O.V. Shindor, E.S. Denisov, Yu.K. Evdokimov
One of the most promising sources of electric energy is hydrogen proton exchange membrane fuel cell.
During strong humidification of the membrane-electrodes assembly, fuel cell electrical fluctuations have clearly marked nonstationary nature. The nonstationarity of the electrical operating mode is expressed by large-scale stochastic variations of the fuel cells electrical potential. The variations are accompanied by fluctuation components appearance, which can be considered as «precursors» of the fuel cell potential large-scale variations.
A method based on wavelet analysis of hydrogen fuel cell electrical fluctuations, which allows to detect the appearance of new components in the structure of the electrical fluctuations is proposed in frame of this paper. It is shown that the proposed method allows to predict of large-scale fuel cell potential fluctuations a time before, which is enough to complete actions directed to exclude negative effect of such fluctuations, for example, changing of the fuel cell operating mode or switching on emergency power source. The optimal base wavelet – Daubechies wavelet 2 (db2) is specified.
The obtained in frame of this work results