This study provides a comprehensive modeling evaluation of the dynamic stability of Long Combination Vehicles (LCVs) that are commonly operated on U.S. highways, using multibody dynamic simulations in MATLAB/Simulink®. The dynamic equations for a tractor with two trailers connected by an A-frame converter dolly (A-Dolly) are developed. The dynamic model is used for running MATLAB® simulations, with parameters that are obtained through measurements or obtained from other sources. The simulation results are verified using track test data to establish a baseline model. The baseline model is used for parametric studies to evaluate the effect of trailer cargo weight, center of gravity (CG) longitudinal location, and trailer wheelbase. The dynamic model is further used to analyze both single-trailer and double-trailer trucks through nondimensionalization. The nondimensionalization method has the added advantage of enabling studies that can more broadly apply to various truck configurations. The simulation results indicate that increasing the trailer wheelbase reduces rearward amplification due to the damping effect of the longer wheelbase. A larger momentum ratio due to increased trailer gross weight increases rearward amplification. The detailed models of pneumatic disc and drum brakes in LCVs, including the airflow delay and thermal characteristics, are also developed and are coupled with the articulated vehicle dynamic models. The disc and drum brake braking performance are evaluated and compared in straight-line braking and combined steering and braking at a 150-ft J-turn maneuver. In straight-line braking, the simulation results indicate that disc brakes provide significantly shorter braking distance than drum brakes at highway speeds on a dry road, mainly due to their larger braking torque. On a slippery road surface, however, the greater braking torque causes more frequent wheel lockup and ABS activation at higher speeds, and disc brakes do not provide a substantially shorter braking distance than drum brakes. The simulations also point out that the disc brakes' cooling capacity is higher than the drum brake, with the cooling efficiency heavily dependent on the airflow speed. At higher driving speeds, the airflow accelerates to a turbulent flow and increases the convection efficiency. For braking in-turn maneuvers, at higher entering speeds, disc brakes decelerate the vehicle slightly sooner and then scrub speed faster, resulting in better roll stability when compared with drum brakes. / Doctor of Philosophy / Long combination vehicles (LCVs) are the combination of a tractor and two or more trailers and have been widely used on U.S. highways for cargo transport. Although LCVs have a larger cargo volume and provide more modularity in transporting goods, at higher speeds, they can be more prone to rollovers and require longer stopping distances and space to maneuver from one lane of travel to another. This study investigates the dynamic stability of an LCV, A-double trailer that includes a tractor, two trailers, and a dolly through modeling and simulation. The dynamic equations of each vehicle unit are derived based on Newtonian Mechanics (i.e., F = ma). The dynamic models are tuned to match the track testing results for similar vehicles, performed by the Center for Vehicle Systems and Safety (CVeSS) at Virginia Tech in the past. A novel evaluation method that nondimensionalizes the equations is used to allow for ease of use for LCVs with different cargo weights, lengths, and other similar variations. The dimensionless parameters are the function of vehicle parameters and express the relationship among the magnitude of vehicle parameters. Using the nondimensionalized model, the study performs a frequency analysis of the effect of trailer cargo weight, CG longitudinal position, and trailer wheelbase on roll stability and rearward amplification. Rearward amplification is the ratio of peak lateral acceleration between the tractor and the rearmost trailer. Slow-sweeping sinusoidal steering from 0.01 Hz to 0.6 Hz is used for the simulation analysis. The simulation results show that by increasing the trailer wheelbase—the distance from the trailer kingpin to the axle—the vehicle is more laterally stable because the longer wheelbases make the trailer more resistive to spinning around. Additionally, the pneumatic disc and drum brake models and thermal models are developed and coupled with the vehicle dynamic model. The disc and drum brake braking performance are investigated during both straight-line braking and combined steering and braking in a curve. The disc brakes generate a greater brake torque compared with drum brakes, and as such can decelerate the vehicle more efficiently on dry road surfaces, particularly at higher speeds such as highway speeds. This improves avoidance during emergency stops and roll stability during traveling in a curve, such as at a highway exit. The disc brakes also have greater cooling capacity because they can transfer the generated heat to the air due to the greater airflow and turbulence caused naturally by their design. This greatly helps to keep the brakes cooler on the track and to improve their stopping efficiency.

