The ideal running rail is continuous. A special joint exists only when another physical requirement is strong enough to justify breaking that continuity.
The mathematical challenge is therefore not “how do we make a gap?” It is “how do we accommodate required movement or separation while making the wheel, track support and maintenance system experience as little discontinuity as possible?”
LTA’s December 2025 Civil Design Criteria state that plain-line running rail is continuously welded and that rail expansion joints on bridge or viaduct track should be avoided. That is an important ownership boundary.
This article does not argue that every MRT line should contain running-rail expansion joints.
It explains the engineering class of special rail discontinuities: places where structural movement, exceptional geometry, electrical separation or another legitimate interface requirement means continuity must be interrupted or transformed.
Ordinary rail continuity, thermal stress and neutral temperature remain with Rail Welding and Continuously Welded Rail. Bridge structural movement remains with Viaducts and Elevated Structures.
The RFE — Why Have a Special Joint at All?
allow a necessary local movement or separation while preserving wheel guidance, load transfer, acceptable dynamic smoothness, track geometry and maintainability across the discontinuity.
If no legitimate movement or separation requirement exists, continuity is usually preferable.
Prompt 1 — How Much Movement Is the Joint Trying to Absorb?
Free thermal movement of a member is:
ΔL = αLΔT
A bridge deck, rail and supporting structure can have different temperatures, coefficients, creep and shrinkage histories.
Relative movement demand can be written conceptually:
Δrel = Δstructure − Δrail + Δcreep + Δshrinkage + Δsettlement + Δconstruction
The signs depend on the chosen coordinate direction.
LTA’s current criteria explicitly require track–structure interaction on bridges and viaducts to consider structure temperature change, residual creep and shrinkage while generally avoiding rail expansion joints.
The reason is simple: it is usually better to distribute compatible force and movement through continuous track than create another wheel-impact location unless a joint is truly justified.
Prompt 2 — What Happens If Movement Is Fully Restrained?
If thermal strain is prevented:
εthermal=αΔT
and a simple fully restrained stress scale is:
σthermal≈EαΔT
For rail area A:
Nthermal≈EAαΔT
A special joint can reduce one local relative-movement problem but it also changes how longitudinal force flows through the system.
The correct question is therefore not “stress or movement?”
It is how much stress and movement each component is intended to carry.
Prompt 3 — Why Is a Gap a Dynamic Problem?
A wheel crossing a perfectly smooth continuous rail sees no geometric step at the weld location.
A special joint can introduce:
- vertical step;
- horizontal misalignment;
- gap;
- local stiffness change;
- different support geometry;
- surface wear.
Let local vertical irregularity be y(x).
At vehicle speed v:
dy/dt = v dy/dx
and vertical acceleration input scales with:
d²y/dt² = v² d²y/dx²
This is why the same joint geometry becomes dynamically harsher at higher speed.
A joint that looks small in millimetres can become large in acceleration because the train crosses it quickly.
Prompt 4 — Why Does Stiffness Transition Matter?
Suppose normal track support stiffness is k1 and special-joint region stiffness is k2.
stiffness ratio ρk = k2/k1
If ρk differs greatly from 1, the wheel load produces a different local deflection.
δ1=F/k1 δ2=F/k2
This creates an additional dynamic transition even when the rail-head geometry is perfectly level.
Joint design therefore coordinates geometry and support stiffness.
Prompt 5 — How Does a Joint Carry Wheel Load Across a Discontinuity?
The wheel load must remain supported as it moves from one rail segment or supporting region to the next.
For equilibrium:
ΣRi = Pwheel
But local distribution Ri depends on rail bending stiffness, joint components and support stiffness.
An excessively flexible local joint can increase rail-end movement.
An excessively stiff joint can create a sharp support transition.
Again, the optimum is continuity of response rather than maximum local stiffness.
Prompt 6 — What About Electrical Separation?
Some railway systems use specialised joints or interfaces when electrical separation is required for signalling, stray-current or other system reasons.
