Research On De-noising With Wavelet Analysis In Damage Identification Of Grid Structure

Structural damage identification is to ensure that no major damage to buildings in the service life, resulting in significant casualties and property damage guarantee of. Due to the impact sensor accuracy and structure of the service environment, the structural response information collected will inevitably contain noise. Use noisy structural damage identification response information directly result in reduced accuracy of damage identification, which led to the identification results to identify errors or mistakes.Aiming at this phenomenon, the use of wavelet threshold de-noising structure of acceleration response signal noise reduction. This paper considers the structural characteristics of network structure, determine wavelet threshold noise reduction methods and decomposition level wavelet function, wavelet threshold function and wavelet threshold evaluation, obtain a suitable network structure of wavelet threshold noise reduction methods; then Numerical simulation of acceleration response information noise damage identification results were observed before and after noise reduction, confirmed the accuracy of this method of wavelet noise; Finally, experimental verification, to test the response of the information collected and compare the observed noise reduction Damage noise before and after the recognition results proved the effectiveness of this method of wavelet noise reduction in the actual structure. The main contents and steps are as follows:1. Using the wavelet threshold noise reduction methods for noise reduction acceleration information grid structure, with the acceleration of high-frequency information obtained by the wavelet analysis method, by contrast enumeration method of high frequency acceleration information obtained in different situations, to determine the appropriate wavelet function and decomposition levels, threshold function and threshold estimation method. Come up with a suitable truss damage identification method wavelet threshold noise.2. Establish a network structure finite element model, the acceleration factor as a response to injury, a two-step method for determining network structure damage identification. Use noisy information identification, determine the noise will affect the precision and accuracy of damage identification of truss structures. Then use the wavelet threshold noise reduction methods for noise reduction acceleration response, then the structural damage identification, the resulting damage identification accuracy of the results and the accuracy has been improved. Attest to the accuracy of wavelet threshold de-noising method proposed in this paper.3. Establishment of experimental model grid structure, the use of experimentally obtained acceleration grid structure before and after damage response signal. First, the acceleration response with noise identification, and then after the response acceleration noise damage identification. Noise damage identification results before and after comparison can be drawn from the results of noise after recognition accuracy is improved. Demonstrate the feasibility of wavelet threshold noise reduction methods proposed.This paper presents a suitable network structure damage identification method wavelet threshold noise reduction, noise reduction and proved this method can effectively improve the precision and accuracy of the grid structure damage identification.