How MRT Power Supply Works Using Mathematics

An MRT train looks like a transport machine.

Electrically, it is also a moving load.

When the train accelerates, its traction motors can demand large electrical power. When it cruises, demand changes. When it coasts, traction demand can fall sharply. When it brakes regeneratively, the direction of electrical power can reverse and part of the train’s kinetic energy can return to the railway electrical system.

Now add another train.

Then another.

Some are accelerating. Some are braking. Some are stopped in stations. Some are running uphill. Some are heavily loaded. Some are drawing auxiliary power while waiting. The physical locations of these electrical loads are changing every second.

The railway power system therefore has to solve a problem that an ordinary building electrical supply does not:

How do we keep voltage, current, power flow, protection and redundancy inside a safe operating envelope while megawatt-scale electrical loads move through the network?

This is why an MRT line can be understood as a moving electrical grid.

Singapore gives us several public examples of this infrastructure. LTA’s completed North-South and East-West Line power renewal replaced large quantities of cables, transformers, switchboards and substation equipment, while adding real-time condition monitoring and fault-isolation capability. In February 2026, LTA, SMRT and SBS Transit announced further measures to strengthen the North East Line and Sengkang-Punggol LRT power system, including additional traction power substations and earlier renewal of the NEL power system.

For the future Jurong Region Line, LTA has publicly described a system containing 66 kV and 22 kV AC networks feeding a 750 V DC traction supply, together with emergency traction tripping and auxiliary systems. This makes the electrical conversion chain unusually visible to the public.

This article continues the eduKateSG MRT mathematics cloud. The permanent synthesis page is How MRT Works | It’s Mathematics. Closely related pillars include How MRT Energy Use Is Optimised Using Mathematics, How MRT Network Resilience Is Measured Using Mathematics and How MRT Predictive Maintenance Works Using Mathematics.

The RFE — Why Does MRT Power Mathematics Exist?

The weakest possible objective would be:

supply as much electrical power as possible

That would be meaningless.

The railway needs neither maximum voltage nor maximum current nor maximum power at every moment.

It needs the right electrical state to support the transport job.

supply enough traction power
+
keep voltage inside usable limits
+
limit current and thermal stress
+
convert and distribute electricity efficiently
+
receive regenerative energy when possible
+
isolate faults before they spread
+
retain redundancy and recovery capability
+
keep stations, signalling and auxiliary systems supplied
+
maintain the electrical assets over decades

The Reason for Existence is therefore not electricity for its own sake.

It is continuity of movement.

The power system exists so that electrical energy can appear at the correct place, at the correct voltage, in the correct amount, with enough protection and redundancy that train movement remains safely available.

Prompt 1 — How Does Electricity Reach an MRT Train?

Electricity typically reaches an electric railway through several stages.

utility grid
→ high-voltage intake
→ railway distribution
→ transformer
→ rectifier where DC traction is used
→ traction substation
→ conductor system
→ train
→ traction converters and motors

The exact arrangement differs by line and railway technology.

The future Jurong Region Line gives a public Singapore example. LTA states that its power-supply system includes:

66 kV AC systems,
22 kV AC systems,
a 750 V DC traction power-supply system,
an Emergency Traction Tripping System,
and associated auxiliary systems.

This shows the hierarchy clearly.

High-voltage AC is useful for distributing large power because a higher voltage allows lower current for the same power.

Electrical power is:

P = VI

So:

I = P/V

Suppose we need to transmit 10 MW.

At 1 kV:

I = 10,000,000 / 1,000
  = 10,000 A

At 20 kV:

I = 10,000,000 / 20,000
  = 500 A

The same power can be transported with far lower current at higher voltage.

Why lower current matters

Resistive power loss in a conductor is:

Ploss = I²R

If current doubles, resistive loss becomes four times as large for the same resistance.

This is why electricity is distributed at high voltage before being transformed closer to the load.