Columns are critical elements of any structure and their failure can lead to the catastrophic consequences of progressive failure. In structural design, procedures to design structures to resist conventional loads are well established. But design for accidental loading conditions is increasingly requested by clients and occupants in modern engineering designs. Among many accidental causes that induce column failure, impact (e.g. vehicular impact, ship impact, crane impact, impact by flying debris after an explosion) has rarely been considered in the modern engineering designs of civil engineering structures such as buildings and bridges. Therefore, most of the design requirements for structural members under vehicle impact as suggested by the current standards and codes such as Eurocode 1 are based on simple equations or procedures that make gross assumptions and they may be highly inaccurate. This research aims to develop more accurate methods of assessing steel column behaviour under vehicle impact.The first main objective of this study is to numerically simulate the dynamic impact response of axially loaded steel columns under vehicle impact, including the prediction of failure modes, using the finite element method. To achieve this goal, a numerical model has been proposed and validated to simulate the behaviour and failure modes of axially loaded steel columns under rigid body impact using the commercial finite element code ABAQUS/Explicit. Afterwards, an extensive parametric study was conducted to provide a comprehensive database of results covering different impact masses, impact velocities and impact locations in addition to different column boundary conditions, axial load ratios and section sizes. The parametric study results show that global buckling is the predominant failure mode of axially unrestrained compressed steel columns under transverse impact. The parametric study results have also revealed that column failure was mainly dependent on the value of the kinetic energy of impact. The parametric study has also shown that strain rate has a minor effect on the behaviour and failure of steel columns under low to medium velocity impact. The parametric study results have been used to develop an understanding of the detailed behaviour of steel columns under transverse impact in order to inform the assumptions of the proposed analytical method.To account for the effect of vehicle impact on the behaviour of steel columns, a simplified numerical vehicle model was developed and validated in this study using a spring mass system. In this spring mass system, the spring represents the stiffness characteristics of the vehicle. The vehicle stiffness characteristics can be assumed to be bilinear, with the first part representing the vehicle deformation behaviour up to the engine box and the second part representing the stiffness of the engine box, which is almost rigid. The second main objective of this research is to develop a simplified analytical approach that can be used to predict the critical velocity of impact on steel columns. The proposed method utilizes the energy balance principle with a quasi-static approximation of the steel column response and assumes global plastic buckling as the main failure mode of the impacted column. The validation results show very good agreement between the analytical method results and the ABAQUS simulation results with the analytical results tending to be on the safe side. A detailed assessment of the design requirements suggested by Eurocode 1, regarding the design of steel columns to resist vehicle impact, has shown that the equivalent static design force approach can be used in the design of moderately sized columns that are typically used in low multi-storey buildings (less than 10 storeys). For bigger columns, it is unsafe to use the Eurocode 1 equivalent static forces. It is acceptable to use a dynamic impulse in a dynamic analysis to represent the dynamic action of vehicle impact on columns, but it is important that both the column and vehicle stiffness values should be included when calculating the equivalent impulse force – time relationship. It is also necessary to consider the two stage behaviour of the impacting vehicle, before and after the column is in contact with the vehicle engine. A method has been developed to implement these changes.

