The train door is a moving boundary between two flow systems: the passenger crowd inside the carriage and the passenger crowd waiting on the platform.
A few seconds of door timing can change dwell, headway, crowding, accessibility and the punctuality of trains several stations later.
Platform screen doors are only half of the station interface.
The other half is carried by the train.
Every stop creates a short operating sequence:
train stops → train and platform openings align → door state becomes safe to open → passengers alight → passengers board → obstruction state is checked → doors close and secure → train becomes ready to depart
The sequence looks simple because it repeats thousands of times.
That repetition is exactly why the mathematics matters.
LTA’s current Thomson-East Coast Line page says its four-car trains have more doors for faster boarding. LTA’s Cross Island Line train procurement says each CRL car will have five doors on each side and wider gangways to improve accessibility within the train. The public design message is clear: doorway count, spacing and internal circulation are capacity variables, not merely styling choices.
This pillar owns the train-side passenger interface: door count, opening geometry, door-cycle timing, obstruction states, passenger distribution, accessibility and train-door reliability. The platform-side boundary belongs to Platform Screen Doors. Whole-stop time belongs to Station Dwell Time.
The RFE — What Is a Train Door Actually For?
The weak answer is:
open so people can get on and off
The real public job is:
create a temporary, correctly aligned and safely controlled opening through which the required passenger exchange can occur fast enough for the railway timetable and accessibly enough for the passengers who depend on it.
Prompt 1 — How Does Door Count Become Passenger Capacity?
Suppose door j has effective passenger flow rate μj.
If n equivalent doors are operating independently, a simple upper flow scale is:
μtotal ≈ Σj μj
If all doors have similar rate μ:
μtotal ≈ nμ
But passenger exchange is rarely perfectly balanced.
One doorway may attract more passengers because it is closest to an escalator or transfer corridor.
So exchange time is often controlled by the busiest door:
Texchange ≈ maxj (Nj/μj)
where Nj is the passenger load using door j.
This is why more doors can help, but only if the carriage and platform distribute passengers across them.
Door count creates potential capacity. Passenger distribution decides how much of that potential is actually delivered.
Prompt 2 — How Does Door Width Change Flow?
A wider opening can support multiple passenger streams.
A simple flow model is:
q = ρ v w
where ρ is effective passenger density at the doorway, v is average movement speed and w is effective usable width.
Increasing w tends to increase possible flow.
But the relation is not unlimited because:
- boarding and alighting directions can conflict;
- passengers pause at the threshold;
- carriage crowding can prevent inward movement;
- luggage, wheelchairs and strollers change effective width;
- passengers do not distribute uniformly.
Door geometry is therefore a coupled crowd-flow problem.
Prompt 3 — What Is a Door Cycle?
Total dwell contains several phases.
Tdwell = Tsettle + Topen + Texchange + Tclose + Tsecure + operational margin
The door mechanism controls only part of the dwell.
If mechanical opening becomes one second faster but passenger exchange still dominates, the whole stop may barely improve.
This is another bottleneck lesson:
optimise the dominant phase not merely the easiest phase to measure
A door that closes too aggressively can create repeat obstruction events.
A door that waits too long wastes line time.
The useful objective is a stable cycle, not the fastest theoretical close.
Prompt 4 — How Does Obstruction Detection Become a Probability Problem?
A closing door must distinguish between a clear opening and an obstruction state.
Any detector has two broad error directions:
| Reality | Detector says clear | Detector says obstructed |
|---|---|---|
| Clear | True clear | False obstruction |
| Obstructed | Missed obstruction | True obstruction |
The costs are asymmetric.
A false obstruction can extend dwell.
A missed obstruction carries a much more serious safety consequence.
A generic threshold θ balances detection sensitivity and nuisance events, but actual door thresholds and logic are train-specific and not reproduced here.
LTA’s older public rolling-stock safety material documents obstruction-detection testing and departure inhibition for a historical train model. The enduring public principle is that door state must be verified before movement.
Prompt 5 — Why Does Passenger Distribution Inside the Carriage Matter?
A train can have empty space and still feel impossible to board.
If passengers cluster near doors while interior space remains underused, doorway density becomes the binding constraint.
Let carriage region k have density ρk.
Define distribution imbalance:
Iρ = variance(ρ1,ρ2,...,ρm)
Higher variance means the same passenger count is distributed less evenly.
Wider gangways and internal circulation can reduce local door congestion by allowing passengers to move deeper into the train.
LTA’s CRL design, with wider gangways, is a current Singapore example of this passenger-interface principle.
Prompt 6 — How Does One Door Delay Become a Line Delay?
Suppose one door adds delay δ at station s.
departure delay = δ
The following train’s headway changes.
More passengers accumulate behind the delayed train’s enlarged gap.
Its next dwell can increase.
door delay → headway distortion → different platform arrivals → different boarding load → changed next dwell → wider propagation
This is why door reliability belongs inside line reliability, not merely rolling-stock maintenance.