For DC traction, rectifiers convert AC into DC:

AC electrical input
→ transformer sets suitable voltage
→ rectifier converts AC to DC
→ DC traction network supplies train

The train then contains its own power electronics to regulate motor voltage, current, frequency and torque.

The railway therefore contains several electrical worlds at once:

utility-grid AC
railway high-voltage AC
traction DC or other line-specific traction supply
vehicle power electronics
three-phase motor currents
low-voltage control and auxiliary systems

The passenger sees one train.

The electrical engineer sees a chain of voltage domains.

The train does not “take electricity from the grid”. Electricity is transformed, protected, distributed, converted and controlled through several layers before motor torque appears at the wheel.

Prompt 2 — Why Does Voltage Drop When a Train Demands Power?

No electrical conductor is perfectly lossless.

A simple conductor has resistance:

R = ρL/A

where:

ρ is electrical resistivity,
L is conductor length,
A is cross-sectional area.

When current flows through resistance, voltage drop is:

ΔV = IR

Suppose a train demands electrical power P at local voltage V.

Current is approximately:

I = P/V

More traction power generally means more current.

More current creates larger IR voltage drop and I²R heating.

A simplified traction-feeder example

Imagine a fictional 750 V DC traction section.

Equivalent electrical path resistance from substation to train is:

R = 0.02 Ω

Train demands:

P = 2.0 MW

If voltage were exactly 750 V, approximate current would be:

I ≈ 2,000,000/750
  ≈ 2,667 A

Voltage drop:

ΔV = IR
   ≈ 2,667 × 0.02
   ≈ 53.3 V

The local voltage would therefore be lower than the ideal source value in this simplified calculation.

But this calculation contains a feedback.

If a converter attempts to maintain constant power while voltage falls:

I = P/V

current rises.

Higher current creates still more voltage drop:

lower V
→ higher I for same P
→ larger IR drop
→ still lower V

Real traction equipment therefore contains voltage and current limits, and the electrical network is designed so operating conditions remain inside validated envelopes.

Why substations are distributed

If all trains were supplied from one distant electrical point, conductor resistance and fault consequences would become difficult to manage.

Distributed traction substations shorten electrical paths and share the load.

For a train between two substations, current can be supplied from more than one direction depending on the system topology and instantaneous electrical state.

Pload = V(Ileft + Iright)

The division depends on network resistance, source conditions and system architecture.

In February 2026, LTA said it intends to augment NEL and SPLRT traction power capacity through additional traction power substations. The public engineering reason is clear: more distributed supply capability can strengthen the power system’s ability to support required loads and improve resilience.

Voltage is not simply “present” on a railway. It is continuously negotiated through distance, resistance, moving load and the number of electrical paths available to support that load.

Prompt 3 — Why Is an MRT Train a Moving Electrical Load?

A station, office or factory usually stays in one place.

A train does not.

Let train i have position x_i(t).

Its power demand is:

Pi(t) = Vi(t) Ii(t)

But P_i(t) depends on its movement state:

Pi(t)
= f(speed,
acceleration,
gradient,
train mass,
resistance,
auxiliary load,
regeneration)

The train’s electrical location also changes.

electrical load = [xi(t), Pi(t)]

For n trains:

Load field(t)
= { [x1(t),P1(t)], [x2(t),P2(t)], ... [xn(t),Pn(t)] }

This is why the railway is a moving grid.

Mechanical demand becomes electrical demand

Train mechanics give:

Ftraction
= meff a
+ R(v)
+ mg sinθ

Mechanical traction power is:

Pmech = Ftraction v

If traction-chain efficiency is η:

Pelectrical ≈ Pmech/η + Paux

where P_aux includes auxiliary loads such as ventilation, lighting and onboard equipment.

This shows why accelerating at higher speed can require very high power.

Suppose traction force remains 100 kN.

At 10 m/s:

P = 100,000 × 10
  = 1.0 MW

At 20 m/s:

P = 2.0 MW

The same force at double speed requires double mechanical power.

Many trains create diversity

Not every train demands maximum power simultaneously.

