Font Size: a A A

Damage Identification Of The Structures Based On Data Fusion

Posted on:2007-03-16Degree:DoctorType:Dissertation
Country:ChinaCandidate:L JiaoFull Text:PDF
GTID:1102360212957640Subject:Disaster Prevention
Abstract/Summary:PDF Full Text Request
Owing to the influence of bearing load and sudden factors, structures have to face the damage problem throughout their lifetime. Small damage will come into being inevitably with the increasing use of the structures. After accumulating to some extent, the damage will reduce the safety, applicability and endurance of the structures, and even cause the breaking down of the whole structures. To detect the occurrence of the damage and protect or repair the structures in time, great attentions have been put to find the effective ways for the health monitoring of the civil engineering structures and base establishment in service. Damage identification technique is the most important one in health monitoring, which bears not only theoretical but also realistic meanings. With all these in consideration, the research achievements in damage identification have been summarized and the data fusion, wavelet analysis and algorithm for damage identification have been studied in this paper. The major work in this paper is list as follow:(1) The consensus algorithm for the data fusion of multisensors is improved. And a improved consensus algorithm based on the existing achievement for data fusion of multisensors is presented. It is very simple and could overcome the shortcomings of the existing consensus algorithm with two sensors, which has different confidence distance for different measuring precision. And the supporting matrix is fuzzified, which can avoid the subjective error in determining the threshold value. The effectiveness of this method is confirmed through numerical examples. It can be concluded that this method could make full use of the data from multisensors and it can reduce the error of fusion result caused by the disturbance.(2) A damage identification method for the data fusion of homogenous sensors is presented. According to Lipschitz index of wavelet analysis, the multiresolution analysis characteristics of white noise and damage signal is analyzed. And the principles of the wavelet denoising and signal identification are introduced.The structural dynamic system model is established by using the state space method. After numerical analysis of a five-layer frame structure, the results show that the sensors in different location bear different capability in denoting the damage feature of the structure. Each sensor could only denote the working status of the surrounding area. With the same distance from the damage, the sensors in lower layers are more sensitive than the higher layers in acceleration responding signal. So it will be more effective to put more sensors in the lower layers, if the number of the sensors is restricted. The damage identification method of homogenous sensors based on data fusion and...
Keywords/Search Tags:Data Fusion, Consensus Algorithm, Damage Identification, Coupled Neural Networks model, Partitioned Fusion
PDF Full Text Request
Related items