We systematically analyze X-ray variability of active galactic nuclei ( AGNs ) in the 7 Ms Chandra Deep Field-South survey .
On the longest timescale ( \approx~ { } 17 years ) , we find only weak ( if any ) dependence of X-ray variability amplitudes on energy bands or obscuration .
We use four different power spectral density ( PSD ) models to fit the anti-correlation between normalized excess variance ( \sigma _ { nxv } ^ { 2 } ) and luminosity , and obtain a best-fit power law index \beta = 1.16 ^ { +0.05 } _ { -0.05 } for the low-frequency part of AGN PSD .
We also divide the whole light curves into 4 epochs in order to inspect the dependence of \sigma _ { nxv } ^ { 2 } on these timescales , finding an overall increasing trend .
The analysis of these shorter light curves also infers a \beta of \sim 1.3 that is consistent with the above-derived \beta , which is larger than the frequently-assumed value of \beta = 1 .
We then investigate the evolution of \sigma _ { nxv } ^ { 2 } .
No definitive conclusion is reached due to limited source statistics but , if present , the observed trend goes in the direction of decreasing AGN variability at fixed luminosity toward large redshifts .
We also search for transient events and find 6 notable candidate events with our considered criteria .
Two of them may be a new type of fast transient events , one of which is reported here for the first time .
We therefore estimate a rate of fast outbursts \langle \dot { N } \rangle = 1.0 ^ { +1.1 } _ { -0.7 } \times 10 ^ { -3 } ~ { } galaxy ^ { -1 } ~ { } yr ^ { % -1 } and a tidal disruption event ( TDE ) rate \langle \dot { N } _ { TDE } \rangle = 8.6 ^ { +8.5 } _ { -4.9 } \times 10 ^ { -5 } ~ { } galaxy ^ { % -1 } ~ { } yr ^ { -1 } assuming the other four long outbursts to be TDEs .