Prompt 7 — How Is Accessibility Built Into the Door Interface?
A useful train-door interface must work for:
- wheelchair users;
- older commuters;
- passengers with strollers;
- passengers with luggage;
- people who need more time to cross the threshold.
The relevant geometry includes:
horizontal gap + vertical step + door width + available clear area + time available to traverse
A dwell-time optimisation that ignores slower passengers is not public-transport optimisation.
The RFE sets a hard condition:
door cycle must remain feasible for the passengers the railway is required to serve
Prompt 8 — How Does the Railway Know the Door System Is Healthy?
Doors are high-cycle assets.
A generic health vector can include:
x(t)=[open time, close time, motor current, position residual, obstruction events, repeat-close events, fault codes]
LTA’s new BPLRT vehicles publicly include condition monitoring for key systems such as doors, brakes and air-conditioning. The broader mathematical principle applies across rolling stock: repeated cycle data allow a healthy baseline to be compared with gradual drift.
For closing time T:
eT = Tmeasured − Texpected
A persistent upward trend can justify inspection.
One slow cycle does not prove a mechanical fault; passenger obstruction or operating context may explain it.
A Fictional Door-and-Dwell Example
Consider a fictional train with 20 active passenger doors on one side.
Total alighting and boarding demand at one stop is 480 passenger movements.
If perfectly balanced:
average movements per door = 480/20 = 24
Suppose each door processes 1.5 passenger movements/s.
ideal exchange time = 24/1.5 = 16 s
But two doors near an escalator each receive 42 movements.
busiest-door time = 42/1.5 = 28 s
The imbalance adds roughly 12 seconds to the exchange phase even though total passenger count did not change.
Now one repeat obstruction adds 7 seconds.
extra dwell ≈ 19 s
At a 120-second planned headway, 19 seconds is almost 16% of one headway.
A local door event has become a meaningful line-timing disturbance.
Deletion Tests
- Remove door count: train capacity ignores how passengers cross the boundary.
- Remove door width: a narrow and wide opening have identical flow.
- Remove passenger distribution: every door receives equal demand automatically.
- Remove obstruction state: doors can close regardless of the threshold region.
- Remove accessibility: the fastest passengers define the only valid dwell.
- Remove door reliability: thousands of daily cycles have no maintenance consequence.
- Remove headway coupling: one slow door never affects another train.
- Remove World Return: actual door-cycle drift never changes maintenance or design.
Door Paradoxes
- More train capacity can be unusable if doorway capacity is too small.
- More doors can fail to help if passengers cluster at the same few openings.
- A faster-closing door can produce a slower stop if repeat obstruction rises.
- One door can delay an entire train even when every other door is clear.
- The safest door cycle can be slower locally while protecting faster system recovery by avoiding repeated incidents.
The Train-Door Audit
- How many usable doors are presented to the platform?
- What effective width does each provide?
- How are passengers distributed across them?
- Which door becomes the busiest?
- What are opening, exchange, closing and securing times?
- What obstruction states can extend the cycle?
- What passengers require more crossing time or space?
- How does carriage interior circulation affect doorway congestion?
- How does one door delay change headway?
- What door-condition signals are monitored?
- What trend indicates degradation rather than one unusual cycle?
- What actual passenger exchange would prove the door-flow model wrong?
The World Return — When the Door Opens to Real Passengers
predict passenger distribution → predict exchange time → open doors → observe flows and obstructions → close and secure → compare with predicted dwell → update station guidance, train design or maintenance
If one doorway repeatedly dominates, platform markings or passenger distribution may deserve attention.
If door closing time drifts across many comparable cycles, condition may be changing.
If more doors do not reduce dwell, the bottleneck may have moved into the carriage interior or platform approach.
An MRT train door works when a temporary opening can exchange real passengers safely, evenly and repeatedly without turning one threshold into the bottleneck of the whole railway.
Key Equations
μtotal≈Σμj Aggregate door-flow capacity q=ρvw Pedestrian flow through an opening Texchange≈maxj(Nj/μj) Busiest-door exchange time Tdwell=Topen+Texchange+Tclose+Tsecure+... Door contribution to dwell Iρ=variance(ρ1,...,ρm) Passenger distribution imbalance eT=Tmeasured−Texpected Door-cycle condition residual
Reader-safety note: This article does not reproduce Singapore MRT door interlock logic, control thresholds, bypass procedures, exact obstruction-detection parameters or restricted recovery instructions. All numerical examples are fictional teaching values.
Sources and Further Reading
- LTA — Thomson-East Coast Line train design and additional doors
- LTA — Cross Island Line trains with five doors per car side and wider gangways
- LTA — BPLRT vehicle condition monitoring for doors, brakes and air-conditioning
- LTA — historical rolling-stock door safety testing
- eduKateSG — Platform Screen Doors
- eduKateSG — How MRT Works | It’s Mathematics