Reduction of the premature aging of PEM Fuel Cell-battery systems by implementing an advanced hybrid control system based on double PID, neural network load predicting algorithm and optimal battery sizing algorithm

Proton Exchange Membrane Fuel Cell (PEM-FC) is a technology that is associated with three main challenges that can be summarised as cost, reliability and lifetime. Current global research in PEM-FC focuses on reducing the use of expensive precious materials to reduce costs. Research is also improving the way fuel cells internal subsystems (water management, thermal management, power conditioning unit, gas supply management) are being controlled, managed and operated to enhance reliability and lifetime. This thesis takes the FC research to a different level of abstraction to reduce FC internal degradation. The proposed work does not focus on current FC global research trends. Instead, it addresses the different fuel cell challenges by proposing new solutions to the external factors (dynamic of the load, battery size, etc.) affecting a FC during operation. The formulated thesis hypothesis is therefore: “ Reduce the FC internal degradation and address the above three FC challenges by looking at the overall FC system (FC, battery, DC/DC converter, load) and ignoring the FC internal subsystems and subcomponents (water, thermal management, power conditioning, gas supply management). This research differs from previous work as it does consider a FC as a black box and only work on how to better control external components that have a negative impact on the fuel cell durability. ” This research work proposed three main interrelated “antidotes” to address the PEM-FC premature failure, thereby reducing cost, reliability and lifetime issues: a) Development of an optimisation technique to size the intermediate battery buffer component within a fuel cell system. b) Devise a method to optimise the fuel cell electrical power output using the proposed double PID control system. c) Devise a method to optimise the fuel cells operation and reduce battery stresses using advanced predictive Artificial Intelligence (AI) technology. The outcome from this research work demonstrates how it is possible to reduce the fuel cell and battery degradation while feeding load requirements.