Paper addresses the problems of information fusion with the goal of estimation of an overall state-of-health indicator. The main idea is based on the fact that various processes governing the operation of fuel cells have inherent randomness. Consequently single frequency components of the impedance characteristic can be regarded as random variables with Rayleigh distribution. Performing information fusion is done by deriving Rayleigh copula function. The results show the effectiveness of the proposed approach for detecting various water management faults in PEM fuel cells.
COBISS.SI-ID: 30736167
The paper addresses the issues of online estimation of remaining useful life of SOFC fuel cells. This paper proposes an integrated approach for SoH estimation based on stack’s Ohmic area specific resistance. A drift model that describes the area specific resistance increase over time enables accurate remaining useful life prediction. The approach is a combination of sequential Monte Carlo approaches. The validation is performed using 10 kW SOFC stack.
COBISS.SI-ID: 30661927
Electrochemical energy systems belong to a special group of so-called fractional-order systems. Time domain identification of such systems offer quite fast characterisation, however current approaches have particular numerical issues. The paper presents methods for time domain identification of fractional-order systems using modulating functions.
COBISS.SI-ID: 32104743
The paper presents a methodology for identification of fractional order systems with algebraic manipulation and its application on SOFCs. The identification process is based only on least squares approach and as such is very computationally efficient. Additionally, the accuracy is preserved even with substantially shorter measurement signals which is important for shortening the charaterisation cycles.
COBISS.SI-ID: 32095271
This book presents newly proposed method for fast impedance spectroscopy that is based on time-frequency analysis. Additionally the complete statistical analysis of the impedance data is presented together with the methods for determining diagnostic thresholds. Furthermore, the concepts of the embedded system for online condition monitoring of electrochemical energy systems is presented. Finally, the proposed algorithms were evaluated on an PEM fuel cell system.
COBISS.SI-ID: 30534439