Every MRT journey contains a moment most passengers barely notice.
The train is moving quickly. Then, long before it reaches the platform, something changes. The motors stop pushing as hard. The train may coast. Braking begins. Speed falls smoothly. The platform appears beside the windows. The train continues slowing until, almost without drama, several hundred tonnes of moving railway equipment become stationary at exactly the place where passengers need to board and alight.
To a passenger, this is simply “the MRT stopping”.
To mathematics, it is a problem involving velocity, acceleration, kinetic energy, stopping distance, gradients, adhesion, control theory, signalling, passenger comfort, energy recovery and uncertainty — all being solved under strict safety constraints.
This article is the braking companion to our larger guide, How MRT Timing Works Using Mathematics. The earlier article asks how an entire railway keeps time. Here, we zoom into one event inside that system: how a train decides when, how hard and how smoothly to stop.
The important question is not “Can the train stop?” It is “Can the train stop safely, predictably, comfortably and at the correct place under the range of conditions the railway expects?”
RFE — Why Does MRT Braking Mathematics Exist?
The Reason for Existence is straightforward: a railway carries people through a world that contains uncertainty.
A train must be able to reduce speed for stations, curves, junctions, trains ahead and changing movement permissions. It must arrive accurately enough for the station environment to work. It must protect passengers who are standing. It must remain dependable across repeated journeys. And, where possible, it should avoid wasting the enormous kinetic energy that had just been used to make the train move.
Braking exists to convert motion → controlled reduction of motion → safe stationary state
But “controlled” is doing a great deal of work in that sentence.
Good MRT braking is not merely maximum braking force. It is the correct braking force, beginning at the correct time, following the correct curve, with enough safety margin, while preserving passenger comfort and the timing of the railway around it.
Prompt 1 — What Is Braking, Mathematically?
Start with the simplest possible description.
If velocity is:
v = dx/dt
then acceleration is:
a = dv/dt
When acceleration acts opposite to the direction of motion, velocity decreases. We commonly describe the magnitude of that reduction as deceleration.
For a deliberately simplified constant-deceleration example, one familiar kinematics equation is:
v² = u² + 2as
If the train stops, final velocity v = 0. If we write the positive braking magnitude as b, then a = −b, giving:
0 = u² − 2bs
and therefore:
s = u² / (2b)
This is the first major equation in train braking.
Stopping distance grows with the square of speed.
That square changes everything.
An illustrative 80 km/h example
The numbers below are an educational model only. They are not published SMRT or LTA operating braking parameters.
Suppose a train is travelling at 80 km/h:
80 km/h = 22.22 m/s
Suppose our illustrative constant braking magnitude is:
b = 1.1 m/s²
Then the simplified stopping distance is:
s = 22.22² / (2 × 1.1) s ≈ 224 m
Stopping time under the same simplified assumption is:
t = u / b t = 22.22 / 1.1 t ≈ 20.2 s
So in this model, the station begins influencing the train more than two hundred metres before the train is stationary.
In a real metro, braking is not adequately described by one constant number. Control systems use speed profiles and braking curves, and safety calculations include many more factors than this classroom equation. But the equation teaches us the shape of the problem.
Prompt 2 — Why Does a Little More Speed Need So Much More Stopping Distance?
The square law is worth slowing down to understand.
For the same illustrative deceleration:
s ∝ v²
If speed doubles:
v → 2v
then:
s → (2v)² = 4v²
So the simplified braking distance becomes four times as large.
This is why railway safety cannot treat speed as a linear variable.
Compare three speeds
Again using b = 1.1 m/s² only as an educational constant:
40 km/h = 11.11 m/s → s ≈ 56 m 60 km/h = 16.67 m/s → s ≈ 126 m 80 km/h = 22.22 m/s → s ≈ 224 m
Speed doubles from 40 to 80 km/h.
Stopping distance in the simplified model rises from roughly 56 metres to roughly 224 metres — about four times as far.
Speed is not merely how quickly the train is travelling. Speed determines how much future railway the train needs.