Este trabalho obtém as frequências naturais e seus respectivos modos de vibrar da massa suspensa e não suspensa de um veículo comercial através de um programa computacional desenvolvido em MatLab Simulink. As rodas do veículo são excitadas no modelo teórico com sinais harmônicos de frequência continuamente variável de 0 a 40 Hz e de 0 a 4 Hz, faixas que contém as frequências naturais da massa suspensa e não suspensa. As frequências naturais foram obtidas analisando-se a amplitude máxima das forças normais nos pneus, ângulo de arfagem e rolamento da massa suspensa e outras variáveis (respostas do veículo à excitação imposta na roda dianteira esquerda). A roda esquerda dianteira do veículo foi excitada em laboratório com movimento harmônico aleatório cujo power spectral density mostra uma faixa relativamente plana na faixa de 0 a 22 Hz. As frequências naturais foram obtidas analisando-se o PSD da resposta do veículo medida nos eixos e na massa suspensa. As frequências naturais da massa suspensa e não suspensa obtidas teoricamente foram comparadas com as obtidas no ensaio em laboratório. O conhecimento das frequências naturais da massa suspensa e não suspensa é o primeiro passo para a análise do ride do veículo. Os resultados práticos e teóricos mostraram-se próximos mostrando que o programa computacional desenvolvido em MatLab Simulink é uma opção para o projeto e otimização da suspensão de veículos comerciais. / This study has the natural frequencies and its respective ways of vibration the lifted mass and not lifted of a commercial vehicle through a computerized program developed at MatLab Simulink. The wheels of this vehicle are excited on the theoretic model with harmonic frequency signals continuously variable from 0 to 40 Hz and from 0 to 4 Hz, ranges that have the natural frequencies of the lifted mass and not lifted. These natural frequencies were obtained by analyses of the maximum amplitude of the normal strength on the tires, pitch angle, riding of the lifted mass and variables (reaction of the vehicle to the excitement stressed on the front left wheel). The front left wheel was excited in lab in harmonic random motion of which the power spectral density shows a range reasonable steady at the range of 0 to 22 Hz. The natural frequencies were obtained by the analyses of the PSD the reaction of the vehicle measured on the axels and the lifted mass. The natural frequencies of the lifted and not lifted mass theoretically obtained were compared with the ones obtained at the lab. The data on the natural frequencies of the lifted and not lifted mass is the first step to the analyses of the ride of the vehicle. The practical and theoretical results appear to close showing that the computerized program developed at the MatLab Simulink is an option for the project and suspension upgrades on commercial vehicles.

Este trabalho obtém as frequências naturais e seus respectivos modos de vibrar da massa suspensa e não suspensa de um veículo comercial através de um programa computacional desenvolvido em MatLab Simulink. As rodas do veículo são excitadas no modelo teórico com sinais harmônicos de frequência continuamente variável de 0 a 40 Hz e de 0 a 4 Hz, faixas que contém as frequências naturais da massa suspensa e não suspensa. As frequências naturais foram obtidas analisando-se a amplitude máxima das forças normais nos pneus, ângulo de arfagem e rolamento da massa suspensa e outras variáveis (respostas do veículo à excitação imposta na roda dianteira esquerda). A roda esquerda dianteira do veículo foi excitada em laboratório com movimento harmônico aleatório cujo power spectral density mostra uma faixa relativamente plana na faixa de 0 a 22 Hz. As frequências naturais foram obtidas analisando-se o PSD da resposta do veículo medida nos eixos e na massa suspensa. As frequências naturais da massa suspensa e não suspensa obtidas teoricamente foram comparadas com as obtidas no ensaio em laboratório. O conhecimento das frequências naturais da massa suspensa e não suspensa é o primeiro passo para a análise do ride do veículo. Os resultados práticos e teóricos mostraram-se próximos mostrando que o programa computacional desenvolvido em MatLab Simulink é uma opção para o projeto e otimização da suspensão de veículos comerciais. / This study has the natural frequencies and its respective ways of vibration the lifted mass and not lifted of a commercial vehicle through a computerized program developed at MatLab Simulink. The wheels of this vehicle are excited on the theoretic model with harmonic frequency signals continuously variable from 0 to 40 Hz and from 0 to 4 Hz, ranges that have the natural frequencies of the lifted mass and not lifted. These natural frequencies were obtained by analyses of the maximum amplitude of the normal strength on the tires, pitch angle, riding of the lifted mass and variables (reaction of the vehicle to the excitement stressed on the front left wheel). The front left wheel was excited in lab in harmonic random motion of which the power spectral density shows a range reasonable steady at the range of 0 to 22 Hz. The natural frequencies were obtained by the analyses of the PSD the reaction of the vehicle measured on the axels and the lifted mass. The natural frequencies of the lifted and not lifted mass theoretically obtained were compared with the ones obtained at the lab. The data on the natural frequencies of the lifted and not lifted mass is the first step to the analyses of the ride of the vehicle. The practical and theoretical results appear to close showing that the computerized program developed at the MatLab Simulink is an option for the project and suspension upgrades on commercial vehicles.

