Re: Hurricane Maria: Summary of communication status - and lack of

[Sean Donelan <sean () donelan com>]

On Wed, 27 Sep 2017, Jean-Francois Mezei wrote:
> > After a week without power, all the stationary batteries throughout the
> > telecommunications network are likely completely drained.
>
> from the point of view of cell sites, wouldn't battery autonomy be
> measured in hours rather than days?  I could see some site having
> autonomy in days due to permanent generator, and when fuel runs out so
> does the cell site.

Yes, long-term power is generators.  But there is always a catch.

What happens during disaster recovery is the batteries are damaged by 
being drained repeatedly, dirty power from generators, and enviromental 
conditions. After too many deep-discharge cycles during the disaster, the 
batteries won't hold a charge any more.  The battery failure rate, 
requiring replacement, goes through the roof after about a week in a 
disaster.  Even those 10-year telco batteries don't last 10-years during 
disaster conditions.

Since a lot of telecommunications gear actually runs off -48 volt battery 
string, and the generators recharge the batteries; when the batteries 
completely fail even with a generator, no more telecom.  You have to 
replace the battery string or run the telecom gear on raw generator power 
(which then damages the telecom gear even more).

Sometimes even the battery starter on the generator fail to start after 
too many refueling stops.  Most backup generators are only rated for 
"stand-by" service, not continuous operation for weeks. Generators need 
more maintenance, and fail more often.

Disaster logistics is a string of dominos. If they start being knocked 
over, it just gets worse.  Stuff that works great during normal 
conditions doesn't anymore. Simple fixes are all complicated now.