Statistical power of goodness-of-fit tests based on the empirical distribution function for type-I right-censored data