<p>Automated Vehicle Path Following Control (PFC) is an advanced control system that can regulate the vehicle into a collision-free region in the presence of other objects on the road. Common collision avoidance functions, such as forward collision warning and automatic emergency braking, have recently been developed and equipped on production vehicles. However, it is impossible to develop a perfectly precise vehicle model when the vehicle is driving. The most PFC did not consider uncertainties in the vehicle model, external disturbances, and parameter variations at the same time. To address the issues associated with this important feature and function in autonomous driving, a new vehicle PFC is proposed using a robust model predictive control (MPC) design technique based on matrix inequality and the theoretical approach of the hybrid $\&$ switched system. The proposed methodology requires a combination of continuous and discrete states, e.g. regulating the continuous states of the AV (e.g., velocity and yaw angle) and discrete switching of the control strategy that affects the dynamic behaviors of the AV under different driving speeds. Firstly, considering bounded model uncertainties, norm-bounded external disturbances, the system states and control matrices are modified. In addition, the vehicle time-varying longitudinal speed is considered, and a vehicle lateral dynamic model based on Linear Parameter Varying (LPV) is established by utilizing a polytope with finite vertices. Then the Min-Max robust MPC state feedback control law is obtained at every timestamp by solving a set of matrix inequalities which are derived from Lyapunov stability and the minimization of the worst-case in infinite-horizon quadratic objective function. Compared to adaptive MPC, nonlinear MPC, and cascade LPV control, the proposed robust LPV MPC shows improved tracing accuracy along vehicle lateral dynamics. Finally, the state feedback switched LPV control theory with separate Lyapunov functions under both arbitrary switching and average-dwell-time (ADT) switching conditions are studied and applied to cover the path following control in full speed range. Numerical examples, tracking effectiveness, and convergence analysis are provided to demonstrate and ensure the control effectiveness and strong robustness of the proposed algorithms.</p>

Nos dias de hoje, a simulação dinâmica de veículos metroviários tem sido um importante método de análise e verificação, tanto no âmbito de projeto, como na engenharia de manutenção. Os pacotes de simulação disponíveis no mercado são altamente desenvolvidos e podem ser adaptados para os mais diversos cenários. O presente trabalho tem por objetivo demonstrar a funcionalidade destes programas, fundamentando a teoria por trás de seu algoritmo computacional, de maneira a expandir a fronteira de conhecimento nas engenharias de projeto e de manutenção. Para isso, são efetuados dois estudos de caso: passagem de veículo em região de abertura de bitola em curva de raio constante; passagem de veículo em sequencia de curvas reversas. Sendo assim, esta dissertação apresenta uma visão geral sobre o estado da arte na simulação da interação veículo e via permanente. / Nowadays, the dynamic simulation of metro vehicles have been an important method of analysis and verification in design engineering as in maintenance engineering. The simulation packages available are highly developed and can be adapted for different scenarios. This paper aims to demonstrate the functionality of these programs, supporting the theory behind its computational algorithm, in order to expand the frontier of knowledge in the national design and maintenance engineerings. For this, two case studies are made of: modelling flexible permanent way; modelling AMVs and crossings; passing vehicle in a region with gauge opening on a constant radius curve; passing a vehicle sequence of reverse curves. Thus, this dissertation presents an overview of the state of the art in the simulation of the interaction vehicle and permanent way of metro.

Nos dias de hoje, a simulação dinâmica de veículos metroviários tem sido um importante método de análise e verificação, tanto no âmbito de projeto, como na engenharia de manutenção. Os pacotes de simulação disponíveis no mercado são altamente desenvolvidos e podem ser adaptados para os mais diversos cenários. O presente trabalho tem por objetivo demonstrar a funcionalidade destes programas, fundamentando a teoria por trás de seu algoritmo computacional, de maneira a expandir a fronteira de conhecimento nas engenharias de projeto e de manutenção. Para isso, são efetuados dois estudos de caso: passagem de veículo em região de abertura de bitola em curva de raio constante; passagem de veículo em sequencia de curvas reversas. Sendo assim, esta dissertação apresenta uma visão geral sobre o estado da arte na simulação da interação veículo e via permanente. / Nowadays, the dynamic simulation of metro vehicles have been an important method of analysis and verification in design engineering as in maintenance engineering. The simulation packages available are highly developed and can be adapted for different scenarios. This paper aims to demonstrate the functionality of these programs, supporting the theory behind its computational algorithm, in order to expand the frontier of knowledge in the national design and maintenance engineerings. For this, two case studies are made of: modelling flexible permanent way; modelling AMVs and crossings; passing vehicle in a region with gauge opening on a constant radius curve; passing a vehicle sequence of reverse curves. Thus, this dissertation presents an overview of the state of the art in the simulation of the interaction vehicle and permanent way of metro.

