Nowadays checking the structural integrity of culturally, economically and strategically significant buildings is a primary problem. This is due both to the increase in the architectural complexity and dimensions of the new buildings, furthermore is necessary for the preservation of economically and strategically important, or historically and culturally valuable, structures. The same importance is shared by the testing of machinery and components that require high reliability , such as in the aerospace field. Structural Health Monitoring is the branch of engineering that is concerned with the analysis of the state of integrity of the structures, with a particular focus on civil structures. In the past, the integrity of the structures was checked through various methods of portion-by-portion inspection resulting in a high costs , high test-time and a lack of reliability of the methods. Therefore, since half of the last century, fostered by the development of computers, has started the search for methods that could give a more accurate indication of the damage from macroscopic investigation of the structure. In particular, in the last ten years, the methods that are based on vibration measurement and analysis have received much attention from the scientific community . The hypothesis of linearity is often used in engineering to simplify problems and models, for the same reason many damage identification methods were built using this hypothesis. However, in the case of damaged structures the assumption of linearity is not always true, so it becomes important to seek methods that do not rely on such assumptions. In the present study will be presented an overview of the major nonlinear damage identification methods. We will also try to verify the effectiveness of some of these methods on analytical models represented by a single degree of freedom system, which will be modeled as damaged and undamaged. For a first check of the systems are tested the methods based on coherence and Harmonic Distortion, while , for more reliable information on the state of damage, are considered methods based on bicorrelation, bicoherence and bispectrum. A model of one degree of freedom linear system has been used to represent the case of undamaged structure. While for cases of damaged structure have been used two models of one degree of freedom non-linear systems the first one with quadratic stiffness and the second one with cubic stiffness. At the end will be presented the results of the methods tested on the models
Al giorno d’oggi il controllo dell’integrità strutturale di edifici e strutture civili, con importanza culturale, economica e strategica, è un problema rilevante. Questo sia per via dell’aumento della complessità architettonica dei nuovi edifici sia per la costruzione di grandi opere, economicamente e strategicamente importanti, sia per la conservazione di edifici di alto valore culturale e storico. La stessa importanza riveste il controllo di macchinari e componenti che necessitino di un’alta affidabilità, come, ad esempio, nel campo aerospaziale. La branca dell’ingegneria che si occupa del controllo dello stato di integrità delle strutture, con una particolare attenzione alle strutture civili, è lo Structural Health Monitoring. In passato, la verifica dell’integrità delle strutture veniva fatta attraverso vari metodi di ispezione puntuale comportando un costo notevole, degli alti tempi di verifica e una non totale affidabilità dei metodi. Pertanto, da oltre metà del secolo scorso a questa parte, è iniziata, favorita dallo sviluppo dei calcolatori, la ricerca di metodi che potessero dare un’indicazione più accurata del danneggiamento a partire da un’indagine macroscopica della struttura; tali sono i metodi di identificazione del danneggiamento. In particolare negli ultimi dieci anni i metodi che si basano sulla misura di vibrazioni e la loro analisi hanno ricevuto molta attenzione dalla comunità scientifica. Come si sa l’ipotesi di linearità viene spesso usata in ingegneria per semplificare problemi e modelli, allo stesso modo molti metodi di identificazione sono stati basati su essa. Tuttavia nel caso di strutture danneggiate l’ipotesi di linearità non risulta sempre rispettata, quindi diventa importante ricercare metodi che non si basino su tali presupposti. Nel presente elaborato verrà presentato un quadro generale dello stato dell’arte sui principali metodi presenti in letteratura per identificare i danneggiamenti con metodologie non lineari. Inoltre si cercherà di verificare l’efficacia di alcuni di questi metodi su modelli analitici rappresentati da un sistema 1 gdl su cui modellare la presenza o meno del danno. Come metodi utili ad una prima verifica del sistema vengono provati i metodi di coerenza e Harmonic Distortion, mentre, per avere informazioni più robuste e affidabili sullo stato del danneggiamento, vengono presi in considerazione metodi basati su bicorrelazione, bicoerenza e bispettro. Per schematizzare il caso di struttura non danneggiata, è stato usato il modello di un sistema ad un grado di libertà lineare. Mentre per i casi di struttura danneggiata sono stati utilizzati due modelli di sistemi a un grado di libertà non lineari: uno con rigidezza quadratica e uno con rigidezza cubica. Al termine del lavoro verranno presentati i risultati dei metodi testati sui modelli.
Metodi non lineari per l'identificazione del danno : studio applicativo
BARBANTI, ALESSANDRO
2012/2013
Abstract
Nowadays checking the structural integrity of culturally, economically and strategically significant buildings is a primary problem. This is due both to the increase in the architectural complexity and dimensions of the new buildings, furthermore is necessary for the preservation of economically and strategically important, or historically and culturally valuable, structures. The same importance is shared by the testing of machinery and components that require high reliability , such as in the aerospace field. Structural Health Monitoring is the branch of engineering that is concerned with the analysis of the state of integrity of the structures, with a particular focus on civil structures. In the past, the integrity of the structures was checked through various methods of portion-by-portion inspection resulting in a high costs , high test-time and a lack of reliability of the methods. Therefore, since half of the last century, fostered by the development of computers, has started the search for methods that could give a more accurate indication of the damage from macroscopic investigation of the structure. In particular, in the last ten years, the methods that are based on vibration measurement and analysis have received much attention from the scientific community . The hypothesis of linearity is often used in engineering to simplify problems and models, for the same reason many damage identification methods were built using this hypothesis. However, in the case of damaged structures the assumption of linearity is not always true, so it becomes important to seek methods that do not rely on such assumptions. In the present study will be presented an overview of the major nonlinear damage identification methods. We will also try to verify the effectiveness of some of these methods on analytical models represented by a single degree of freedom system, which will be modeled as damaged and undamaged. For a first check of the systems are tested the methods based on coherence and Harmonic Distortion, while , for more reliable information on the state of damage, are considered methods based on bicorrelation, bicoherence and bispectrum. A model of one degree of freedom linear system has been used to represent the case of undamaged structure. While for cases of damaged structure have been used two models of one degree of freedom non-linear systems the first one with quadratic stiffness and the second one with cubic stiffness. At the end will be presented the results of the methods tested on the modelsFile | Dimensione | Formato | |
---|---|---|---|
Metodi non lineari per identificazione del danno studio applicativo.pdf
non accessibile
Descrizione: Testo della tesi
Dimensione
1.81 MB
Formato
Adobe PDF
|
1.81 MB | Adobe PDF | Visualizza/Apri |
I documenti in POLITesi sono protetti da copyright e tutti i diritti sono riservati, salvo diversa indicazione.
https://hdl.handle.net/10589/87164