posted on 2023-05-26, 23:42authored byBecker, Steffen
Electrolytic hydrogen production can be seen as the binding element in utilising and storing renewable energies towards a sustainable and environmentally compatible energy supply. In this thesis comprehensive literature review on hydrogen production with emphasises on electrolysis of water and various conventional models has been conducted. Furthermore, literature survey on applied Artificial Neural Networks (ANN) and Adaptive Neuro-Fuzzy Inference Systems (ANFIS) elucidates architecture, functionality and application in detail. This study provided predictive hydrogen production performance models for a commercial PEM-electrolyzer. Two different approaches using intelligent techniques have been conducted. The first employs ANN and the second uses a hybrid model ANFIS as time series prediction combining fuzzy logic and Neural Networks. An experimental apparatus has been developed to measure and model specific performance parameters such as hydrogen flow rate, system-efficiency and stack-efficiency. A comprehensive range of experimental conditions were tested as part of the investigation that covers a wide range of input variables and their influence on the output performance. The various parameters have been obtained using the electrolyzers' internal software (windows diagnostic) and additional sensors measuring power and feed water parameters, such as water quality, water pressure, system temperature, stack current, stack voltage, system power consumption, system pressure, product pressure and lower explosive limit. Synchronous data-acquisition of all parameters was carried out with National Instruments LabVIEW software to build a database. The database formed the foundation for the predictive models, where experimental data were used to train and test the developed hydrogen production performance models. Verification of those models was carried out by comparison of predicted and measured data. It is argued that, due to the high costs associated with the hydrogen measuring equipment; these reliable predictive models can be implemented as virtual sensors.
History
Publication status
Unpublished
Rights statement
Copyright 2010 the Author - The University is continuing to endeavour to trace the copyright owner(s) and in the meantime this item has been reproduced here in good faith. We would be pleased to hear from the copyright owner(s). Thesis is available for downloading from 23/4/20. This thesis contains confidential information and was previously not to be disclosed or made available for loan or copy without the express permission of the University of Tasmania. Once released the thesis became available for loan and limited copying in accordance with the Copyright Act 1968. Authority from School of Engineering on 10th February 2012. Thesis (PhD)--University of Tasmania.