The aim of this thesis work is to develop an analysis dealing with the tyre contact patch, its shape and deformation during specific manoeuvres, such as free rolling and cornering. Furthermore, an innovative approach concerning the lateral force estimation is proposed. The context in which this study is performed is the project brought by Pirelli Tyre S.p.A. known as Cyber TyreTM, which is one of the most advanced smart tyre system. It is intended to process and analyse the great amount of information that can be extracted from the tyre and combined with the ones available from the other vehicle sensors, in order to implement new logics for improving both active and passive control systems, guaranteeing the vehicle safety. The investigation is based on the data acquired by three three-axial accelerometers positioned on the inner liner of the tyre, in correspondence of the tyre ribs. This solution guarantees the possibility of estimating with quite accurate results how the tyre contact patch deforms and which is the influence of vertical load, inflation pressure and tyre characteristic angles (i.e. camber and tyre slip angle) on the contact patch itself. The analysis exploits synthetic indices evaluated by Cyber TyreTM, especially in cornering conditions both in terms of radial displacement and patch length, investigating their reliability and, in some cases, suggesting some possible corrections based on the theoretical expected behaviour. As far as the lateral force estimation is concerned, a regression model by means of Artificial Neural Network, based on the already studied regressors describing the tyre deformation, is proposed. The effects of different training methods are investigated comparing an indoor training, carried out using MTS Flat TracR machine data, with an outdoor training based on measurements provided by dynamometric hubs. Finally, the lateral force estimations obtained with the ANN and with a simple physical model, such as the Single Track Vehicle Model, are combined thanks to a fuzzy logic, which is implemented in order to guarantee the best lateral force estimation in different tyre working conditions.
Lo scopo del seguente lavoro di tesi consiste nell’analizzare l’impronta a terra assunta dallo pneumatico, la sua forma e la relativa deformazione in diverse condizioni di manovra, quali dinamica di puro rotolamento e di scorrimento laterale. Viene, inoltre, presentato un metodo innovativo per la stima della forza laterale. Il lavoro è svolto nell’ambito del progetto sviluppato da Pirelli Tyre S.p.A. denominato Cyber TyreTM, uno dei più avanzati sistemi di smart tyre. Tale sistema permette di estrarre ed analizzare le informazioni relative alla dinamica dello pneumatico, le quali, combinate con quelle provenienti da altri sensori installati sul veicolo, vengono utilizzate per l’implementazione di nuove logiche di controllo in grado di migliorare la sicurezza del veicolo agendo sia in modo attivo che passivo sulla dinamica del veicolo stesso. Lo studio si basa sulle acquisizioni effettuate tramite tre accelerometri triassiali disposti sul liner interno dello pneumatico, in particolare, in corrispondenza dei ribs. Questa soluzione permette di stimare con buoni risultati la forma dell’impronta a terra assunta dallo pneumatico e gli effetti che il carico verticale, la pressione e gli angoli caratteristici dello pneumatico (angolo di camber e di deriva) hanno sull’impronta stessa. L’analisi sfrutta gli indici di sintesi inerenti alla deformazione dello pneumatico in direzione radiale e alla lunghezza di impronta, calcolati tramite Cyber TyreTM, in modo particolare in condizioni di scorrimento laterale. Vengono, inoltre, proposte delle possibili strategie di correzione degli indici qualora questi manifestassero andamenti non attesi e teoricamente non giustificabili. Viene descritto, quindi, un metodo di stima della forza laterale basato sull’utilizzo di una Rete Neurale Artificiale che compie un processo di regressione tra gli indici di deformazione dell’impronta precedentemente studiati e la forza laterale stessa. Vengono, quindi, analizzate le prestazioni di stima ottenute tramite due approcci di addestramento della rete neurale: i risultati ottenuti addestrando la rete su prove indoor vengono confrontati con i risultati ottenuti addestrando la rete su prove outdoor in cui la stima della forza di riferimento viene misurata tramite mozzi dinamometrici. Infine, per garantire una stima ottimale della forza laterale in ogni condizione, è stata sviluppata una logica fuzzy per combinare i risultati di stima ottenuti tramite l’utilizzo sia della Rete Neurale che di un semplice modello fisico quale quello Mono-Traccia.
Tyre contact patch shape analysis and tyre lateral force estimation through Cyber TyreTM technology
VIGNARCA, DANIELE;TERENGHI, TOMMASO
2018/2019
Abstract
The aim of this thesis work is to develop an analysis dealing with the tyre contact patch, its shape and deformation during specific manoeuvres, such as free rolling and cornering. Furthermore, an innovative approach concerning the lateral force estimation is proposed. The context in which this study is performed is the project brought by Pirelli Tyre S.p.A. known as Cyber TyreTM, which is one of the most advanced smart tyre system. It is intended to process and analyse the great amount of information that can be extracted from the tyre and combined with the ones available from the other vehicle sensors, in order to implement new logics for improving both active and passive control systems, guaranteeing the vehicle safety. The investigation is based on the data acquired by three three-axial accelerometers positioned on the inner liner of the tyre, in correspondence of the tyre ribs. This solution guarantees the possibility of estimating with quite accurate results how the tyre contact patch deforms and which is the influence of vertical load, inflation pressure and tyre characteristic angles (i.e. camber and tyre slip angle) on the contact patch itself. The analysis exploits synthetic indices evaluated by Cyber TyreTM, especially in cornering conditions both in terms of radial displacement and patch length, investigating their reliability and, in some cases, suggesting some possible corrections based on the theoretical expected behaviour. As far as the lateral force estimation is concerned, a regression model by means of Artificial Neural Network, based on the already studied regressors describing the tyre deformation, is proposed. The effects of different training methods are investigated comparing an indoor training, carried out using MTS Flat TracR machine data, with an outdoor training based on measurements provided by dynamometric hubs. Finally, the lateral force estimations obtained with the ANN and with a simple physical model, such as the Single Track Vehicle Model, are combined thanks to a fuzzy logic, which is implemented in order to guarantee the best lateral force estimation in different tyre working conditions.File | Dimensione | Formato | |
---|---|---|---|
Tesi_Terenghi_Vignarca.pdf
non accessibile
Descrizione: Tesi_Terenghi_Vignarca
Dimensione
22.59 MB
Formato
Adobe PDF
|
22.59 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/10589/164907