Redundancy upgrade for ASM diskgroups can only be done through downtimes. Best tip seems to be to restore the database on a higher redundancy diskgroup. But as disk loss and disk resync does already work with normal or high redundancy diskgroups during the repair of disk failures, it does not seem to be that hard to just add a new failure group to a diskgroup, run the rebalance and so upgrade the redundancy in an online fashion.
The same applies for dropping failure groups to downgrade redundancy.