Artificial neural network control of hybrid fuel cell/ultra-capacitor vehicular power system

The effect of vehicles has an important role in global gas emission. Besides, the internal combustion engines of current vehicle systems have poor efficiency ratios particularly in rated rpm. Alternatively, the usage of electric motors has drawn attraction due to the environmental factors and high efficiency in lower rpms. In near future, fuel cell technologies are expected to be a sustainable, alternative and environmental friendly solution for especially vehicular applications in feeding the electric motors. However, using only a fuel cell may not be sufficient for conditions such as start-up, transient events and instantaneous peak power demands (rapid acceleration or hard breaking) which the power source may face in urban traffic conditions. In transient conditions, insufficient humidification with gas starvation or flooding problems may occur due to the instant load changes. The main reason for this problem is the gas (oxygen and hydrogen) flow ratios are relatively slow when compared to a transient load increment or decrement. Gas starvation has a major impact in safe fuel cell operation and performance. Thus, operating the fuel cell in steady state loading conditions and preventing it from subjecting to the rapid load changes will keep secure operation and increase its lifetime and decrease its hydrogen consumption. Also, commercially available fuel cells are not reversible and they do not have capability of capturing the braking energy. Ultra-capacitors, a recently popular energy storage system, have the capability of transferring and capturing high amounts of energy very fast. Considering these issues, utilizing fuel cell and ultra-capacitor systems together will meet all the load demands of a vehicle. Such a hybrid system will not only reduce the size and cost of the fuel cell power technology but also provide a better potential of energy efficiency.In this thesis, the energy management of a hybrid fuel cell/ultra-capacitor vehicular system is realized utilizing a control algorithm based on artificial neural networks. With the usage of the performed control algorithm, providing a steady state operating condition for fuel cell is realized. Besides, controlling the state-of-charge of the ultra-capacitor and promoting the vehicle performance and energy savings is aimed.