Wirkungsgradoptimierung eines hybriden Einzylindergenerators mittels Optimalsteuerung und Extremum-Seeking-Control

Description

The focus of this thesis is on efficiency maximization of hybrid power units that consist of an internal combustion engine which is mechanically coupled to an electric generator. Systems like that are widely used as isolated power plants or within hybrid-electric propulsion systems. In the dimensioning as well as the operation of such systems, there exist a lot of different design and operating variables that affect the machines’ efficiency and make the optimal design a challenging task. Besides, the hybrid setup generates new dependencies within the energy system that can be used for efficiency optimization. One of those variables that is thoroughly analyzed in this dissertation is the torque trajectory of the electric generator: since the torque influences the dynamic motion profile of the engine piston there is an immediate impact of the operating strategy of the electric generator on the thermodynamic processes. The main contribution of the dissertation is the development of a concept that allows a systematic determination of the most efficient operating strategy under consideration of the various design and operating variables, particularly the dynamic torque trajectory. The study is conducted using the example of a single cylinder compression-ignition internal combustion engine and a permanent magnet synchronous generator.

The proposed concept consists of two major steps. In the first part of the thesis the question of the best design and the best operating strategy in the sense of optimal efficiency is analyzed in a model-based framework. For this investigation a mathematical description of the hybrid power system is developed and methods from optimal control theory are applied to solve the underlying optimization problem. One important result of this study is that the degrees of freedom associated with the torque trajectory of the electric generator should mainly be chosen so that current dependent losses in the electrical subsystem are reduced. Besides, the contribution contains recommendations for an efficient design of the hybrid system and a discussion of the robustness of the proposed operating strategy.

The second part focuses on the experimental validation of the previously calculated strategy. For this purpose, an extremum-seeking based real-time optimization concept that maximizes the efficiency and simultaneously regulates the electrical output power on a constant level is developed. The local stability of the presented feedback loop is proven by using singular perturbation as well as averaging theory and is also analyzed by simulations and experimental studies. Overall, the investigations on the test bench confirm the results obtained from the model-based optimization but also show the benefit of the model-free approach.