Eddy Current Brake Control for Test Cycles Simulation

On-vehicle (rolls dynamometer or road) tests are usually more expensive and time-consuming than test bench ones. Furthermore, sometimes results would be useful during vehicles design phase. The paper aim is to present a methodology that allows simulating the vehicle on an engine test cell, by properly controlling the bench actuators. Engine operating conditions mainly depend on speed and load, which are determined by the vehicle driving conditions: the speed-time trend assigned for the vehicle must be converted into equivalent speed-time and load-time trends for the engine, and used for feedback control of brake and accelerator actuators. To evaluate the engine load torque it is necessary to know vehicle characteristics (mass, gear ratios, wheels radius, drag coefficient, frontal area, etc.) and driving conditions: the real vehicle can thus be substituted with a virtual vehicle. The methodology has been applied to simulate an ECE-EUDC driving cycle, which is usually carried out on the rolls dynamometer, as imposed by regulations. During such test the vehicle has to follow an assigned speed-time trajectory, while road load and vehicle inertia are simulated and calibrated using a standard procedure. The test is subject to human error, since the driver does not follow exactly the theoretical speed trend, while using robot-drivers increases the setup cost. The same test has been reproduced on a standard engine bench. This setup would be useful to tune the engine correctly and to study the effects of vehicle characteristics variation, thus allowing to determine the correct strategy for emissions reduction, or to estimate the vehicle emission performance, before it is available for chassis dynamometer tests. The same system could be used for real time implementation of control strategies involving both the vehicle and the engine, such as traction control algorithms. Furthermore driving conditions simulations, executed by electronically controlling engine speed and load trajectories, would be more repeatable than human driving on the chassis dynamometer, and their cost would be substantially smaller. The paper shows how the vehicle speed trend can be converted into engine speed and load trends with a physical system model, and then used to control the bench using a real time control system, thus performing a vehicle driving cycle simulation.