Effective sensor clustering for online monitoring in transmission towers

Health monitoring systems are essential to investigate the performances during the service life of a structure such as the three-dimensional transmission tower. however,many real-world sensors and transducers require sensor positioning drafts before a computer based measurement procedure can effectively and accurately acquire the signal. this paper develops a direct physical property adjustment method,named as cross-model cross-mode method. in dealing with spatially incomplete situations,model reduction schemes were used. the selection procedure of the inactive degrees of freedom in process of the model reduction evaluated with a reasonable criterion by using the sensitivity analysis of system response under base excitation. also,the noisy data measurements are the other crucial factors. the success rates based on the correct detection probability factor were defined in order to evaluate the noise effect on the accuracy of the method. the efficiency of the method is validated by different damage scenarios. the results show that the developed methods are suitable for damage classification. but in the cases of the less used sensors than the degrees of the freedoms,the location of the sensors must be considered an important factor influencing the success rates. © school of engineering,taylor’s university.