2012, Article in monograph or in proceedings ()During the previous HVTT conference in Melbourne, results have been presented of a project FORWARD, including eleven different trailer manufacturers, involving testing and modeling activities to establish a practical basis for realistic fatigue assessment as a step towards an improved light-weight trailer design. The present paper describes the next step as part of the follow-up project LIFE (LIfetime Fatique Enhancement), starting with the model-based assessment of a realistic loading history for a semi-trailer. Representative loading conditions in this history, being derived from three weeks of field testing, are then used in Finite Element analyses to estimate fatigue life at a critical high loaded welded connection in the aluminum
trailer chassis. Results are discussed with respect to the order of different loading cycles (Palmgren-Miner number, based on fatigue tests), 3D loading at the weld, the varying weld quality, and the impact of bonded connections in the trailer. These ‘lessons learned’ are interpreted, to result in improved general guidelines in light-weight trailer design.
2012, Article / Letter to editor (Journal Vehicle System Dynamics, vol. 50, iss. 6, (2012), pp. 939-959)Path tracking driver models assume the observed path deviation ahead of the vehicle to be proportionally transferred to a corrective steering input. The most simple version of such a model includes three parameters, a single preview distance, a steering gain and a delay time, being examined in this paper, in dependency of vehicle properties, driver characteristics, velocity and path. It is shown that, for different and bounded preview lengths, a driver can follow any path with almost the same minimum path error, if the gain is adapted appropriately. The upper boundary is path-dependent but driver and path characteristics have only a minor effect on the resulting relationship between preview length and gain. Consequently, gain and preview length may well vary along some path. This has been examined, experimentally, for different drivers. A too small preview length conflicts with closed-loop stability, explicitly described in terms of vehicle parameters, vehicle speed and driver delay time. The results of this paper provide a basis for enhanced understanding of human driver behaviour.