How MRT Platform Screen Doors Work Using Mathematics

An MRT train can stop safely at a station and still be in the wrong place for the platform screen doors.

That sounds strange until we remember that a modern platform is not simply an open edge beside a track.

The train doors and the platform doors are two separate mechanical systems.

They must line up closely enough for passengers to pass through the combined opening.

They must enter compatible open and closed states.

They must do so at the correct time.

They must respond safely when something obstructs a door.

They must process hundreds of boarding and alighting passengers without turning every station stop into a bottleneck.

And after millions of operating cycles, they must remain reliable enough that a door fault does not repeatedly become a railway delay.

A platform screen door is not a wall with a motor. It is a geometric, timing, passenger-flow and safety interface between a moving train and a fixed station.

Singapore has made this interface a normal part of MRT travel. The Ministry of Transport stated in 2021 that all MRT stations have platform screen doors. Above-ground stations originally opened without them, but Singapore began retrofitting those platforms in 2009 and completed the programme in 2012.

The system is more integrated than it appears. In a July 2026 parliamentary reply, the Ministry of Transport explained that platform screen doors require power supply, communications and signal-control equipment so that they can operate safely and in synchronisation with train doors. That systems requirement is one reason retrofitting platform screen doors to Singapore’s smaller existing LRT stations is much more difficult.

Future railways preserve the same integration. LTA’s Jurong Region Line contract pairs CBTC signalling with half-height platform screen doors, while the Cross Island Line systems contract covers both CBTC signalling and the PSD system.

This article continues the eduKateSG MRT mathematics cloud. The permanent synthesis page is How MRT Works | It’s Mathematics. The closest specialist pillars are How MRT Signalling and Train Regulation Work Using Mathematics, How MRT Braking Works Using Mathematics and How MRT Station Dwell Time Works Using Mathematics.

The RFE — Why Do Platform Screen Doors Exist?

The weakest answer is:

stop people falling onto the track

That is central.

But the full Reason for Existence is larger.

separate passenger space from railway space
+
open only where a train door is correctly presented
+
allow fast, accessible passenger exchange
+
prevent unsafe departure states
+
reduce track-intrusion risk
+
support station environmental control where applicable
+
remain reliable through repeated daily cycles

A perfectly strong barrier that never opens would be safe from track intrusion but useless for transport.

A very wide opening that moves passengers quickly but opens without the train correctly aligned would be unacceptable.

A door that detects every possible obstruction but produces constant nuisance alarms could damage service reliability.

The optimum is therefore not maximum barrier, maximum speed or maximum sensitivity.

The RFE is controlled access: keep the track unavailable to the passenger until the train creates a safe, correctly aligned and usable doorway through the barrier.

Prompt 1 — How Does a Train Line Up with a Platform Door?

Begin with one-dimensional geometry along the platform.

Let the target train stopping position be:

x*

Let the actual stopping position be:

xstop

Stopping error is:

e = xstop − x*

If e is positive, the train stopped beyond the target in the chosen coordinate direction.

If e is negative, it stopped before it.

The train is correctly presented to the PSD system only when the alignment remains inside an authorised operating tolerance.

|e| ≤ eallowed

The actual Singapore tolerance is not reproduced here.

The mathematics is the important part.

Many doors must align at once

Suppose the centre of train door j relative to the train reference point is d_j.

Its absolute platform position after stopping is:

xtrain,j = xstop + dj

Let matching platform-door centre be p_j.

Door-pair alignment error is:

ej = xtrain,j − pj

If the train and platform were manufactured and maintained perfectly, every e_j would equal the same train stopping error.

Real systems also have manufacturing, installation, wear and measurement tolerances.

A more realistic conceptual error budget is:

ej,total
= estop
+ etrain-door,j
+ eplatform-door,j
+ emeasurement,j

The system should not allocate the entire permissible error to one component.

It needs an error budget.

Opening overlap

Suppose both train and platform openings have nominal usable width w in a simplified model.

If their centres are offset by |e|, geometric overlap is approximately:

woverlap = max(0, w − |e|)

This is deliberately simplified because real door leaves, frames, gaps and accessibility geometry are more complicated.

But it makes the idea visible.

Stopping error does not merely create a prettier or uglier alignment.

It changes the physical opening through which passengers must pass.