At one instant:

Train A accelerating: +2.2 MW
Train B cruising:     +0.8 MW
Train C coasting:     +0.2 MW auxiliaries
Train D braking:      −1.4 MW regenerative
Train E in station:   +0.15 MW auxiliaries

Ignoring losses and network limits, net instantaneous train demand would be:

Pnet
= 2.2 + 0.8 + 0.2 − 1.4 + 0.15
= 1.95 MW

The railway power system therefore depends on coincidence.

Peak demand occurs when many high-power events overlap.

Regenerative opportunity occurs when braking and receptive demand overlap.

The timetable changes the electrical problem.

train timetable
→ positions and motion states
→ electrical load field
→ current and voltage distribution
→ power losses and peaks

This is why timetable, energy and power-supply engineering cannot be fully separated.

The railway timetable is also an electrical schedule, because every acceleration and braking event moves a large load or generator through the power network.

Prompt 4 — How Does Regenerative Braking Reverse the Power Flow?

A moving train contains kinetic energy:

Ek = ½mv²

Braking must remove that energy.

Traditional friction brakes convert much of it into heat.

Regenerative braking allows traction motors to operate as generators, converting part of mechanical energy back to electrical energy.

train kinetic energy
→ motor/generator
→ vehicle power electronics
→ traction electrical network
→ receptive train or other usable load

LTA describes regenerative braking on Downtown Line trains, and its Green Mark material notes that the Circle Line can reuse energy from one braking train to power another.

Power sign convention

Define train traction power as:

Ptrain > 0  during traction
Ptrain < 0  during regeneration

For a local electrical section:

Psection(t)
= Σi Ptrain,i(t)
+ Paux(t)
+ Ploss(t)

If one train brakes at −2 MW while another accelerates at +2 MW, ideal timing suggests strong opportunity for local energy exchange.

But real recovery is limited by:

conversion efficiency,
electrical connectivity,
instantaneous receptive demand,
voltage limits,
network losses,
and line-specific equipment capability.

Regenerative receptivity

Let available regenerative power be P_regen.

Let local receptive demand be P_load.

A simplified directly usable regenerative amount is:

Pusable,regen
≤ min(Pregen, Pload)

If a braking train can return 2 MW but the local network can use only 0.5 MW at that instant, all 2 MW cannot automatically become useful recovered traction energy.

This is why train timing matters.

The energy pillar explains how Green CBTC and coordinated speed profiles can reduce traction energy. The power-supply layer explains the network through which those gains are physically realised.

Voltage rise during regeneration

During traction, current draw tends to create voltage drop.

During regeneration, current injection can push local voltage upward.

The power system must therefore manage both directions:

traction
→ current drawn
→ local voltage tends downward

regeneration
→ current injected
→ local voltage tends upward

This makes the traction network genuinely bidirectional in energy terms, even if individual equipment and pathways have specific constraints.

A braking train is not merely a load that became smaller. For a short period, it can become an electrical source moving through the same network.

Prompt 5 — How Do Peak Demand and Train Frequency Interact?

Increase train frequency and more trains are active.

More active trains can increase electrical demand.

But not proportionally in every moment because their acceleration, cruising and braking phases differ.

Let train headway be H.

Train flow is:

q = 3600/H

Shorter headway generally means more train movements per hour.

For average energy per train journey E_trip, average traction energy rate scales roughly with:

Ehour ≈ q Etrip

But electrical infrastructure must also survive the peak, not merely the hourly average.

Define:

Ppeak = max_t Psystem(t)

Two timetables can have the same number of trains and the same total energy but different P_peak.

A simple coincidence example

Suppose four trains each need 2 MW while accelerating.

If they accelerate at exactly the same time:

Ppeak = 4 × 2 MW
      = 8 MW

If acceleration phases are staggered and only two overlap:

Ppeak ≈ 4 MW

The same four trains can create a very different peak.

