SQL Server Alwayson Failover process - Three node cluster

I have 3 node windows cluster spread across 2 sites.
OS Windows 2012
SQL 2012


SITE2 (DR Site)

SQLServer1 and 2 are part of the SQL FCI, whereas the SQLServer3 has Alwayson Availaibilty group setup as Ashyrnchrous between the SQL FCI with Listener setup to route traffic.

Site 2 is the DR site and intended for manual failover when Site one is not available.

I can manually failover the Availability group between the sql fci(site1) and site 2(SQLserver3) but my question here is:
If site one goes down what are the steps to failover to site2 as the Windows cluster will not be available and also what are the recommended settings in windows failover cluster manager particularly for the quoram.
And like wise failback process?