Automatic stopping

Automatic Train Operation controls the train’s final approach.

The target is not merely zero speed.

v(tstop) = 0
and
x(tstop) ≈ x*

A control objective can conceptually penalise stopping error:

Jstop
= α(xstop−x*)²
+ βvstop²
+ γjerk
+ ηenergy

subject to safe braking and movement authority.

Stopping accuracy is therefore the last part of the braking problem and the first part of the platform-door problem.

A centimetre of train position belongs simultaneously to signalling, braking, passenger accessibility and platform-door geometry.

Prompt 2 — How Do the Train Doors and Platform Doors Synchronise?

Door synchronisation is easier to understand as a state machine.

A simplified public state sequence is:

train approaching
→ train stopped
→ position/alignment accepted
→ safe opening state established
→ train doors and platform doors open
→ passenger exchange
→ close command
→ doors confirmed closed/secure
→ departure becomes available

The exact signalling logic, interfaces and interlocks are system-specific and not published here.

The mathematical principle is that opening is conditional.

OpenAllowed
= TrainStopped
AND AlignmentValid
AND DoorSystemHealthy
AND OpeningAuthorityValid

Departure is also conditional:

DepartAllowed
= TrainDoorsSecure
AND PSDsSecure
AND TrainMovementAuthorityValid
AND other required conditions

Boolean logic is mathematics too.

Safety-critical control often depends on logical conjunction:

A AND B AND C

rather than an average score in which one good variable can compensate for one unsafe variable.

Timing difference

Let train-door opening time be t_T.

Let platform-door opening time be t_P.

Δtsync = tP − tT

The opening processes need to remain sufficiently coordinated for safe passenger use.

Again, real timing limits are not disclosed.

The important concept is that door opening is a coupled event, not two unrelated motors happening to move near the same time.

Why platform screen doors need signalling integration

The Ministry of Transport’s 2025 and 2026 LRT replies say PSDs require power, communications and signal-control equipment.

That statement tells us exactly what a PSD system is not.

It is not independent station furniture.

train control state
↔ communication
↔ platform-door controller
↔ physical door state
↔ train departure state

This is also why the JRL and CRL contracts couple PSD delivery with signalling-system integration.

A platform screen door becomes part of the railway only when its mechanical state is trusted by the train-control state.

Prompt 3 — How Do Platform Doors Change Passenger Flow?

The train stops correctly.

The doors synchronise.

Now the station becomes a bidirectional flow problem.

Passengers alight from the train while other passengers want to board through the same door zone.

For door j, let:

Aj = number alighting
Bj = number boarding
μj = effective passenger flow rate

A first-order exchange-time estimate is:

Tj ≈ (Aj + Bj)/μj

If all doors must finish before departure, train exchange time is approximately constrained by the slowest important doorway:

Texchange ≈ maxj Tj

This is why passenger distribution matters.

Suppose ten doors each need to process 20 passengers.

That is very different from nine doors processing 10 while one door processes 110.

The total passenger count can be identical while dwell is controlled by a local bottleneck.

Door-load imbalance

Let passenger exchange demand across m doors be n_j.

Mean demand is:

n̄ = (1/m)Σnj

Door-load variance is:

σdoor² = (1/m)Σ(nj−n̄)²

Higher variance means passenger exchange is less evenly distributed.

This can lengthen the maximum doorway exchange time even if total passenger demand is unchanged.

LTA uses priority-queue markings at platform screen doors for passengers such as older commuters, families with strollers and commuters with disabilities. This makes an important systems point: the door zone must not only maximise raw passengers per second. It must preserve equitable and accessible boarding.

Congestion is more than density

Research on platform-door passenger flow has shown that congestion depends on passenger position, movement speed and direction, not merely the number of people in an area.

Two groups moving in opposite directions through one doorway can interfere strongly even at moderate density.

A useful conceptual congestion index is:

Cj = f(density, speed variance, counterflow, direction disorder)

Then effective flow rate becomes:

μeffective = μfree × g(Cj)

with g declining as congestion and interference increase.

The platform-screen-door line therefore also organises passenger behaviour:

queue markings
→ expected door position
→ alighting corridor
→ boarding queues
→ lower counterflow interference
→ shorter and more predictable dwell

The Station Dwell pillar develops this passenger-flow layer in depth.