Add regenerative braking and the scheduling problem becomes richer.

accelerating train
+
braking train
→ lower net substation demand if timing and network receptivity align

This creates a power-aware timetable objective.

minimise
energy + peak demand + journey-time penalty + delay risk

subject to
safe train separation
passenger capacity
station dwell
fleet
power-system constraints

Again, the point is not that operators use this exact public equation.

It is that power supply and timetable are coupled by the motion of trains.

Capacity needs electrical capacity

A signalling upgrade can permit more trains per hour.

But if the power system cannot supply the additional simultaneous traction demand, technical headway capability does not automatically become usable transport capacity.

LTA’s description of NSEWL renewal makes this relationship explicit: higher rail capacity needs a power system capable of supporting the job, which was one reason the power supply system was comprehensively upgraded.

Rail capacity is not only trains per hour. It is trains per hour that the tracks, stations, signalling, fleet and electrical system can all sustain together.

Prompt 6 — How Does the Power System Prevent One Fault from Becoming a Line-Wide Failure?

Power networks are designed not only to deliver energy but to contain abnormal electrical states.

Possible electrical faults include short circuits, insulation failures and equipment failures.

The public article should not reproduce detailed protection settings or fault-clearing procedures.

The mathematical principle is enough.

For a fault current I_f, protection must distinguish:

normal load
vs
abnormal fault state

and disconnect the smallest justified affected electrical region quickly enough to protect equipment and prevent propagation.

Sectionalisation

Imagine a simple line divided into electrical sections:

Section A | Section B | Section C | Section D

If a fault occurs in Section C, the ideal protection objective is not:

turn off the entire railway

It is closer to:

detect C
→ isolate C
→ preserve A, B and D where safe and system architecture permits
→ restore or reconfigure

This is fault containment.

LTA says the NSEWL power renewal replaced the former touch-voltage protection arrangement with Voltage Limiting Devices designed to isolate power faults and prevent line-wide disruptions. LTA also describes an intelligent Fault Isolation System that helps locate cable faults for faster isolation and recovery.

The public lesson is powerful:

Electrical resilience depends on making a fault small enough that the rest of the railway still has a future state.

Fault location as an inverse problem

Suppose voltage and current are measured at several points.

A fault changes those measurements.

The system observes:

Y = [V1,V2,...,I1,I2,...]

and asks which fault location x_f and fault type best explain them:

find xf such that
model(xf) ≈ observed electrical disturbance

This is an inverse problem.

The better the monitoring and model, the faster engineers can narrow the physical search.

Protection must discriminate

Too sensitive:

normal transient
→ unnecessary trip
→ avoidable service loss

Too insensitive:

real fault
→ delayed isolation
→ larger damage or propagation

This is the same false-positive/false-negative geometry that appears in predictive maintenance, now operating inside electrical protection.

Prompt 7 — What Does Electrical Redundancy Actually Mean?

Redundancy means the system has more than one viable way to preserve a necessary function.

But duplication alone is not enough.

Two power feeds are weak redundancy if both depend on the same upstream failure.

Two substations are weak redundancy if neither can carry enough load after the other is unavailable.

So resilience needs both independence and spare capacity.

Parallel reliability

For two ideal independent parallel sources with reliabilities R₁ and R₂:

Rparallel
= 1 − (1−R1)(1−R2)

If each has reliability 0.99:

Rparallel
= 1 − 0.01²
= 0.9999

But this improvement assumes independence.

If both feeds fail when one upstream switchboard fails, the independence assumption is false.

Capacity redundancy

Suppose two substations each have usable capacity C.

Normal total demand is D.

If one substation fails, remaining supply is only C.

Full service can continue only if:

D ≤ C

If:

C < D ≤ 2C

the intact system may support full normal demand when both substations are available but not after one is lost.

Resilience therefore requires more than installed total capacity.

It asks:

What demand can the network support after a component is unavailable?

This can create degraded operating modes.

normal power network
→ one element lost
→ remaining electrical capacity reduced
→ train service may need lower power demand or different operating pattern
→ passengers are managed through a degraded but viable mode

In February 2026, LTA announced plans to upgrade the existing power intake at the NEL depot in Sengkang and augment traction power capacity through additional substations. The same announcement said the NEL power-supply renewal would be brought forward to begin in 2026.

