EuropeLeonardo Da VinciEducation Culture
Partners
Alfano Primo DIMEC E.R.F.A.P. ElettroSannio Budapest University of Tecnology and Economics Universite de Haute Alsace Dunarea de Jos  University Galati Istanbul Teknik Universitesi PST Auto Consulting DIIIE
Graphics by Marco Coraggio
Timeline
Progetto finanziato con il contributo della Comunità Europea
Untitled Document
Link in rosso: Cerca su Wikipedia
Untitled Document

Control system (for HEV and HSV)

J.Bokor, P.Gaspar, P.Bauer, T.Peter, T. Becsi (*)
(*) BUTE, Budapest

Model based control

It becomes important to control complicated hybrid systems that consist of not only a powertrain but also vehicle systems such as regenerative braking. Model-based control and calibration enables both control strategy optimization and control system development efficiency improvement. Recently, the hybrid vehicle, which requires the integration of various control systems, has been introduced to the market. Model based control offers the potential to derive an accurate controller rapidly and systematically. It is especially useful for developing a coordinated control system like those present in a hybrid vehicle.

Model based control is characterized by a plant model that describes the dynamics of the system to be controlled and allows closed loop simulation. A high fidelity powertrain model can make it easy to introduce advanced control technologies including recent developments in robust, adaptive and nonlinear control. Advanced, high-speed data processing will provide sophisticated filtering technologies to extract important information from measured signals. However, fundamental questions remain for the current control theories. For an example, control designs are often subject to unrealistic restrictions such as the need to neglect time delays and manipulation constraints, the linearization of plants and so on.

For controller design on a given hybrid vehicle, we need to know the physical layout of the powertrain. Basically it can have series, parallel or series/parallel structure. We need to know the whole structure containing internal combustion engine (ICE), electrical motor / generator (EM), battery and perhaps solar cells too. We must have, or construct the characteristics (rpm, torque, efficiency, etc.) and models (MATLAB / SIMULINK) of all elements.

In case of parallel structure the device which joins the ICE and electrical motor EM is the most interesting component of the powertrain. It affects seriously the optimal energy management of the vehicle. It could allow arbitrary rpm rate between ICE and EM, or only a fixed rate (as a timing belt). These two cases are very different. In the first case we can change both the torques and rpms of the motors to achieve optimum performance split, while in the second case we can change only the torques of the motors (if cluch is closed).

In literature we can find several control strategies for hybrid electric vehicles. The global, optimal control strategy could be calculated only with dynamic programming or other backward algorithms. The main problem is the lack of knowledge about the whole drive cycle, which could last hundreds or thousands of seconds. So in real applications only a suboptimal solution can be found which still requires the use of drive cycle prediction. For this purpose, several prediction methods can be used for example Neural Networks [2] or Model Predictive Control.

Another problem is how to create the cost function which reflects the fuel equivalent of battery power besides the ICE fuel consumption. In some articles [1] equivalence factors which are control parameters are used in the cost function. So we have to optimize the drive cycle dependent equivalence factors too. The searching for cost functions, without additional factors, can be a second goal of controller design (besides minimal fuel consumption in global sense). For controller tests and comparison we need to know the driving schedules (FUDS = Federal Urban Driving Schedule, FHDS = Federal Highway Driving Schedule etc. [1,2] ) used by other authors.

References

[1] Cristian Musardo, Giorgio Rizzoni, Yann Guezennec és Benedetto Staccia: A – ECMS: An Adaptive Algorithm for Hybrid Electric Vehicle Energy Management, European Journal of Control, 2005, pp. 509-524

[2] Ivan Arsie, Marco Graziosi, Cesare Pianese, Gianfranco Rizzo, Marco Sorrentino: Optimization of Supervisory Control Strategy for Parallel Hybrid Vehicle with Provisional Load Estimate, AVEC ’04 (Department of Mechanical Engineering – University of Salerno)

Inizio Pagina - Modulo Precedente - Modulo Successivo