That is a useful way to understand signalling. A fast-moving train has a larger claim on the track ahead because it requires more distance in which to reduce its motion safely.
The same square appears in energy
Kinetic energy is:
Ek = ½mv²
Again we see v².
Double speed and kinetic energy becomes four times as large for the same mass.
Braking therefore has two square-law stories happening at once:
higher speed → much longer stopping-distance requirement higher speed → much more kinetic energy to remove or recover
This is one reason maximum speed alone tells us remarkably little about how good a metro railway is.
Prompt 3 — What Does Mass Change?
This part contains a subtle mathematical lesson.
If we assume a train can maintain exactly the same deceleration b regardless of its mass, then the idealised stopping-distance formula:
s = v² / (2b)
does not explicitly contain mass.
It might therefore look as though a heavily loaded train and an empty train have identical braking problems.
That conclusion would be too simple.
The force required to produce a chosen deceleration is:
F = ma
If mass rises while desired deceleration stays the same, braking force must rise proportionally.
Kinetic energy also rises with mass:
Ek = ½mv²
So a heavier train at the same speed carries more kinetic energy that must be handled by regenerative and/or friction braking systems.
Mass may disappear from the simplest stopping-distance equation only because its effect has been hidden inside the assumption that the same deceleration can always be produced.
Real engineering must deal with the force, heat, electrical capacity, adhesion and component limits required to make that assumption true enough for safe operation.
Prompt 4 — Why Gradient, Adhesion and Passenger Comfort Change the Braking Curve
A railway is not a perfectly flat mathematical line.
If a train is travelling on a slope, gravity contributes a component along the track:
Fg = mg sin θ
On an uphill approach, gravity assists the reduction of forward speed. On a downhill approach, gravity works against braking.
A simplified effective deceleration idea might therefore be written:
beffective ≈ bbrake ± g sin θ
The exact railway calculation is more sophisticated, but the lesson is clear:
A braking curve that is appropriate on level track is not automatically the same braking curve required on a gradient.
Adhesion: force has to reach the rail
Braking force ultimately has to be transmitted through the train–rail interface. A simple friction-limited model looks like:
Fmax ≈ μN
where μ represents an effective coefficient of adhesion and N the normal force.
Again, real rail adhesion is a specialist subject and cannot be reduced to one classroom coefficient. The conceptual point is that the train cannot demand unlimited braking force merely because an equation asks for it.
Physical contact imposes constraints.
Passenger comfort: deceleration is not the whole story
Suppose a train instantly changes from no braking to strong braking. Even if the final deceleration is safe, passengers feel the sudden transition.
That brings us to jerk:
j = da/dt
Good passenger braking therefore tends to use a shaped curve rather than an abrupt mathematical step.
Conceptually:
0 braking → braking builds → useful deceleration is maintained → braking eases near stop → 0
The final easing is particularly important to what passengers feel. A poorly controlled final stop can produce a noticeable lurch even when the train stops at the correct location.
This is another RFE lesson: the railway is not moving abstract masses. It is moving people who may be standing, carrying bags, using mobility aids or holding a child’s hand.
Prompt 5 — How Does Signalling Decide When Braking Must Begin?
Now braking connects back to the whole railway.
The train cannot choose its braking point only by measuring the distance to the next station. It must also respect the safe movement authority available ahead.
Singapore’s North-South and East-West Lines operate with Communications-Based Train Control. LTA explains that the upgraded CBTC signalling tells trains when to stop or move and enables trains to operate more closely together than the earlier system. SMRT states that on the NSEWL, acceleration, braking and coasting are controlled by the Automatic Train Operation system.
Mathematically, this means the train is not following one speed limit. It is following a moving permitted-speed envelope.
We can represent that conceptually as:
vtrain(x,t) ≤ vpermitted(x,t)
and the permitted curve itself reflects the need to remain able to reach a safe target state.
