moving load using the Monte Carlo method is given for the comparison.