Le trafic routier devenant de plus en plus dense, la sécurité active des véhicules doit être sans cesse améliorée. Ceci nécessite notamment une amélioration de la dynamique des véhicules. La majeure partie des efforts transmis par la chaussée au véhicule passant par les pneumatiques, une connaissance précise de l’interface pneu-sol est primordiale. Ceci passe notamment par le développement de modèles de pneumatiques et leur identification, à partir de mesures spécifiques, lors des phases de conception et de mise au point d’un véhicule. Le projet de thèse a pour objectif principal de contribuer à une meilleure détermination du point de fonctionnement dynamique de l’ensemble pneumatique/roue. Plus précisément, il s’agit d’améliorer la détermination de l’angle de dérive du pneumatique, dont la précision est encore aujourd’hui insuffisante pour une analyse des performances des pneumatiques à partir d’essais sur piste. Cette détermination est réalisée simultanément à la détermination des autres grandeurs nécessaires à l’étude de l’interface roue-sol : glissement longitudinal du pneumatique, angle de carrossage de la roue et torseur des efforts roue-sol. Ce projet s’est appuyé sur la collaboration entre la Direction de la Recherche de Renault et le laboratoire MIPS de l’Université de Haute-Alsace. Des études préliminaires menées par Renault et par le MIPS sur ce sujet ont servi de support à ce projet.Une étude menée en début de thèse a permis de rédiger un cahier des charges exhaustif, en termes de performances de mesure de l’angle de dérive du pneumatique, permettant d’étudier avec suffisamment de précision la dynamique du véhicule, en se basant sur indicateurs de performance. Afin de répondre au cahier des charges, une solution de mesure novatrice de l’angle de dérive du pneu a été conçue et développée au cours de la thèse. Celle-ci se base sur l’intégration de mesures inertielles et GPS au niveau du plan de jante, à l’aide d’un filtre de Kalman étendu (EKF). Par la suite, un capteur de mesure (INS/GPS) dédié a été conçu, développé et enfin intégré à un véhicule d’essais. Les performances de mesure du nouveau capteur INS/GPS ont été validées vis-à-vis du cahier des charges, en suivant une procédure de caractérisation mise au point spécifiquement durant la thèse. L’angle de dérive pneu est ainsi mesuré avec une précision de 0,1° (à 3o) sur une plage de 360°, tout en étant insensible à la texture et à la granularité de la chaussée ; contrairement aux solutions industrielles existantes. Lorsque ce moyen de mesure novateur est couplé à d’autres mesures standard de la dynamique du véhicule, le mouvement complet du plan de jante est alors déterminé également avec la précision requise. Ceci en ajoutant uniquement un capteur au niveau du plan de jante, ce qui permet de limiter l’inertie ajoutée à la roue, en comparaison aux moyens de mesures existants.Pour valider le nouveau moyen de mesure développé et montrer tout son intérêt, une méthodologie d’identification d’un modèle de pneumatique (Magic Formula 5.2) a été proposée et validée expérimentalement. Un algorithme d’optimisation Simplex a été développé afin de pouvoir identifier un modèle de pneu, aussi bien à partir de mesures (sur piste ou sur banc de test) qu’à partir de données de simulation. Ainsi, les paramètres du modèle de pneumatique peuvent être optimisés durant la phase de conception d’un véhicule, afin de répondre à des indicateurs de performance du véhicule bien définis. Aujourd’hui, les mesures précises du capteur INS/GPS, conjuguées à l’algorithme d’optimisation développé, sont notamment utilisées pour identifier des modèles de pneu à partir de mesure sur piste, et pour définir les paramètres du modèle de pneu nécessaires, en phase de conception de véhicule. […] / The road traffic becomes more and more dense so that the active safety of road vehicles must continuously improved. This especially requires the improvement of the vehicle dynamics. The major part of the forces applied by the road to the vehicle are issued of the tyres. So, an accurate knowledge of the tyre-road interaction is primordial. This especially requires the development of the modelling and the identification of tyre models in the phases of design and tuning of a vehicle. The main goal of the thesis project is the contribution to a better determination of the dynamic functioning point of the system tyre/wheel. More accurately, is aims at improving the tyre slip angle determination accuracy, which is still insufficient for the analysis of tyre performance from vehicle-on-track tests. This determination is performed simultaneously to the determination of the other variables required in study of the tyre-road interface: longitudinal slip ratio, camber angle, tyre forces and moments. This project has been made in collaboration with the Research Department