This is resilience being built before the next failure.

Power resilience is network resilience

Loss of traction power affects more than one train.

It can change:

train movement,
headway,
platform queues,
route choice,
alternative-line loading,
service-recovery requirements.

So the power system sits underneath the network-resilience pillar.

A redundant cable is not resilience until the surviving electrical network can still carry the transport function that passengers require.

Prompt 8 — How Does the MRT Know Its Power System Is Healthy?

Electrical assets degrade.

Cables age.

Insulation deteriorates.

Switchgear accumulates operating cycles.

Connections heat and cool repeatedly.

Transformers experience thermal and electrical stress.

A modern railway therefore monitors electrical condition rather than waiting for failure.

LTA says the renewed NSEWL power system constantly monitors traction-cable insulation so operators can assess condition and intervene pre-emptively when appropriate.

This is condition-based maintenance applied to electrical infrastructure.

Insulation condition

An insulation-health indicator can be represented:

Hins(t)

The system looks for change:

ΔH = H(t) − Hbaseline

or rate of change:

dH/dt

A slow trend can be more important than one noisy measurement.

Thermal loading

Electrical heating is strongly related to I²R.

A simple thermal-state model is:

Cth dT/dt
= I²R − h(T−Tamb)

where:

C_th is effective thermal capacity,
h is an effective cooling coefficient,
T_amb is ambient temperature.

High current over time raises conductor or equipment temperature.

Thermal history therefore matters, not only instantaneous current.

Load forecast versus actual load

Suppose predicted substation load is P̂(t).

Observed load is P(t).

eP(t) = P(t) − P̂(t)

Persistent positive error can indicate:

higher-than-expected train demand,
changed timetable or passenger load,
lower regenerative recovery,
increased electrical losses,
or model drift.

The power model should therefore return to the same learning loop as the rest of the MRT:

forecast
→ operate
→ measure voltage, current and condition
→ compare
→ diagnose
→ maintain, reinforce or recalibrate
→ operate again

That is World Return in electrical form.

A Complete Fictional MRT Power-Flow Example

Consider a simplified DC traction section supplied by two substations.

These values are fictional and are not Singapore MRT design parameters.

Step 1 — Define the trains

Train A accelerating: +1.8 MW
Train B accelerating: +1.4 MW
Train C braking:      −1.0 MW
Auxiliary section load: +0.3 MW

Ignoring losses:

Pnet
= 1.8 + 1.4 − 1.0 + 0.3
= 2.5 MW

Step 2 — Estimate current

At an illustrative local voltage of 750 V:

I ≈ 2,500,000/750
  ≈ 3,333 A

Suppose equivalent network resistance is 0.012 Ω.

Resistive loss:

Ploss = I²R
≈ 3,333² × 0.012
≈ 133 kW

Approximate voltage drop:

ΔV = IR
≈ 3,333 × 0.012
≈ 40 V

The simplified calculation shows why high current matters even when conductor resistance is small.

Step 3 — Regeneration disappears

Train C finishes braking.

Now:

Pnet
= 1.8 + 1.4 + 0.3
= 3.5 MW

At 750 V:

I ≈ 4,667 A

With the same 0.012 Ω:

Ploss
≈ 4,667² × 0.012
≈ 261 kW

Demand increased by 40 per cent, but resistive loss almost doubled because current is squared.

Step 4 — Add another accelerating train

Train D begins accelerating at +1.5 MW.

Pnet = 5.0 MW

At 750 V:

I ≈ 6,667 A

Resistive loss:

Ploss
≈ 6,667² × 0.012
≈ 533 kW

The example is intentionally simple, but it shows why coincident acceleration creates electrical peaks and why local voltage and thermal limits matter.

Step 5 — Stagger Train D

Suppose the timetable allows Train D to begin strong acceleration after Train A has reduced traction demand by 1.0 MW.