The platform door is simultaneously a safety barrier and a passenger-flow nozzle. Its geometry changes how quickly the train can become ready to move again.

Prompt 4 — What Does Obstruction Detection Have to Solve?

A door should not close through a passenger or object.

That sounds like a simple binary problem:

obstruction
or
no obstruction

In real sensing, observations are imperfect.

A detection system receives some measurement z and must decide whether the door path is clear.

There are four possible outcomes.

                    Actual obstruction   Actually clear
Detect obstruction        TP                  FP
Detect clear              FN                  TN

A false positive means the system believes there is an obstruction when the doorway is actually clear.

That can create unnecessary re-opening, delay or maintenance investigation.

A false negative means a real obstruction is not detected.

Its safety consequence can be much more serious.

This makes the decision asymmetric.

Cost(FN) ≫ Cost(FP)

for safety-relevant obstruction states, while still recognising that excessive false positives can damage service availability.

Detection threshold

Suppose an obstruction score S is produced.

if S ≥ θ
→ treat as obstruction

Lower θ catches more ambiguous events but may create more false positives.

Higher θ reduces nuisance detections but can increase missed detections.

No real MRT threshold is disclosed here.

The educational point is that safety detection is a statistical decision, not magical certainty.

Sequential confirmation

A robust public conceptual architecture is:

close command
→ monitor doorway state
→ detect obstruction if present
→ respond according to authorised logic
→ confirm final secure state
→ permit departure only when required conditions are satisfied

The door does not simply move from “open” to “closed”.

It moves through a verified sequence.

That sequence is safety.

The mathematical challenge is not detecting everything. It is refusing to declare the doorway safe until the evidence is strong enough for the next railway state.

Prompt 5 — How Reliable Must Hundreds of Platform Doors Be?

One station can have many door modules.

A line has many stations.

Each door opens and closes repeatedly throughout the operating day.

This creates a large number of component cycles.

If one door has probability p of a relevant failure in one period, and n doors behaved independently, probability of at least one failure would be:

P(at least one failure)
= 1 − (1−p)^n

For very small p:

1 − (1−p)^n ≈ np

So even rare component faults can become noticeable when multiplied across many components and cycles.

Actual service impact is more nuanced because door architectures can isolate individual modules and operators have authorised response procedures.

The statistical lesson remains:

Mass transit reliability is created by making small component failure probabilities extremely small before they are multiplied by a very large number of repetitions.

Availability

For a repairable door subsystem:

A ≈ MTBF/(MTBF + MTTR)

High Mean Time Between Failures reduces fault frequency.

Low Mean Time To Restore reduces the duration of each loss.

Both matter.

Preventive maintenance

In a 2019 parliamentary reply, MOT said rail operators had a monthly preventive-maintenance regime for PSDs at that time, including component inspections and functional tests of mechanical, electrical and pneumatic systems, with servicing and replacement carried out when required.

That historical public description reveals the maintenance logic:

inspect repeatedly
→ test function
→ identify degradation
→ service or replace
→ return to reliable cycle operation

A current maintenance regime may differ in detail, so the 2019 interval should not be treated as a present-day operating instruction.

Cycle-based deterioration

Suppose door module j completes N_j cycles.

Health can be modelled as:

Hj = H(Nj, age, environment, fault history)

A heavily used central station may accumulate cycles faster than a quiet station.

Calendar age and operating age are not identical.

This connects PSD maintenance to the Predictive Maintenance pillar.

Prompt 6 — Why Do Platform Screen Doors Affect Airflow, Pressure and Energy?

A train moving through a tunnel pushes air.

The pressure field around a platform door therefore changes as trains approach, arrive, depart or pass nearby.

A simple dynamic-pressure scale is:

q = ½ρv²

where ρ is air density and v is flow speed.

Actual station pressure is much more complex because it depends on:

train geometry,
tunnel cross-section,
station openings,
shaft configuration,
train speed,
other trains,
and door state.

Research on metro platform screen doors has measured transient aerodynamic pressure and vibration loads produced by train movement. Finite-element and fatigue studies then translate those loads into stress and expected structural life.

Structural response

Let aerodynamic pressure on a door panel be p(t).

Resulting structural stress is conceptually:

σ(t) = Fstructure[p(t), geometry, supports, material]

Repeated stress cycles can produce fatigue.