of Renault and the MIPS laboratory of the University of Haute-Alsace. Preliminary studies carried out by Renault and MIPS on this topic have served as a support to this project. A study carried out at the beginning of the thesis has helped to the determination of exhaustive requirements, in terms of tyre slip angle measurement performance, allowing the study of the vehicle dynamics with a sufficient accuracy, using performance indicators. A novel tyre slip angle measurement means has been developed in the thesis, in order to fit these requirements. It is based on the combination of inertial and GPS data near the wheel rim edge plane, using a extended Kalman filter (EKF). A dedicated measurement sensor (INS/GPS) has then be designed, developed and embedded in a test vehicle. The measurement performance of the novel INS/GPS has been validated according to the requirements, following a characterization procedure which has been specifically designed in the thesis. The tyre slip angle is then measured with an accuracy of 0.1° (at 3σ) in a range of 360°, with a insensitivity to the texture and the granularity of the road, contrary to the existing industrial solutions. When this novel measurement means is coupled with other standard measurements in vehicle dynamics, the whole motion of the wheel rim is then also determined with the required accuracy. This is made possible by mounting only one sensor at the level of the wheel rim edge plan, which helps to limit the inertia added to the wheel, in comparison with the existing measuring systems.A strategy of identification of a tyre model (Magic Formula 5.2) has been proposed and validated experimentally, in order to validate the novel measurement means developed. A Simplex optimization algorithm has been developed for the identification of a tyre model, from measurements (on a track or on a test rig) and/or simulation data. So, the tyre model parameters can be optimized in the phase of vehicle design, in order to fit well know vehicle performance indicators. The accurate INS/GPS measurements, combined with the optimization algorithm, are now especially used for the identification of tyre models from vehicle-on-track measurements and for the determination of the tyre model parameters required in the design phase. The novel INS/GPS is also considered as a reference for the tyre slip angle measurement, whatever the road grip (dry, wet, snow covered, ice covered ...). A number of research collaborations, especially with a worldwide tyre manufacturer, are been launched following the thesis results, especially for the study of the correlation between the physical tyre properties and the vehicle dynamics.

Le véhicule est un système dynamique complexe, composé de différents sous-systèmes de nature différente (moteur, système de freinage, suspension ...). Chaque sous-système est décrit par un modèle mathématique dépendant d’un nombre important de paramètres, très souvent incertains (méconnaissance, manque de mesures,…). L’incertitude sur les paramètres se propage à travers le modèle et se retrouve sur la sortie. Cette dernière représente les forces et moments mis en jeu dans le véhicule. L’incertitude sur la sortie n’est pas toujours tolérable pour des raisons de sécurité, précision,…Situé dans ce contexte, les travaux de la thèse consistent à proposer des méthodes d’analyse de sensibilité permettant de déterminer les paramètres dont les incertitudes ont un effet significatif sur le comportement d’un système donné, en particulier le véhicule. Dans une première partie, le cas des modèles à paramètres dépendants et suivant une distribution arbitraire est étudié. Une méthode, basée sur la décorrélation des paramètres par la décomposition de Cholesky, a été proposée. Pour résoudre le problème de la distribution arbitraire, l’approximation par polynôme du chaos arbitraire est adoptée, en construisant une base orthonormale en termes de moments statistiques non centrés des paramètres. Les indices de sensibilité, permettant de quantifier la contribution de chaque paramètre à la variance de la sortie, sont obtenus directement à partir des coefficients des polynômes du chaos ainsi obtenus. La méthode proposée est appliquée et validée sur un modèle de pneumatique. Dans la deuxième partie, le cas des modèles dynamiques est traité. Une méthode basée sur les dérivées partielles est explorée, puis une approche alternative est proposée. Elle utilise de façon originale des outils de l’Automatique, les grammiens d’atteignabilité et d’observabilité. L’influence des paramètres sur l’énergie consommée en entrée et restituée par le système en sortie est ainsi déterminée. L’avantage de cette technique est que les paramètres peuvent être classifiés selon leurs influences sur l’énergie consommée ou restituée, tout au long de la dynamique du système. D’autre part, l’étude