[mmaimcal] reconfiguration thoughts

[Simon Radford]

After last Monday's discussion (July 12), I had another look at
reconfiguration times with the idea of refining earlier estimates 
(Memos 147, 199, & 265) to allow for the differences in configuration
sizes. Also, in the spirit of presenting a baseline plan that might be
improved later, I suggest an operating cycle for the ALMA and the
necessary number of transporters.

My estimates of reconfiguration times are slightly more pessimistic than
those in Memo 147 and 199, but a bit more optimistic than those in Memo
265. Because it will take significantly longer to move antennae into the
largest (10 km) configuration than into the smaller configurations, I
suggest we deploy the 10 km configuration less often. I suggest an 18
month cycle where the 10 km configuration is used once each cycle and
the smaller ones twice. This means a reconfiguration every 2 months.

The relevant specifications are 15 minutes for antenna pickup or
dropoff, including cables connections (RFP); transporter speeds of 10
km/h on flat terrain, 5 km/h on a 10% grade, and 20 km/h unloaded on
flat terrain (Project Book); 16 m/s maximum wind speed during transport
(Project Book); and five configurations, with diameters of 150 m (40%
filled), 400 m, 1.1 km, 3 km, and 10 km (current baseline plan). 

Between the small configurations, 150 m - 1.1 km, it should be possible
to move an antenna in an hour: 15 min each for pickup and dropoff plus
15 min of low speed maneuvering with the transporter plus 15 min general
overhead.

For the large configurations, the transport time dominates. As a first
approximation, assume the stations are uniformly distributed along a
road about the circumference of a circle. Then starting from a compact
configuration tangent to the large configuration, the mean travel
distance is one quarter of the circumference, pi D/4 or approximately
the diameter of the circle, D, since pi = 4 to the necessary accuracy.
If the average transporter speed is 7 km/h (15 km/h unloaded), then it
should take about 1.5 h to move the average antenna to the 3 km
configuration (pickup, dropoff, and overhead as above plus 30 min
transport plus 15 min transporter return) and 2.25 h for the longest
move, which is half the circumference (overheads plus 60 min transport
plus 30 min return).  For the 10 km configuration, the average move will
be 3 h (overheads plus 90 min transport plus 45 min return) and the
longest move will be 4.5 h (overheads plus 3 h transport plus 90 min
return). For simplicity, I ignore whether it is quicker to go from
between the 3 km to 10 km configurations than between the 150 m
and 3 km configurations. 

High winds may complicate reconfiguration. The wind is strongest during
the afternoon and during winter months, the median afternoon wind speed
approaches the wind speed limit for the transporter (16 m/s). If
antennas are not moved at night and cannot be moved during the
afternoon, total reconfiguration times will be proportionately longer.
This can be avoided by operating more transporters simultaneously.
Moreover, winter snowstorms might preclude antenna transport until roads
are plowed, a further delay.

For safety reasons, if nothing else, a transport crew should have at
least 3 people, one to operate the transporter (driver) and two to guide
the driver, connect cables, etc. (flappers). If several transporters
operate simultaneously, the necessary number of people could be reduced
by having two flaggers for antenna pickup, two flaggers for dropoff, and
one driver per transporter, rather than a complete crew for each
transporter. This arrangement might be especially suitable for moves
between small configurations, when the pickup and dropoff times are
comparable to the transport times. 

The time required for pointing, baseline determination, etc., should be
independent of the configuration size. Previous estimate of this
overhead seem well considered, but I have not included them in my totals
below. (By the way, Min, we must sample the Az-El *hemisphere*, not
*plane*!)

With five transporters operating 6 hours per day with no delays, the
ALMA could be redeployed between small configurations in just over 2
days (6 antennae per day per transporter), into the 3 km configuration
in just over 3 days (4 antennae per day per transporter), and into the
10 km configuration in 6.5 days (2 antennae per day per transporter).
With five transporters, reconfiguration times are still reasonably short
even if weather delays, etc., reduce the effective operating day to only
4 or 4.5 hours (3.3, 4.5, and 13 days for small, 3 km, and 10 km
reconfigurations). An abbreviated work day is equivalent, of course, to
a full work day for fewer transporters, three in this case.

For an 18 month (548 day) cycle that uses the 10 km configuration once
and the others twice, there are four small moves, three intermediate
moves, and two long moves (Figure below). With five transporters, these
can be accomplished in 30 moving days, about 5% of the cycle. If the
work day is abbreviated, then 53 moving days are required (10% of the
cycle).

Configuration cycle (first line) and moving days for five transporters
(second line) and for abbreviated work days (third line):
150 - 450 - 1100 - 3 k - 10 k - 150 - 450 - 1100 - 3 k - 150 - 450 ----
    2     2      3    6.5    6.5    2     2      3     3     3     ---
   3.3   3.3    4.5    13    13    3.3   3.3    4.5   4.5   3.3    ---