Peak becomes:

Ppeak ≈ 4.0 MW instead of 5.0 MW

The trains still complete the same broad sequence, but electrical coincidence has improved.

Whether that timing change is acceptable depends on journey time, headway and operational constraints.

Step 6 — One supply path is lost

Suppose two substations normally share the 4.0 MW peak.

Each can supply an illustrative 2.5 MW under the scenario.

With both available:

total usable capacity = 5.0 MW

If one is unavailable:

remaining capacity = 2.5 MW
required peak = 4.0 MW

Full normal electrical demand cannot be supported in this fictional case.

A degraded operating plan may therefore reduce simultaneous traction demand.

electrical failure
→ lower usable power capacity
→ change train operating pattern
→ lower line capacity or longer journey
→ protect a viable transport core

This is exactly where electrical resilience becomes passenger resilience.

The Power-Supply Deletion Tests

Remove voltage

Power has no electrical potential through which current can be driven.

Remove current limits

Conductors and equipment are assumed able to carry unlimited electrical load without heating.

Remove conductor resistance

Voltage drop and I²R losses disappear unrealistically.

Remove train position

Every electrical load is treated as though it remains at one fixed point.

Remove the timetable

Electrical peaks and regenerative overlap cannot be connected to actual train movements.

Remove regeneration

Every braking event is treated only as dissipated energy rather than a possible electrical source.

Remove fault isolation

One local electrical fault can unnecessarily remove a much larger part of the railway.

Remove redundancy

Every essential load depends on one electrical path.

Remove spare capacity

Backup equipment exists nominally but cannot carry the transferred load.

Remove condition monitoring

Cable and equipment deterioration becomes visible only when a fault reaches operation.

The power system is not simply wires and substations. It is a continuously observed network of electrical limits, moving loads, protection boundaries and recovery paths.

The MRT Power Paradoxes

Paradox 1 — Higher voltage can reduce energy loss

For the same transmitted power, higher voltage means lower current, and I²R losses fall strongly.

Paradox 2 — A braking train can become a power source

Regenerative braking reverses electrical power flow.

Paradox 3 — More frequent trains can lower some energy waste yet raise peak power

More trains create more regenerative matching opportunities but also more coincident traction events.

Paradox 4 — A power system can have enough average energy but insufficient instantaneous power

Energy over an hour and peak megawatts are different constraints.

Paradox 5 — A backup can exist and still provide weak resilience

If it shares the same failure mode or lacks capacity, nominal redundancy does little.

Paradox 6 — More sensitive protection can reduce reliability

If normal transients are misclassified as faults, unnecessary trips create avoidable service loss.

Paradox 7 — Electrical infrastructure can limit rail capacity without passengers ever seeing it

A signalling system may permit more trains than the power network can sustain simultaneously.

Paradox 8 — Maintenance closures can improve future availability

Taking part of the railway out of service deliberately creates short-term inconvenience to reduce longer-term failure risk.

The Electrical Failure Cascade

A power problem can travel through the MRT in several ways.

Electrical event	Immediate effect	Passenger shadow
Local traction fault	One electrical section constrained	Train movement reduced or rerouted
Substation unavailable	Reduced local supply capacity	Lower service capability if remaining network cannot absorb load
Excessive voltage drop	Train traction performance constrained	Longer running time, lower recovery margin
Protection trip	Electrical section isolated	Service interruption until safe reconfiguration
Fault not isolated	Damage or trip propagates farther	Wider service loss
Regeneration not receptive	Less recovered energy	Higher net electrical demand and energy use
High simultaneous traction load	Peak current and voltage drop rise	Power can become a capacity constraint
Condition-monitoring blind spot	Degradation reaches failure unnoticed	Reactive rather than planned disruption

The failure mechanism should therefore be traced from source to receiver:

electrical trigger
→ affected section
→ remaining power capacity
→ train movement constraint
→ headway and timetable change
→ passenger queue and route change
→ recovery

This keeps electrical engineering connected to the public transport job.