Mechanically, such a location still has to transmit wheel load and preserve rail geometry.
Electrically, the intended separation can be represented conceptually by high impedance between the two sides:
|Zjoint| ≫ |Zrunning path|
Detailed signalling circuits, insulated-joint locations, detection logic and electrical acceptance criteria are intentionally outside this public article.
Prompt 7 — Why Is the Conductor Rail Different?
LTA’s current criteria explicitly require expansion joints in the third-rail conductor to accommodate longitudinal expansion and contraction while maintaining electrical continuity across the joint.
This is not the running rail.
The contrast is useful:
running rail: continuity strongly preferred; expansion joints generally avoided conductor rail: controlled expansion joints are an explicit part of thermal movement management
The moving electrical-contact consequences remain owned by Current Collection.
Prompt 8 — How Does a Special Joint Age?
A discontinuity concentrates maintenance attention because every wheel repeatedly crosses the same location.
Useful state variables can include:
- gap or overlap geometry;
- vertical step;
- lateral alignment;
- joint/support stiffness;
- fastener condition;
- impact vibration;
- wear profile;
- temperature and movement history.
Let measured joint movement be Δm(T) and expected movement Δ̂(T).
eΔ = Δm − Δ̂
If the joint no longer follows temperature as expected, restraint, wear or structural movement may have changed.
A Fictional Special-Joint Example
Imagine a fictional 120 m structural segment with effective thermal coefficient 10×10⁻⁶/°C and temperature change 25°C.
ΔL=αLΔT =10×10⁻⁶×120×25 =0.030m =30mm
Suppose the rail/support interaction absorbs most of this compatibly, leaving a fictional local relative movement demand of 6 mm at a special interface.
Now imagine the joint develops a 0.5 mm vertical step spread over a 100 mm transition length.
A train at 20 m/s crosses 100 mm in:
t=0.1/20=0.005s
The structural movement happened over hours; the wheel experiences its geometric consequence in milliseconds.
That difference in time scale is the heart of special-joint engineering.
Deletion Tests
- Remove relative movement: structures, rails and supports expand identically forever.
- Remove thermal stress: restrained movement creates no force.
- Remove geometric discontinuity: the wheel cannot detect joint shape.
- Remove speed: joint dynamic severity is unchanged from walking speed to line speed.
- Remove stiffness transition: all supports deflect equally.
- Remove electrical purpose: specialised separation has no systems reason.
- Remove wear: repeated wheel crossings do not change joint geometry.
- Remove World Return: actual movement versus temperature is never compared with design.
Special-Joint Paradoxes
- The best expansion joint in running rail is often the one the design successfully avoids.
- A joint exists to permit movement but must feel almost motionless to the wheel.
- Millimetres of slow structural movement can produce millisecond dynamic events.
- A perfectly level joint can still be dynamically harsh if its support stiffness changes abruptly.
- Electrical discontinuity can require mechanical continuity.
The Special-Joint Audit
- Why is continuity insufficient at this location?
- What relative movement or separation is genuinely required?
- How much comes from temperature, creep, shrinkage or settlement?
- What rail stress would occur without accommodation?
- What local geometry does the wheel cross?
- What support stiffness exists before, through and after the joint?
- How is wheel load transferred across it?
- Does the joint have an electrical function?
- What wear modes are concentrated there?
- How does movement vary with temperature?
- What measurements prove the special interface remains smoother than the failure it was designed to prevent?
World Return — The Joint Must Behave Like an Exception, Not a New Problem
predict structural movement → design special interface → install and survey → trains cross → measure movement, geometry and impact → compare with temperature and baseline → maintain / adjust / replace
A railway special joint works when it solves the exceptional movement or separation problem without making the wheel repeatedly pay a larger price for the discontinuity.
Reader-safety note: This article does not publish the locations, dimensions, acceptance limits or maintenance procedures of Singapore MRT special rail joints, electrical separation joints or bridge–track transition devices. Numerical examples are fictional.