7716

7188

كارشناسي ارشد

الكترونيك

اصفهان: دانشگاه صنعتي اصفهان، دانشكده برق و كامپيوتر

1391

نه،108ص.: مصور،جدول،نمودار

ص.ع.به فارسي و انگليسي

بهزاد نظري، رسول امير فتاحي

حميدرضا ميردامادي

22/2/92

محمدرضا احمدزاده، محمدصادق فاضل

مهندسي برق و كامپيوتر

ID7188

به فارسي و انگليسي: قابل رويت در نسخه ديجيتالي

Abstract Due to increasing awareness of the importance of the maintenance topic in different industries guided ultrasonic wave propagation technique has been widely used for structural healthmonitoring SHM of structures in recent years SHM systems are one of the most important toolsfor evaluating and ensuring the safety of structural systems and constructions Several researcheshave been done to design SHM systems for diagnosing damages in different mechanical aerospaceand civil structures In this thesis guided ultrasonic wave propagation based SHM is studied Atfirst SHM is introduced as an online and suitable method for the maintenance of structures andguided wave based SHM is explained Afterwards signal processing of guided wave signals as themost important step in a damage detection process is presented Different signal processing aspectsincluding pre processing signal processing and feature extraction pattern recognition and makingthe damage detection process intelligent are described Next identification of damage in a thicksteel beam is studied as a sample SHM application By using available finite element simulationand experimental signals identification of the location and severity of created damage isinvestigated In this study different algorithms are introduced and are used for processing thesignals and determining the damage characteristics At first de noising and compressing of guidedwave signals by means of discrete wavelet transform DWT and wavelet packet transform WPT respectively is studied and the effect of choosing different orthogonal and bi orthogonal motherwavelets in the quality of de noising and compressing processes is investigated The results areshowing the better performance of bi orthogonal wavelets in guided wave signals de noising andcompression Next an algorithm is proposed based on some general statistical features of thesignals for classifying the damage severities This algorithm is based on WPT and support vectormachines SVM and classifies the damage severities into four classes no damage low mediumand high severity damages The proposed algorithm could classify the damage conditions in finiteelement signals perfectly The performance of the algorithm is verified by comparing it with somesimilar methods Furthermore an algorithm is proposed based on continuous wavelet transform CWT and fixed grid wavelet network FGWN for detecting the damage location and severity This algorithm uses time of flight ToF amplitude and area of damage reflected wave as damagefeatures By comparing the proposed method with an algorithm based on artificial neural network ANN and two other existing methods called DCP and DDF the results are showing superiorityof the proposed method At the end as online processing is a key factor in SHM systems hardwareimplementation of convolution method for damage localization on FPGA cores is simulated Adamage localization system is created in Active HDL software and corresponding simulations areperformed The simulation results are showing a good damage localization precision Keywords Structural health monitoring SHM Guided ultrasonic wave propagation Damageidentification Signal processing Wavelet transform WT De noising Compressing Neuralnetworks NN Support vector machines SVM Wavelet networks WN Hardwareimplementation FPGA

بهزاد نظري، رسول امير فتاحي

حميدرضا ميردامادي

محمدرضا احمدزاده، محمدصادق فاضل