Blueprint does not have technical replicates, you will find that very few sequencing based experiments ever have technical replicates. As biological replicates: each individual is regarded as a replicate.
This makes sense for the sort of experimental design which blueprint was designed to answer questions such as "Do individuals from group A have higher or lower H3K4me3 than individuals from groups B" or "Does cell type A have higher accessibility than cell type B". In each of these cases the within group variability to be considered is encapsulated in the difference between individuals.
Overall technical variation is assumed to be similar in all samples or at least not correlated with the contrast of interest (this may well not be true in the second example design I gave) it is safe to directly measure it, as long we know its distribution.
As an aside, one of the reasons that ENCODE got away with only two reps is that their reps are effectively technical (repeated preparations from the same cell lines) and thus there is next to no biological variation (which is what our algorithms are set up to find).