Miner’s linear cumulative-damage rule is:

D = Σi ni/Ni

where:

n_i is experienced cycles at stress level i,
N_i is estimated cycles to fatigue failure at that level.

D approaching 1 is often used as a classical fatigue-damage indicator under the model’s assumptions.

Real PSD design uses validated standards and engineering analysis, not this teaching equation alone.

Full-height doors and station cooling

At underground stations, platform screen doors also help separate the conditioned platform environment from the tunnel.

LTA’s Green MRT material explicitly says underground platform screen doors help reduce air-conditioning energy consumption.

This gives the PSD another job:

safety barrier
+
passenger opening
+
environmental boundary

Above-ground MRT stations use half-height PSDs. LTA’s original master-plan rationale noted that this height protects commuters while preserving natural airflow and avoiding the energy cost of air-conditioning an open elevated platform.

The design therefore depends on station environment.

The same door geometry that keeps a passenger away from the track can also become part of the station’s airflow and energy strategy.

Prompt 7 — What Happens When a Platform Door Fails?

A platform-door fault is not automatically a line-wide failure.

That depends on the fault, the system architecture and the authorised degraded operating mode.

The public reasoning chain is:

door fault detected
→ identify affected module or system
→ place it in a known safe condition
→ determine whether passenger use and train departure remain authorised
→ isolate or manage the fault according to procedures
→ restore normal operation

This article deliberately does not reproduce operational bypass or isolation procedures.

Service-delay probability

Let:

P(F) = probability of a door-system fault
P(D|F) = probability the fault produces a relevant service delay

Then:

P(delay from PSD fault)
= P(D|F)P(F)

A maintenance programme can improve reliability by reducing P(F).

A resilient operating architecture can reduce P(D|F) by containing some faults locally when safe and authorised.

Those are different interventions.

In 2019, MOT reported that there had been ten PSD faults causing delays of more than five minutes from 2015 to 2017, followed by none in the 19 months to July 2019. That historical snapshot illustrates why both equipment reliability and maintenance matter, though it is not a current reliability statistic for 2026.

Door failure can become headway failure

Suppose one PSD fault adds Δd seconds of dwell.

Headway behind the delayed train changes.

Passenger accumulation changes.

The fault can propagate:

door fault
→ extra dwell
→ headway distortion
→ more passengers accumulated
→ longer downstream dwell
→ wider delay

The door is mechanically local.

The delay it creates can become network-wide.

This connects directly to How MRT Delays Propagate and Recover Using Mathematics.

A PSD is one of the smallest station assets that can place a time signature onto every train following behind it.

Prompt 8 — How Do Platform Doors Work During Emergencies?

The normal-day problem is:

keep passenger and track separated
until a correctly aligned train creates a safe opening

Emergency conditions can reverse the priority.

The system may need to support evacuation, emergency access or other authorised response states.

The public mathematical lesson is that an evacuation bottleneck is controlled by available opening width and passenger flow.

If opening width is w and effective evacuation flow per unit width is q:

Flow capacity ≈ q w

For N passengers:

Tevac ≳ N/(qw)

This simplified relationship explains why available doorway width matters strongly during evacuation.

Research on underground station evacuation has shown that restricted PSD opening conditions can significantly increase evacuation time under heavy passenger loads, reinforcing the importance of correct train stopping and coordinated door availability.

Real emergency procedures are far richer and are not reproduced here.

The important design principle is:

A barrier designed for normal safety must also possess authorised states for the abnormal day when movement through the barrier becomes part of safety.

A Complete Fictional Platform-Door Example

Consider a fictional metro train with six passenger-door pairs per car-side section under analysis.

All numbers are invented and do not represent Singapore MRT PSD tolerances, speeds or control thresholds.

Step 1 — Stopping error

Target stop:

x* = 0

Actual stop:

xstop = +0.08 m

Stopping error:

e = +0.08 m

Suppose fictional authorised alignment tolerance is ±0.15 m.

The train is inside this teaching tolerance.

Step 2 — Opening overlap

Suppose nominal combined doorway width is:

w = 1.40 m

Simple overlap estimate:

woverlap = 1.40 − 0.08
         = 1.32 m

Again, real usable clear width is more complex than this simplified centre-offset calculation.

