The present work was carried out at the Department of Mechanical Engineering at Politecnico di Milano. The objective consists in the implementation and consecutive application of methodologies aimed at estimating the lateral and roll rail track irregularities from running dynamics measurements of the rail vehicle. The estimation of rail track irregularity turns out to be an important problem in the maintenance of rail networks. In fact, the continuous monitoring of track irregularity is one of the most relevant factor for the purpose of ensuring running safety and a proper level of ride comfort for the passengers. For these reasons, some methods, able to provide irregularities estimations, have been developed. The first method described is developed in the frequency domain and it is a procedure of pseudo-inversion of the frequency response matrix. Due to the restrictive assumptions that must be satisfied in order to work in the frequency domain, it was considered appropriate to pursue the possibility to develop different methodologies. Therefore the work continues with the implementation of some methods formulated in the time domain: a deterministic Unknown Input Observer type, and a stochastic Kalman Filter methodology. In order to apply these methods, an effort was made to develop a rail vehicle model system able to reproduce, with good approximation, the real behaviour of the train. To this aim, a 17-degrees-of-freedom model of a railway vehicle was developed. The identification methods performances were evaluated through numerical experiments, achieving with the software AD3S virtual running dynamics measurements of the rail vehicle, corresponding to track irregularities of High Speed railway lines belonging to Rete Ferroviaria Italiana. The results, obtained through the application of the methods to the 17-degrees-of-freedom system, were analysed and compared with each other, thus being able to provide an evaluation of the different identification methodologies developed, based on the accuracy of the irregularity estimate provided by each of them.
Il presente lavoro di tesi è stato svolto presso il Dipartimento di Ingegneria Meccanica del Politecnico di Milano. L’obiettivo prefissato consiste nell’implementazione e successiva applicazione di metodi finalizzati alla stima dell’irregolarità laterale e di rollio del binario a partire da misure di dinamica di marcia del veicolo ferroviario. La stima dell’irregolarità del binario ferroviario risulta essere un problema importante nell’ambito della manutenzione delle reti ferroviarie. Infatti il monitoraggio dell’irregolarità è uno dei fattori maggiormente rilevanti ai fini di garantire un livello sufficiente di comfort percepito a bordo treno. Per questi motivi sono state sviluppate differenti metodologie capaci di fornire una stima delle irregolarità. Il primo metodo descritto è sviluppato nel dominio delle frequenze e risulta essere una procedura di pseudo-inversione della matrice di risposta in frequenza. A causa delle ipotesi restrittive che è necessario formulare per procedere nel dominio delle frequenze, si è ritenuto opportuno investigare la possibilità di sviluppare metodologie differenti. Perciò si è proseguito con l’implementazione di alcuni metodi nel dominio del tempo: una metodologia deterministica di tipo Unknown Input Observer, e una stocastica di tipo Kalman Filter. Per applicare tali metodi è stato definito un modello matematico del veicolo ferroviario in grado di riprodurre, con buona approssimazione, il comportamento reale del treno. Si è così giunti ad una modellazione attraverso un sistema a 17 gradi di libertà. Le prestazioni dei metodi di identificazione sono state valutate sulla base di esperimenti numerici, realizzando, mediante il simulatore AD3S, misure virtuali della dinamica di marcia del veicolo, corrispondenti a profili di irregolarità misurati sulle linee ad Alta Velocità appartenenti a Rete Ferroviaria Italiana. I risultati, ricavati dall’applicazione dei metodi al sistema a 17 gradi di libertà, sono stati analizzati e confrontati tra loro, potendo così fornire una valutazione delle diverse metodologie di identificazione sviluppate, sulla base dell’accuratezza della stima dell’irregolarità fornita da ciascuno di essi.
Metodi per la stima dell'irregolarità laterale e di rollio del binario mediante misure di dinamica di marcia del veicolo ferroviario
DE ROSA, ANNA
2014/2015
Abstract
The present work was carried out at the Department of Mechanical Engineering at Politecnico di Milano. The objective consists in the implementation and consecutive application of methodologies aimed at estimating the lateral and roll rail track irregularities from running dynamics measurements of the rail vehicle. The estimation of rail track irregularity turns out to be an important problem in the maintenance of rail networks. In fact, the continuous monitoring of track irregularity is one of the most relevant factor for the purpose of ensuring running safety and a proper level of ride comfort for the passengers. For these reasons, some methods, able to provide irregularities estimations, have been developed. The first method described is developed in the frequency domain and it is a procedure of pseudo-inversion of the frequency response matrix. Due to the restrictive assumptions that must be satisfied in order to work in the frequency domain, it was considered appropriate to pursue the possibility to develop different methodologies. Therefore the work continues with the implementation of some methods formulated in the time domain: a deterministic Unknown Input Observer type, and a stochastic Kalman Filter methodology. In order to apply these methods, an effort was made to develop a rail vehicle model system able to reproduce, with good approximation, the real behaviour of the train. To this aim, a 17-degrees-of-freedom model of a railway vehicle was developed. The identification methods performances were evaluated through numerical experiments, achieving with the software AD3S virtual running dynamics measurements of the rail vehicle, corresponding to track irregularities of High Speed railway lines belonging to Rete Ferroviaria Italiana. The results, obtained through the application of the methods to the 17-degrees-of-freedom system, were analysed and compared with each other, thus being able to provide an evaluation of the different identification methodologies developed, based on the accuracy of the irregularity estimate provided by each of them.File | Dimensione | Formato | |
---|---|---|---|
2016_04_DeRosa.pdf
non accessibile
Descrizione: Testo della Tesi
Dimensione
8.11 MB
Formato
Adobe PDF
|
8.11 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/10589/118803