Re: Sydney Metro control roomand 24 hour service

On 11/10/24 23:43, Geoffrey Hansen wrote:

> I sometimes think that the Metro is like a large model train layout
> controlled from a remote centre.


It is basically that. The Sydney Metro M1 trains are not smart or in any
way autonomous. The train simply follows speed codes transmitted from
the ground equipment. It's basically an extension of 100 year old cab
signalling technology, except a human isn't looking at the signal and
translating that to a power/brake setting, the indication is being feed
straight to the speed controller.

The signalling system works out which of a limited number of speeds
ensures safe movement and that is transmitted to the train. The train
blindly follows that instruction.

Sydney Metro 'regulation' is performed by the control center staff
withholding departure authority for trains at a stations. It's possible
the master train planning computer can do this with human intervention.
The trains them selves have no idea of route profile or operating
diagram, they just follow the speed code sent to them.

The simple minded system frequently results in the trains 'hunting' for
an ideal speed. Particular up the grade from Epping to Cherrybrook, the
trains hunt badly, with pulsed application of power. The train powers
for the grade, reaches line speed, over speeds slightly, cuts off, slows
down below target, then powers again. 'bang bang control'. With their
modern electronic traction drive, any level of power could be applied
for traction, but no, it goes full on, full off and back to full on again.

I've also observed multiple times trains charging up the grade when they
go onto the viaduct section further west at full power, crest the grade
still at full power then go to full brake. A decent control system (and
a competent human driver) would cut off before the top of the grade to
avoid suddenly speeding up when the grade eases.

I hope Siemens provides a more intelligent system for Metro West. I've
certainly been on GoA4 metros where the train does exhibit route
knowledge uses momentum properly.

There are any number of systems referred to as 'distance to go' where
the signalling sends the 'distance' to the next speed / authority change
and the train itself calculates the appropriate speed profile to meet
that authority. This is NOT what Alstom supplied for 'M1', instead it's
an old school 'send speed code to train and train runs at that speed'. I
think Alstom also have Distance-to-Go systems, but I suspect we got the
old speed-code system because TfNSW specified 'tried and tested system'.
So we got a rehash of a 40 year old control system instead of 'state of
the art'.

Note i've seen similar tech applied to trams - Dubai's trams have to
line up with the platform doors at stops, so the trams auto-brake into
the stations. When the tram hits the station beacon, the ATP takes over
from the human and brings the tram to a stop within the 10cm tolerance
for the platform doors.

A good human driver is more than capable of this accuracy, but not all
drivers employed by operators care about their 'craft'.