太深奥了 没几个知道吧

Stationary: 1 weak stationary: conventionally first two moments are independent of time
                   2 strictly stationary: the joint density (y_1........y_t) for any finite "t" is independent of "t"(marginalized joint density f(y_1.......y_t) constant over time)

Hence, stationary 和 nonstationar 的前提不是autocorrlated: independent processes might be stationary and they are not autocorrelated

1, Martingale difference: not autocorrelated but can be nonstationary if the variance are different over time, can be weakly stationary if the variance is constant over time
2,  a sequence(processes) of independent Normal distribution but with different standard deviation: not autocorrelated but nonstationarI

I(1) process is only one special case of nonstationarity