An approach for the online optimized control of a hybrid powertrain

The basic objective for the control of a hybrid electric vehicle is the powertrain's operation with minimum energy consumption. At a parallel hybrid, the fundamental problem is the optimal split of the torques of combustion engine and electric motor at an optimum engine speed. It has to be solved "online" at every point of time during the operation. At actual hybrid vehicles, this function is realized by an operation strategy, which is based on a static specification, developed on empirical considerations with "optimized" strategy parameters. This kind of control works fast and reliably, but often produces results far away from all optimal control. To meet higher requirements, the paper discusses direct use of optimization methods to solve this problem. In the past, fundamental research on the Autark Hybrid's powertrain control had been done by computing the optimum control vectors for given driving cycles by a special offline optimization method. The objective was the overall fuel consumption at the end of a cycle. For online optimization, however, it is necessary to specify an objective which relates to each point of time and which defines the problem of latching electrical energy. The paper presents a new approach with an objective based on the minimization of the powertrains amount of power dissipation.