Longitudinal Train Dynamics Model for CC203/CC206 Locomotive Simulator

Abstract

This paper presents train modeling used in a simulator platform for driver training. It was developed for the CC203/CC204 locomotive. The driver will gain experience as in a real locomotive from the perceived platform movements if the movements match real conditions as accurately as possible, including the distance travelled. To this aim, a longitudinal model of the train was developed based on measurement data obtained from the Argo Parahyangan train traveling from Bandung to Jakarta. A second-order linear time invariant model was obtained by a black box identification approach, in which the input and the output of the model are the resultant force (a traction and a slope-friction force) and the train’s position, respectively. While the speed is directly obtained from measurement data, the traction force of the locomotive is predicted using the traction characteristic of the locomotive, train’s measured speed, and latitude time history during a train trip. The model is then validated by running a simulation for one complete trip of the train. In the simulation, the same input as in the model identification is applied and the mileage obtained from simulation result is compared to data of the real train trip with a fitness level of 94.09%.