From the Windows failover cluster manager,
select the group listener, stop it, and delete it.
Do the same thing for the database.
Pick a node that will "own" the database.
Then, on the remaining nodes, execute the following command at a CMD prompt:
oradim -SID {sid} -DELETE
On all nodes execute the following command
sc delete {Windows service name for listener}
Edit the listener.ora and remove the declaration for the group listener.
Do this on all nodes.
Edit the tnsnames.ora on the nodes
that won't own the db and remove all entries that reference the database.
Edit the tnsnames.ora on the node that owns the database
and modify the address list for the database so that it contains the address of the local node.
That should clean things up.