The Power-Supply Audit

What electrical load must be supported?
At what train positions and times?
What is the traction-voltage architecture of this line?
What high-voltage distribution feeds it?
Where does AC-to-DC conversion occur where applicable?
What is the expected peak rather than only average power?
How much current does that power imply?
What conductor resistance and voltage drop follow?
What thermal loading follows from I²R?
How does train headway change simultaneous load?
Which braking trains can supply regenerative energy to receptive loads?
What happens when no receptive load is available?
Which electrical section is the bottleneck?
What happens if one substation or feed is unavailable?
Does the backup have enough spare capacity?
Are the backup paths genuinely independent?
How are faults detected and localised?
How small a region can be isolated without unnecessary wider loss?
What condition-monitoring evidence exists?
What maintenance window is required?
What passenger service remains under a degraded electrical state?
How will residual train and passenger disruption be cleared after power is restored?
What measurements would prove the electrical model wrong?

How the Mathematics Grows from School to Research

Primary Mathematics and Science

energy and electricity,
rates,
simple ratios,
graphs of changing demand.

Secondary Mathematics and Physics

Ohm’s law,
power P=VI,
resistance and I²R heating,
series and parallel circuits,
energy and efficiency.

Junior College

calculus and power integration,
complex AC power,
electromagnetic induction,
differential equations and thermal models,
probability and reliability.

University and Research

power-system load flow,
traction-power simulation,
power electronics,
protection engineering,
optimal train control,
regenerative-energy management,
stochastic load forecasting,
reliability and resilience optimisation,
condition monitoring and digital twins.

The school equation P=VI survives all the way into a city-scale railway.

What changes is the number of moving variables attached to it.

The World Return — When Voltage, Current and Passenger Service Answer the Model

The electrical model predicts:

substation load,
local voltage,
peak current,
regenerative recovery,
cable temperature,
fault response,
and degraded capacity after component loss.

The railway returns measurements.

predicted voltage V̂(t)
observed voltage V(t)
error eV(t)=V(t)−V̂(t)

predicted power P̂(t)
observed power P(t)
error eP(t)=P(t)−P̂(t)

If observed voltage is repeatedly lower than predicted during peak service, the model may underestimate load, resistance or simultaneous acceleration.

If regenerative recovery is lower than predicted, braking overlap or network receptivity assumptions may be wrong.

If one cable’s insulation condition degrades faster than expected, maintenance planning should change.

If a fault-isolation event removes a larger area than anticipated, the resilience model needs review.

And after the electrical system is restored, passenger recovery still has to be measured.

power restored
≠
passenger journey restored instantly

Trains may be out of sequence.

Queues may have accumulated.

Alternative lines may still be crowded.

World Return therefore closes at the receiver, not at the switchboard.

The electrical system is recovered when usable train service returns—not merely when voltage returns to a cable.

RFE Return — What Does Good MRT Power Supply Owe the Passenger?

The passenger should not need to know which substation is feeding the train.

They should not need to calculate voltage drop.

They should not need to care whether one braking train is electrically helping another accelerate.

That complexity should disappear into a simple public experience:

train accelerates when required
+
train can brake safely
+
service frequency is electrically sustainable
+
power faults remain bounded where possible
+
assets are monitored and renewed
+
recovery preserves the passenger journey

The best railway power system is therefore not the one with the largest installed electrical capacity.

It is the one whose capacity, topology, protection, monitoring and recovery are properly matched to the movement system above it.

The RFE of MRT power is to make electricity boring: always sufficient, correctly controlled, intelligently reused, locally containable when faulted and almost invisible to the passenger.

Conclusion — The MRT Is Also an Electrical Network That Moves

A train leaves a station.

Current rises.

Voltage changes along the traction network.

Cables heat slightly.

A nearby train begins braking.

Its motors become generators.

Part of that energy flows back.

Another train accelerates and becomes the receptive load.

Farther along the line, another substation supplies a different part of the moving load field.