Step 3 — Passenger distribution

Door exchange demands are:

Door 1: 22 passengers
Door 2: 24
Door 3: 46
Door 4: 25
Door 5: 20
Door 6: 23

Total:

160 passengers

Mean per door:

n̄ = 160/6
  ≈ 26.7

Door 3 is the clear local bottleneck.

Step 4 — Flow rate

Suppose effective exchange rate at uncongested doors is:

μ = 2.0 passengers/s

Door 3 requires roughly:

T3 = 46/2.0
   = 23 s

Most other doors need around 10–13 seconds.

Dwell is therefore constrained by Door 3, not the average door.

Step 5 — Counterflow congestion

Suppose boarding passengers begin moving into Door 3 before alighting passengers fully clear.

Effective flow rate falls to:

μeffective = 1.6 passengers/s

Then:

T3 = 46/1.6
   ≈ 28.8 s

Five to six seconds of dwell were created without any train fault or door fault.

The bottleneck was passenger interaction at one door.

Step 6 — Obstruction event

During closing, an obstruction is detected.

Suppose the authorised response adds 7 seconds before a secure closed state is confirmed.

Total extra station time relative to the uncongested ideal can become approximately:

passenger counterflow penalty ≈ 5.8 s
obstruction response          = 7.0 s
combined additional time      ≈ 12.8 s

The train itself is healthy.

Yet one platform-door event has now changed the headway.

Step 7 — Headway propagation

Suppose target headway is 120 seconds.

The 12.8-second delay creates:

new gap ahead ≈ 132.8 s

At passenger arrival rate λ=3 passengers/s, the extra 12.8 seconds can accumulate:

ΔN ≈ 3×12.8
   ≈ 38 extra passengers

Those passengers may increase dwell at the next station.

A platform-door event has become a train-regulation problem.

Step 8 — World Return

Suppose models predicted Door 3 exchange in 24 seconds.

Observed exchange was 28.8 seconds.

eT = 28.8 − 24
   = 4.8 s

Repeated positive error suggests the passenger-flow model is underestimating counterflow or door-local crowding.

The response might involve passenger guidance, queue markings or other station-flow measures rather than changing the PSD mechanism itself.

The correct fix follows the cause. A slow doorway caused by passenger geometry should not be misdiagnosed as a slow motor.

The Platform Screen Door Deletion Tests

Remove stopping accuracy

Train and platform openings can no longer be assumed to overlap correctly.

Remove signalling integration

The platform door cannot reliably know whether the train state permits opening or departure.

Remove door-state confirmation

A command is mistaken for a successfully completed physical action.

Remove obstruction detection

The system loses evidence that the doorway may not be clear during closing.

Remove passenger flow

The door is treated as instantly usable regardless of crowding and counterflow.

Remove accessibility

A doorway is judged only by maximum flow rate rather than whether different passengers can use it safely.

Remove preventive maintenance

Repeated mechanical, electrical and control cycles continue until faults emerge in service.

Remove aerodynamic load

Door structures are assumed to experience no pressure or vibration from moving trains.

Remove emergency states

A barrier designed for normal operation has no safe role when evacuation or emergency access is required.

Remove World Return

Predicted dwell, door faults and passenger flow are never compared with what the station actually experienced.

The Platform Screen Door Paradoxes

Paradox 1 — A door can be mechanically perfect and operationally unusable

If the train stops outside the authorised alignment, the PSD mechanism itself may be healthy but the combined interface cannot operate normally.

Paradox 2 — Better safety can improve capacity

A clear protected platform edge can support more predictable passenger organisation and reduce track-intrusion disruption.

Paradox 3 — A wider door can still be slow

Counterflow, passenger clustering and local geometry can reduce effective throughput even when physical width is generous.

Paradox 4 — More sensitive obstruction detection can create more delays

Safety sensitivity and false-positive availability must be engineered together.

Paradox 5 — Half-height doors can be better than full-height doors in one environment

At naturally ventilated elevated stations, half-height doors preserve airflow without requiring a fully conditioned platform enclosure.

Paradox 6 — Full-height doors can save station energy

At underground stations, separating conditioned platform air from the tunnel can reduce air-conditioning load.

Paradox 7 — A local door fault can become a line-level delay

Extra dwell changes headway, passenger accumulation and downstream station timing.

