Avhandling: Electrified engine air intake system: Modeling, optimization and control

torsdag, februari 4, 2016

Sava Marinkov försvarade nyligen sin avhandling Electrified engine air intake system: Modeling, optimization and control vid TU Eindhoven. Avhandlingen finns att ladda ner här >>

Summary

Despite recent advances in electric and fuel cell vehicle technology, the internal combustion engine is still seen as a key facilitator of ground transportation for the next decade. This is mainly due to superior energy density and storage properties of carbon-based liquid fuels compared to, e.g., electric batteries.  However, increasing societal concerns for natural resource depletion and environmental pollution put ever-tightening constraints on the vehicle fuel economy. This thesis addresses the improvement of the vehicle fuel economy using two novel engine technologies: electric supercharging and regenerative throttling.  The research conducted within this domain resulted in the following main developments.

The first new development is a theoretical investigation of a downsized engine concept, where short-duration, high-power demands are delivered by means of a standalone electric supercharger. In this context, the supercharger consists of a compressor and a high-speed electric motor, which is powered from a car battery. This research presents a novel, convex method for the battery and engine sizing such that they match the supercharger energy requirements and power-enhancing capabilities.  A simulation-based case study is also provided, showing that, over a specic driving cycle, the investigated powertrain conguration can yield up to 10% savings in fuel costs, w.r.t. a naturally-aspirated engine powertrain scenario.

The second new development concerns a theoretical investigation of regenerative throttling for gasoline engines.  Namely, by replacing a throttle valve with a high-speed generator-turbine throttle unit (GTU), the engine intake airflow can be controlled while simultaneously producing electricity. In this development a new computational method is proposed for studying the effect of such a device on the vehicle fuel consumption. The analysis has shown that regenerative throttling has a potential to deliver 2-4% fuel cost savings compared to a conventional throttle valve situation.

The third new development relates to Switched Reluctance Machine (SRM) control. The SRM is an electric machine free of brushes, rotor windings and permanent magnets. Its simple, low-cost design and high-speed capability make it suitable for both electric supercharging and regenerative throttling applications. However, the same design also imposes a considerable challenge for control, as it results in inherently nonlinear switched system dynamics.  This issue has been addressed by means of three novel control strategies. The first introduces a four-quadrant speed tracking controller for 4-stator/2- rotor pole SRMs.  The second provides an Explicit Model Predictive Controller for the SRM output voltage tracking. The third, however, explores the possibility of the SRM speed control using only the DC-link voltage and current measurements, i.e., without speed, position, and phase voltage and current sensors. The effectiveness of all presented control algorithms has been veried in simulations, whereas the first has also been validated experimentally.

Finally, the fourth new development is a method for the GTU auto-calibration.  To maximize the GTU fuel-saving potential the turbine speed needs to be suitably matched to the conditions present in the engine air-intake system, at all times. However, often the exact optimal speed value, which yields maximal energy recovery, is unknown or dicult to derive. This research proposes a non-model-based solution to the problem of finding the optimal turbine rotational speed.  The algorithm is based on a novel Extremum Seeking Control (ESC) law, with a disturbance-based optimal input parametrization. The proposed method allows adaptive reconstruction of the unknown relationship between the measured disturbance signals, i.e., the turbine pressure ratio and its vanes position, and the optimal turbine speed – in an initial, automated calibration step. The usefulness of the presented auto-calibration scheme, however, extends beyond this particular application.  It can be applied to any similar ESC situation, where the disturbances leading to the changes in the optimal inputs can be measured or estimated.