de la sensibilité des paramètres sur les échanges de l’énergie, permet de déterminer un placement optimal des paramètres pour une optimisation de l’énergie consommée et restituée en sortie. Les deux méthodes proposées sont appliquées et validées sur un modèle bicyclette décrivant le comportement dynamique d’un véhicule. Dans la dernière partie, des tests sur véhicule d’essais ont été réalisés sur circuit. Les différentes approches d’analyse de sensibilité ont été appliquées sur les résultats d’essais, afin de recaler des modèles de pneumatique. / The vehicle is a complex dynamic system, composed of various subsystems of different kind (engine, braking system, suspension, etc.). Each subsystem is described by a mathematical model depending on a significant number of parameters, very often uncertain (unknown, lack of measures, etc.). The uncertainty on the parameters is propagated through the model and takes place on the model output. The model output represents the forces and moments involved in the vehicle. The uncertainty on the model output is not always tolerable for safety reasons, precision, etc. In this context, the aim of the thesis is to propose sensitivity analysis methods allowing to determine parameters whose uncertainties have a significant effect on the behavior of a given system. In the first part, the case of models with dependent parameters which follow an arbitrary distribution is studied. A method based on the decorrelation of the parameters using the decomposition of Cholesky, is proposed. To solve the problem of the arbitrary distribution, an approximation using arbitrary polynomial chaos is adopted and an orthonormal data basis is constructed in terms of non central statistical moments of parameters. Sensitivity indices, allowing to quantify the contribution of every parameter to the model output variance, is directly obtained from the polynomial chaos coefficients.The proposed method is applied and validated on a tyre model. In the second part, the case of the dynamic models is studied. A method based on partial derivative is explored. Then a new alternative approach is proposed. This method uses in an original way the control theory tools, the reachability and observability Gramians. The influence of the parameters is formulated in terms the energy consumed and restored by the system. The advantage of this technique is that the parameters can be classified according to their influences on the consumed or restored energy throughout the system dynamics. On the other hand, the study of the parameters sensitivity based on ratio energy exchanged, allows to determine an optimal placement of the parameters for an optimization of consumed and/or restored energy. Both proposed methods are applied and validated using bicycle model describing vehicle dynamic behavior. Finally, the various sensitivity approaches are applied to adjust tyre model parameters using vehicle measurements acquired during a steady-state maneuver.

Pervasive applications of the vehicle simulation technology are a powerful motivation for the development of modern automobile industry. As basic parameters of road vehicle, vehicle dynamic parameters can significantly influence the ride comfort and dynamics of vehicle, and therefore have to be calculated accurately to obtain reliable vehicle simulation results. Aiming to develop a general solution, which is applicable to diverse test rigs with different mechanisms, a novel model-based parameter identification approach using optimized excitation trajectory is proposed in this paper to identify the vehicle dynamic parameters precisely and efficiently. The proposed approach is first verified against a virtual test rig using a universal mechanism. The simulation verification consists of four sections: (a) kinematic analysis, including the analysis of forward/inverse kinematic and singularity architecture; (b) dynamic modeling, in which three kinds of dynamic modeling method are used to derive the dynamic models for parameter identification; (c) trajectory optimization, which aims to search for the optimal trajectory to minimize the sensitivity of parameter identification to measurement noise; and (d) multibody simulation, by which vehicle dynamic parameters are identified based on the virtual test rig in the simulation environment. In addition to the simulation verification, the proposed parameter identification approach is applied to the real test rig (vehicle inertia measuring machine) in laboratory subsequently. Despite the mechanism difference between the virtual test rig and vehicle inertia measuring machine, this approach has shown an excellent portability. The experimental results indicate that the proposed parameter identification approach can effectively identify the vehicle dynamic parameters without a high requirement of movement accuracy.