Paradox 8 — The safest barrier must be designed to open

Public transport requires controlled permeability: sealed most of the time, open exactly when the railway state makes crossing safe.

The PSD Systems Audit

Where is the train supposed to stop?
What is the measured stopping error?
How much uncertainty belongs to the train, platform and measurement system?
Are all door pairs geometrically aligned inside authorised tolerance?
Is the train fully stopped?
Is the opening state authorised by the signalling/control system?
Are train-door and platform-door states synchronised?
How much usable opening width remains?
Which doorway carries the heaviest passenger exchange?
What is the local effective passenger-flow rate?
Is counterflow or queue disorder reducing throughput?
Are priority and accessible boarding needs represented?
What evidence indicates an obstruction?
What false-positive and false-negative risks exist?
Has the final secure state been physically confirmed?
How many operating cycles has the door accumulated?
What failure trend or maintenance history exists?
What aerodynamic and structural loads does the door experience?
How does the door type affect ventilation or station energy?
What happens if one door module is unavailable?
What degraded safe state is authorised?
How much delay would the fault add to the train?
How does that delay propagate into headway?
What emergency states must the barrier support?
What observation after maintenance proves the problem is actually fixed?

How the Mathematics Grows from School to Research

Primary Mathematics

length and position,
time intervals,
counting passengers,
simple rates.

Secondary Mathematics and Science

coordinate geometry,
error and tolerance,
probability,
flow rates,
force and pressure.

Junior College

statistics and classification,
optimisation,
feedback,
mechanics and fluid pressure,
reliability probability.

University and Research

automatic train control,
mechatronics,
formal safety logic,
pedestrian-flow modelling,
computer vision and congestion detection,
finite-element structural analysis,
aerodynamics,
fatigue and reliability engineering,
maintenance optimisation,
emergency-evacuation modelling.

A child measures centimetres.

A railway engineer asks whether those centimetres still leave every train door, platform door, passenger and safety state compatible across millions of stopping events.

The World Return — When the Platform Answers the Model

The train-control model predicts stopping position.

The train returns an actual stop.

estop = xactual − xpredicted

The passenger model predicts door exchange time.

The station returns actual dwell.

edwell = dobserved − dpredicted

The maintenance model predicts failure probability.

The door returns actual faults and inspection findings.

predicted fault risk
vs
observed condition and failure history

The aerodynamic model predicts pressure and structural response.

Measurements and inspections return physical evidence.

The loop is:

design
→ stop train
→ align
→ open
→ exchange passengers
→ close
→ confirm
→ depart
→ record errors and faults
→ maintain or recalibrate
→ repeat

World Return prevents the PSD from being judged only as a successful mechanical design.

It asks whether it remained:

safe,
aligned,
available,
accessible,
fast enough for passenger exchange,
and maintainable after repeated real operation.

The platform screen door works only when the train, the passenger and the door all return the state that the model expected.

RFE Return — What Does a Platform Screen Door Owe the Passenger?

The passenger should experience the door as almost obvious.

The train stops.

The openings line up.

The doors open.

People alight.

People board.

The doors close.

The train leaves.

What disappears into that ordinary event is:

stopping control
+
position tolerance
+
communications
+
signalling logic
+
mechatronic motion
+
obstruction evidence
+
passenger-flow geometry
+
accessibility
+
reliability
+
maintenance
+
emergency readiness

The passenger does not need maximum technical sophistication.

They need trustworthy controlled access.

The RFE of the platform screen door is to make the dangerous space unavailable until the exact moment the railway has turned it into a safe doorway.

Conclusion — A Station Door Is a Precision Interface

A train approaches the platform at speed.

The signalling system knows its movement state.

ATO shapes the final braking trajectory.

Velocity reaches zero.

Position must reach the correct place too.

The PSD system receives a compatible train state.

Train and platform doors open into one shared passage.

Passenger flow begins.

One doorway becomes crowded.

Dwell stretches.

An obstruction is detected.

The door sequence responds.

The final secure state is confirmed.

Departure becomes possible again.

braking
→ stopping accuracy
→ alignment
→ signalling permission
→ door synchronisation
→ passenger flow
→ obstruction handling
→ secure closure
→ departure
→ headway
→ maintenance feedback

Singapore’s platform-screen-door history shows the system growing from infrastructure improvement into normal expectation. Above-ground MRT stations were retrofitted between 2009 and 2012. Today all MRT stations use PSDs, while the 2026 LRT discussion makes clear how much hidden power, communications and signalling infrastructure is required to make synchronised barriers work safely.