A braking curve looks backwards from the future
Suppose the train must reach a target point at speed vtarget. A simplified backward-looking relation is:
vallowable² ≈ vtarget² + 2bd
If the target speed is zero:
vallowable ≈ √(2bd)
As remaining distance d becomes smaller, allowable speed falls.
This produces one of the most beautiful ideas in railway mathematics:
The train’s present speed is constrained by a future state that has not happened yet.
The train begins braking now because the mathematics says where it must be later.
Service braking and emergency intervention are different jobs
Normal passenger braking is designed around regular operation: safe, smooth, repeatable stopping while respecting the timetable and signalling system.
Emergency intervention exists for a different reason: protecting safety when normal assumptions are no longer sufficient.
SMRT’s public commuter guidance notes that station emergency-stop equipment can prevent a train at a platform from moving off and can be used to stop an approaching train in an emergency. We do not need operational thresholds or system internals to understand the important principle: railway safety is layered rather than dependent on a single everyday braking command.
LTA has also publicly described extensive brake-system testing for C151A trains before service, noting that the brake system is one of the most safety-critical train subsystems because it allows the train to stop within a safe stopping distance.
Prompt 6 — How Does an MRT Stop at the Right Place, Not Merely Somewhere in the Station?
Stopping safely is only half the platform problem.
The train also needs to stop where the station expects it.
Modern Singapore MRT stations use platform screen doors or platform barriers on many lines. LTA’s newer rail-system contracts explicitly pair CBTC signalling with platform screen-door systems. That gives us a clue to the systems problem: train movement and the platform interface have to agree about where “stopped” is.
Mathematically, we can define a stopping-position error:
eposition = xactual − xtarget
The control system wants:
eposition → 0
while velocity simultaneously goes to:
v → 0
This is a two-part target:
correct position + zero speed
Arriving at zero speed ten metres too early would still be a poor stop. Arriving at the correct location with non-zero speed would be worse.
So the controller progressively corrects the train trajectory as it approaches the platform.
Why the final metres matter disproportionately
At high speed, the main problem is shedding kinetic energy while staying inside a safe braking envelope.
Near the platform, the problem changes. Speed is already low. Positional precision and passenger comfort become increasingly important.
The control objective can therefore change emphasis during one braking event:
far from stop: energy reduction + safe speed envelope near stop: position accuracy + smooth final deceleration
The train is solving a different version of the same problem as the seconds pass.
Prompt 7 — Where Does the Train’s Energy Go When It Brakes?
This is where braking stops being merely a safety story and becomes an energy story.
A moving train carries kinetic energy:
Ek = ½mv²
When the train stops, that kinetic energy cannot simply vanish.
Some braking systems dissipate motion as heat through friction. Electric trains can also use regenerative braking, in which the traction motors operate in a generating mode and convert part of the train’s kinetic energy back into electrical energy.
LTA states that Downtown Line trains use energy-efficient regenerative brakes that recover kinetic energy and convert it for other uses. SMRT has likewise described regenerative braking on its trains, where electricity generated while slowing can be returned to the power distribution system.
A simple energy-recovery model
If the train begins braking with kinetic energy:
Ek = ½mv²
and an overall fraction η can be usefully recovered, then a simplified recovered-energy model is:
Erecovered = η(½mv²)
The unrecovered portion must be handled elsewhere in the braking and electrical system.
η is not one universal number. It depends on the equipment and on whether the electrical network can make useful use of the regenerated energy at that moment.
Now one train’s braking can matter to another train
Suppose Train A is braking while Train B is accelerating elsewhere on the same traction-power system.
Train A is potentially producing electrical energy while Train B needs electrical energy.
The timetable can therefore affect the opportunities for regenerated energy to be used.
SMRT’s 2023/24 sustainability reporting described Green CBTC work using advanced coasting, adaptive slow running and coordinated arrivals and departures to maximise regenerative braking. Its current decarbonisation information says Phase 1 of Green CBTC is delivering around 17 GWh of annual energy savings, with the overall project aiming for up to a 15% reduction in traction energy on the NSEWL.
