Power spectrum analysis to identify exotic events in ultra-relativistic heavy-ion collisions

The usefulness of the power spectrum method has been studied for identifying non-statistical fluctuations artificially injected in a small subset of normal simulated events in S + Au and Pb + Pb collisions at CERN SPS energy. Number of peaks in the power spectrum above a threshold is used as the selection criterion. In most cases the efficiency of selection is high and the impurity in the filtered sample due to the statistical fluctuation of normal events is quite low. Special treatment is suggested for weak non-statistical component at a level of admixture comparable to the statistical fluctuation.