The train-control engineer sees a stopping coordinate.

The PSD engineer sees alignment and verified door states.

The station planner sees passenger flow through a bottleneck.

The reliability engineer sees millions of repeated cycles.

The energy engineer sees an environmental boundary at underground stations.

The passenger sees two doors opening together.

A platform screen door works when centimetres, seconds, logic, people and machinery agree so completely that the passenger experiences only an opening.

Key Equations

e = xstop − x*
Train stopping error

xtrain,j = xstop + dj
Absolute train-door position

ej = xtrain,j − pj
Door-pair alignment error

ej,total = estop + etrain-door + eplatform-door + emeasurement
Conceptual error budget

woverlap = max(0,w−|e|)
Simplified aligned opening width

Jstop = α(xstop−x*)² + βvstop² + γjerk + ηenergy
Conceptual precision-stop objective

OpenAllowed = A AND B AND C ...
Boolean safety-condition structure

Δtsync = tPSD − ttrain
Door synchronisation timing difference

Tj ≈ (Aj+Bj)/μj
Door passenger-exchange time

Texchange ≈ maxj Tj
Door-bottleneck exchange time

σdoor² = (1/m)Σ(nj−n̄)²
Door-load imbalance

Recall = TP/(TP+FN)
Obstruction-detection sensitivity

Precision = TP/(TP+FP)
Alarm precision

P(at least one fault) = 1−(1−p)^n
Multi-component failure probability under independence

A ≈ MTBF/(MTBF+MTTR)
Generic repairable-system availability

q = ½ρv²
Dynamic-pressure scale

D = Σni/Ni
Miner cumulative fatigue damage

P(delay from fault)=P(D|F)P(F)
Fault-to-service-delay probability

Flow ≈ qw
Simplified evacuation-flow capacity

Tevac ≳ N/(qw)
Simplified evacuation lower-bound time

estop = xactual−xpredicted
edwell = dobserved−dpredicted
World Return errors

Reader-safety note: All stopping tolerances, opening widths, obstruction thresholds, timing values, door-control states and degraded-operation examples in this article are fictional or generic educational abstractions unless explicitly attributed to a public source. This article does not reproduce Singapore MRT PSD safety tolerances, interlock logic, sensor thresholds, bypass procedures, emergency-control sequences or other restricted operating details.

Continue the MRT Mathematics Cloud

How MRT Works | It’s Mathematics — the permanent whole-system hub.
How MRT Timing Works Using Mathematics — system timing.
How MRT Braking Works Using Mathematics — stopping trajectories and platform accuracy.
How MRT Headway Works Using Mathematics — train spacing and capacity.
How MRT Station Dwell Time Works Using Mathematics — doors, passengers and station time.
How MRT Delays Propagate and Recover Using Mathematics — how local station delays spread.
How an MRT Timetable Is Built Using Mathematics — planned service events.
How MRT Energy Use Is Optimised Using Mathematics — energy and regeneration.
How MRT Passenger Capacity Is Calculated Using Mathematics — delivered passenger capacity.
How MRT Routes and Transfers Are Optimised Using Mathematics — route and transfer costs.
How MRT Network Resilience Is Measured Using Mathematics — disruption and recovery.
How MRT Predictive Maintenance Works Using Mathematics — condition monitoring and failure probability.
How MRT Track Inspection Works Using Mathematics — spatial inspection.
How MRT Wheel–Rail Contact Works Using Mathematics — wheel–rail forces and wear.
How MRT Noise and Vibration Work Using Mathematics — waves and vibration.
How MRT Power Supply Works Using Mathematics — the moving electrical grid.
How MRT Signalling and Train Regulation Work Using Mathematics — CBTC, ATO and real-time regulation.

The next natural pillar is How MRT Tunnel Ventilation and Airflow Work Using Mathematics: piston effect, pressure waves, ventilation, heat, smoke control, station cooling, tunnel cross-sections and why a moving train also behaves like a giant piston inside an underground tube.

How MRT Platform Screen Doors Work Using Mathematics: Why Centimetres of Stopping Error Matter