Development of Madymo-based Model for Simulation of Laboratory Rollover Test Modes

Types of vehicle rollovers can be classified into two categories: untripped and tripped. Untripped rollovers are relatively rare events resulting from high lateral friction forces between the tires and road. Tripped rollovers are the result of lateral forces caused by the tire or wheel digging into the road or ground or from striking a curb or other obstacles. As reported in the open literature, various test methods for conducting rollover events such as SAE J2114, Side Curb Trip, Critical Sliding Velocity, and Corkscrew have been used. This paper presents the development of MADYMO-based models for simulating vehicle kinematics in these four modes. The CAE methodologies using MADYMO is interactively developed with the test methodologies. Experimental data obtained from these test modes are used for developing rollover CAE models for replicating vehicle motions under similar test conditions. Analyses of simulated results provide feedback to improve the test procedures. Testing with improved procedures provide additional new data for continued model refinements. MADYMObased CAE tools thus provide quality models with better simulated and/or predicted results. MADYMO rollover models consist of sprung and un-sprung masses, suspension systems and tires, whose characteristics are extracted from ADAMS-based vehicle handling model. Use of the MADYMO-based models to support rollover testing, rollover sensing algorithm development, and rollover protection system development will be described. Since MADYMO modeling described in this paper is a rigid-body based approach, model limitations and issues associated with rollover simulation will also be discussed. In addition, model correlations with test data in these four modes and future areas of improvement will be presented. INTRODUCTION For many years, NHTSA has conducted research investigating the underlying causes of vehicle rollover accidents, developing rollover test procedure, and developing vehicle and roadway design criteria to help reduce both the number and the severity of rollover accidents. The rollover process, which involves a complex interaction of forces from suspension systems, tires, power-trains, and road surface, is one of the most complicated types of safety analysis. To study the vehicle and/or occupant kinematics during rollover crashes, mathematical models are useful tools for understanding essential rollover mechanics and evaluation of restraint system performance in mitigating occupant ejection. Tools available for such analysis include vehicle dynamic handling models, occupant gross-motion simulators [1-3] and finite element (FE) analysis programs. A BRIEF REVIEW OF ROLLOVER MODELS Rollover models are basically mathematical analyses which describe equations of motion derived for a simplified vehicle system consisting of one rigid body or two/three rigid bodies connected by joints and springs. Models are specifically developed for studying rollover mechanics under specific conditions. Jones [4] used a simple one-degree-offreedom model to study the mechanics of vehicle rollover as a result of curb impact by treating the contact force at the curb as impulse forces in determining the vehicle kinematics. Ford and Thompson [5] developed a two-dimensional model as an initial attempt to predict the rollover characteristics of a vehicle. Their model is basically a 2D rigid-body of an automobile to allow simulation of vehicle ground contact and airborne motion. Lund and Bernard [6] developed a one-rigid-body model for analysis of simple rollovers to study the mechanics of the tilt table test and critical sliding velocity. Rollover simulation using a nonlinear model was reported by Eger et al. [7], using two rigid bodies