Artificial Neural Networks based thermodynamic and economic analysis of a hydrogen production system assisted by geothermal energy on Field Programmable Gate Array

In this study, the thermodynamic and economic analysis of a geothermal energy assisted hydrogen production system was performed using real-time Artificial Neural Networks on Field Programmable Gate Array. During the modeling of the system in the computer environment, a liquid geothermal resource with a temperature of 200 °C and a flow rate of 100 kg/s was used for electricity generation, and this electricity was used as a work input in the electrolysis unit to split off water into the hydrogen and oxygen. In the designed system, the net work produced from the geothermal power cycle, the overall exergy efficiency of the system, the unit cost of the produced hydrogen and the simple payback period of the system were calculated as 7978 kW, 38.37%, 1.088 $/kg H₂ and 4.074 years, respectively. In the second stage of the study, Feed-Forward Artificial Neural Networks model with a single hidden layer was used for modeling the system. The activation functions of the hidden layer and output layer were Tangent Sigmoid and Linear functions, respectively. The system was implemented on Field Programmable Gate Array using the Matlab-based model of the system as a reference. The maximum operating frequency and chip statistics of the designed unit of Field Programmable Gate Array based geothermal energy assisted hydrogen production system were presented. The result can be used to gain better knowledge and optimize hydrogen production systems.