So the braking problem becomes:
safe stop + correct stop + comfortable stop + timetable-compatible stop + energy-aware stop
That is much richer than “apply the brakes”.
Prompt 8 — Why a Safe Stop Is a System, Not a Single Brake
By now we can see that braking is distributed across the railway.
There is braking hardware on the train. SMRT’s 2026 public engineering exhibition, for example, highlighted the tread brake unit as a component that enables trains to stop at stations.
There is propulsion and regenerative control.
There is Automatic Train Operation controlling acceleration, braking and coasting on the NSEWL.
There is CBTC signalling controlling how the train is permitted to move relative to the railway around it.
There are station systems that expect the train to arrive at the correct location.
There are emergency safety layers for unusual events.
There are maintenance and testing regimes because a mathematical safety assumption is only useful if physical equipment continues to perform as expected.
LTA’s published account of C151A safety testing illustrates this systems philosophy. Brake subsystems were type-tested; brake-shoe material was tested on a dynamometer using simulated train run profiles; completed trains then underwent further functional, performance and integrated testing before mainline operation.
A safe braking equation is not merely written in software. It has to remain true in steel, electronics, wheels, rails, power systems, sensors, maintenance and operations.
A Worked Braking Journey: From 80 km/h to a Platform Stop
Let us assemble the concepts into one educational example. Again, every numerical operating value here is illustrative rather than a published Singapore MRT operating parameter.
Imagine a train travelling at:
u = 80 km/h = 22.22 m/s
The target station stop is ahead.
Stage 1 — The target exists before braking begins
The system already knows the train must eventually satisfy:
x = xstation v = 0
This future boundary condition shapes the permitted speed before the station.
Stage 2 — Coasting may begin
If the timetable has enough time and the route profile permits it, the train may stop applying traction before active braking is needed.
Speed begins to fall naturally from resistance and gradient effects while little or no traction power is being used.
This converts timetable flexibility into energy savings.
Stage 3 — Main controlled braking
Suppose, only for illustration, the average braking magnitude through this phase behaves near 1.1 m/s².
The idealised full-speed stopping distance would be roughly 224 metres. But a real controller uses a time-varying braking profile rather than treating the event as one perfectly constant deceleration.
The train’s allowed state can be represented as:
vactual(t) ≤ vpermitted(t)
while maintaining sufficient future stopping capability.
Stage 4 — Low-speed approach
As velocity becomes smaller, the remaining kinetic energy falls sharply because:
Ek ∝ v²
The control problem increasingly becomes one of stopping-position accuracy and comfort.
Stage 5 — Final stop
The required end state is:
v = 0 eposition ≈ 0
and braking force is eased so passengers do not experience an unnecessary final jolt.
What looked from the platform like a single event — “the train stopped” — was actually a trajectory created from a chain of future constraints.
Why Braking Earlier Can Sometimes Be Faster for the Railway
This sounds contradictory.
How can braking earlier possibly make a railway faster?
Because “faster railway” does not mean “one train stays at maximum speed for the longest possible time”.
Suppose a train rushes towards a slower train ahead and then has to brake sharply. It may consume more energy, create greater control variation and still arrive at the same signal constraint.
A more intelligent trajectory may coast or reduce speed earlier, reach the same constrained point at nearly the same time, and do so with less energy and smoother headway behaviour.
This is the railway equivalent of arriving at a red traffic light after racing towards it.
If the future is already constrained, maximum speed in the present may buy nothing.
The best control decision is therefore contextual rather than absolute.
The Braking Curve Is Also a Time Curve
Braking determines more than stopping distance.
It determines arrival time.
If a train begins braking too early and too strongly, it may arrive late.
If it begins too late for the required trajectory, it may violate the permitted speed profile.
So braking is part of timetable control.
Let desired station-arrival time be:
Tschedule
and actual arrival:
Tactual
Timing error is:
etime = Tactual − Tschedule
The movement controller is therefore simultaneously managing:
position error speed error time error energy use comfort constraints safety constraints
One braking event can affect all of them.
