Condition monitoring of polymer electrolyte membrane fuel cells
Doctoral Thesis
2014
Permanent link to this Item
Authors
Journal Title
Link to Journal
Journal ISSN
Volume Title
Publisher
Publisher
University of Cape Town
Department
License
Series
Abstract
As the global demand for energy continues to grow new technologies and systems must be developed to supply the market. This includes renewable energy generation, storage and conversion systems. The primary storage technology in use today in the portable electronics, the automotive sector and to a lesser extent power networks is battery based systems. To overcome some of the limitations inherent in batteries, fuel cell based power generators and converters have been developed. Fuel cells act as electrochemical energy converters that convert a fuel source such as natural gas directly into electrical power without any secondary phases. For systems running on Hydrogen generated via renewable or natural sources, the input/output cycle becomes completely sustainable. Out of the different fuel cell types available and under development, the Proton Exchange Membrane or Polymer Electrolyte Membrane (PEM) fuel cell has emerged as the technology of choice, and currently owns more than 80% of the commercial fuel cell market. This has spurred further research in the field to increase performance and life expectancy of the cell materials. A promising development in the form of High Temperature PEM (HT-PEM) fuel cells has recently emerged and addresses some of the shortcomings of the low temperature counterparts. A critical field of research is the condition monitoring strategies and technologies for the electrochemical device that ties in with the power conditioning sub-systems. This thesis presents the development of condition monitoring systems by conducting detailed studies on the fault/degradation mechanisms prevalent in the cell materials for the purpose of detection, classification and implementation of possible mitigation strategies. Specific consideration is given to the detailed analysis of the fault mechanisms in HT-PEM fuel cells that are not yet fully understood and commercialized. In particular, electrochemical equivalent circuit models and reduced order semi- empirical models are developed to facilitate fault detection. Based on these models, mitigation strategies for specific faults are proposed and experimentally verified. New systems and methods are developed for rapid online impedance signature mapping that provide a basis for early fault prediction that can increase system performance and life expectancy. The findings in this research provide valuable insight into the effect that most prevalent faults have on the internal electrochemistry and the impact on electrical performance. From the experimental results, a semi-empirical electrochemical model is developed to assist with life time estimation and system optimization. The model is integrated with a real time emulator platform that can reproduce single cell voltage levels at the high output currents and transient characteristics. A detailed analysis is conducted on CO poisoning and the resulting effects on key equivalent circuit parameters that enable quantification of the fault condition. It is shown that the catalyst at the higher operating temperature is still susceptible to a certain degree of semi-permanent degradation. To mitigate these effects, a new active current control strategy is proposed to enforce electro-oxidation of the CO to recover the lost active area that delivered superior results compared to current pulsing strategies. New rapid online detection strategies are proposed by using small voltage transients in an operational HT-PEM fuel cell. The method makes use of the discrete S-transform that overcomes some of the limits in other signal processing methods used in fuel cell diagnostics. To enable detailed parameter calculation, a population based incremental learning algorithm is implemented in the developed method. A new condition monitoring system is developed that makes use of Optimized Broadband Impedance Spectroscopy. The hardware is designed to accommodate both single cell and stack level implementation. It is shown that the proposed system is able to deliver measurements under extreme non-linear conditions that can occur in PEM fuel cells in a fraction of the time associated with normal EIS based systems.
Description
Includes bibliographical references.
Keywords
Reference:
De Beer, C. 2014. Condition monitoring of polymer electrolyte membrane fuel cells. University of Cape Town.