Protection systems watch for abnormal electrical states.

Condition monitoring watches insulation and equipment health.

The timetable decides when the next acceleration will occur.

The headway decides how many trains may occupy the electrical network within one hour.

The station decides when the train can depart and demand power again.

The power system is therefore never separate from operations.

grid intake
→ high-voltage distribution
→ traction substation
→ traction conductor
→ moving train load
→ motor force
→ train movement
→ timetable state
→ braking regeneration
→ electrical return
→ monitoring and protection
→ maintenance
→ reliable movement again

The passenger sees acceleration.

The power engineer sees current.

The control engineer sees voltage and limits.

The energy engineer sees regenerative opportunity.

The reliability engineer sees protection zones and redundancy.

The maintenance engineer sees insulation condition and thermal history.

The passenger sees the train leave.

An MRT power system works when thousands of moving electrical demands can appear, disappear and reverse direction without the electrical network losing the ability to keep the railway moving.

Key Equations

P = VI
Electrical power

I = P/V
Current required for a given power and voltage

R = ρL/A
Conductor resistance

ΔV = IR
Voltage drop

Ploss = I²R
Resistive power loss

Pmech = Ftraction v
Mechanical traction power

Pelectrical ≈ Pmech/η + Paux
Simplified train electrical demand

Ek = ½mv²
Train kinetic energy

Ptrain > 0 traction
Ptrain < 0 regeneration
Power-flow sign convention

Psection = ΣPtrain + Paux + Ploss
Section power balance

Pusable,regen ≤ min(Pregen,Pload)
Simplified regenerative receptivity

q = 3600/H
Trains per hour from headway

Ehour ≈ qEtrip
Approximate hourly traction energy scaling

Ppeak = max_t Psystem(t)
Peak electrical demand

Rparallel = 1−(1−R1)(1−R2)
Ideal independent parallel reliability

Cth dT/dt = I²R − h(T−Tamb)
Simple electrical thermal model

eV(t)=V(t)−V̂(t)
Voltage model error

eP(t)=P(t)−P̂(t)
Power model error

Reader-safety note: All fictional voltages, resistances, train powers, capacities and fault examples in this article are educational abstractions unless explicitly attributed to public LTA material. This article does not reproduce sensitive traction-power topology, actual substation spacing, protection settings, switching procedures, fault-current thresholds or security-relevant electrical vulnerabilities.

Continue the MRT Mathematics Cloud

How MRT Works | It’s Mathematics — the permanent whole-system hub.
How MRT Timing Works Using Mathematics — the system of seconds.
How MRT Braking Works Using Mathematics — stopping distance and regenerative braking.
How MRT Headway Works Using Mathematics — train separation, frequency and throughput.
How MRT Station Dwell Time Works Using Mathematics — doors and passenger exchange.
How MRT Delays Propagate and Recover Using Mathematics — delay feedback and recovery.
How an MRT Timetable Is Built Using Mathematics — train events, fleet and operating day.
How MRT Energy Use Is Optimised Using Mathematics — traction, coasting and regenerative energy.
How MRT Passenger Capacity Is Calculated Using Mathematics — passengers per train and line capacity.
How MRT Routes and Transfers Are Optimised Using Mathematics — graph theory and route choice.
How MRT Network Resilience Is Measured Using Mathematics — redundancy and recovery.
How MRT Predictive Maintenance Works Using Mathematics — condition monitoring and remaining life.
How MRT Track Inspection Works Using Mathematics — track geometry and health mapping.
How MRT Wheel–Rail Contact Works Using Mathematics — force, friction and wear.
How MRT Noise and Vibration Work Using Mathematics — decibels, resonance and vibration transmission.

The next natural pillar is How MRT Signalling and Train Regulation Work Using Mathematics: position uncertainty, movement authority, safe separation, CBTC, automatic train operation, headway regulation, control loops and why a railway needs both a timetable and a real-time referee.

How MRT Power Supply Works Using Mathematics: Why a Railway Is Also a Moving Electrical Grid