What Happens When the Train Ahead Brakes?
Now place our train back into the railway.
Train A is ahead.
Train B follows.
If Train A slows, the movement authority or safe speed available to Train B may also change.
Train B therefore does not merely respond to stations. It responds to the evolving state of the railway.
A conceptual spacing variable is:
dAB(t) = xA(t) − xB(t)
The safe requirement is not simply:
dAB > 0
It is that spacing remains compatible with the safe braking and protection model.
So two trains can be physically far apart while a following train is already reducing speed because the future safe envelope is shrinking.
This is another reason passengers sometimes feel a train slow between stations even when there seems to be no obvious reason. The visible station is only one of the train’s constraints.
What Mathematics Does a Brake Engineer Actually Care About?
Not one equation.
A braking system has to reconcile a family of relationships.
Velocity: v = dx/dt Acceleration: a = dv/dt Jerk: j = da/dt Idealised stopping distance: s = v²/(2b) Kinetic energy: Ek = ½mv² Braking force: F = ma Gradient component: Fg = mg sin θ Adhesion-limited force concept: Fmax ≈ μN Mechanical power: P = Fv Recovered-energy concept: Eregen = η(½mv²) Position error: eposition = xactual − xtarget Timing error: etime = Tactual − Tschedule
And all of these sit inside constraints imposed by the route, signalling system, equipment and operating rules.
This is why engineering mathematics often feels different from examination mathematics.
The exam asks:
Find the stopping distance.
The railway asks:
Find a stopping trajectory that is safe, physically achievable, comfortable, accurately positioned, timetable-compatible, energy-aware, robust to uncertainty and repeatable all day.
The RFE Returns — What Is a Good Stop?
At first we might define a good stop as:
vfinal = 0
That is insufficient.
A train could stop too early.
It could stop too late.
It could stop with an uncomfortable final lurch.
It could waste recoverable energy.
It could arrive at the wrong time and reduce the headway behind it.
It could satisfy normal conditions but leave too little robustness when conditions vary.
A more complete definition is:
Good stop = safe speed reduction + sufficient stopping margin + physically achievable braking + correct platform position + passenger comfort + timetable compatibility + energy efficiency + reliability
This is the real reason braking mathematics exists.
It turns motion into a controlled promise:
The train will be where it is supposed to be, at the speed it is supposed to have, when the railway needs it there.
Conclusion — A Train Starts Stopping Before You Think It Does
The next time your MRT begins slowing before a station, notice what is happening.
Its velocity is falling.
Its kinetic energy is being removed and, where possible, partly recovered.
The route gradient is affecting the force balance.
The control system is keeping the train within a permitted speed envelope.
The train is moving towards two targets at once:
the right place and zero speed
At the same time, the railway is protecting the headway behind it and the schedule ahead of it.
The braking event therefore begins conceptually before the brakes themselves are applied.
It begins when the future state becomes a constraint on the present.
The train does not stop because it has reached the station. It reaches the station correctly because it started solving the stop earlier.
That is the mathematics hidden beneath an ordinary MRT ride.
And once we understand braking, we can move one level outward again — to the mathematics of 100-second headways, where the stopping capability of one train becomes part of the capacity of the entire line.
Continue the MRT Mathematics Series
Sources and Further Reading
- Land Transport Authority — Rejuvenating the Railway: NSEWL Core Systems
- SMRT — Automatic Train Operation, Braking, Coasting and Green CBTC
- SMRT — Decarbonisation and Green CBTC
- SMRT — Green CBTC Next Gen Energy Savings
- Land Transport Authority — Downtown Line Regenerative Braking
- SMRT — Engineering Your Commute and Tread Brake Units
- Land Transport Authority — Testing of Safety Features for C151A Trains
- Land Transport Authority — Cross Island Line CBTC and Platform Screen Door Systems
Return to the complete MRT Mathematics hub: How MRT Works | It’s Mathematics.