I'm just trying to work something out in my head... if a terminus station was receiving a pretty frequent service of electric multiple unit trains, say one arriving every 2-3 minutes, what would be the ideal number of platforms to accommodate such a service, without trains having to wait for a platform to be available?
Assuming for the moment that such a service was headway based, rather than timetable based, and including allowances for reasonable crew turnaround time, boarding/alighting of passengers, the occasional train fault etc. and assuming no grade seperation of arrival and departure lines, so some services would have to cross the paths of others?
Theoretically one platform, but this of course depends upon many factors such as the length and speed of the train in and out of the station (primarily acceleration/deceleration), then the speed of moving pax on and off, then the speed of getting a crew from one end to the other.
The latter can be optimised by having a crew at the ready and the speed of pax can be optimised by using two platforms but one track, eg: the Gatwick terminal shuttles, and also by strictly controlling where the pax stand and the size of the groups getting on and off, plus a few other mobility factors.
The speed at which a train enters the station and more importantly its deceleration is critical, along with the length of the train. I can't do the calculus in my head but let's say it takes 20s to enter the station and 20s to leave with 30s for pax and crew - this would give a theoretical upper limit of around 51 trains per hour.
Adding a second platform would optimise this a little in that you have additional time, around 70s in the above example to get the train, crew and pax ready, but this is more resiliance than anything. An educated guess could probably give you an addition 10 trains per hour - again, need to do the maths (queueing theory this time).
The main bottleneck would be any interference between the incomming and outgoing trains. This can be optimised by having, say, a cross-over or balloon loop. If trains at this point could be spaced as close together as possible in that as one train clears the platform and another enters, I guess you could get to over 70 trains per hour easily (for some definitions of "easily"!)
However as soon as anything goes wrong you've got delays with a single track/platform. Increasing the number of platforms at least keeps things moving albeit with a reduced capacity (from over 60 down to 50, using above theoretical maximums).
You can optimise further by treating the terminal station as a through station by having sidings for preparing and turning trains after the terminal.
Another good rule of thumb is that 70% capacity is a good limit to guarantee some kind of smooth running. So, my guess with 2 platforms and the above figures, 40 trains per hour is the practical limit. 3 platforms gives additional resilience just in case things go really badly wrong.
A more knowledgeable person will confirm that some London Underground stations are equipped with sidings to take trains out of the system to reduce the load when things are getting badly delayed.
The maths is fairly straightforward, but gets complexquickly. I'm not an expert in this area but
https://en.wikipedia.org/wiki/Queueing_theory should give you a flavour.
t.
Ian
NB: maths done in head and with some BIG assumptions (YMMV) and the above probably contains mistakes