Reimagining Crossing Maintenance

Switches and Crossings (S&C) are essential to any rail network but are also among the fastest degrading assets. Due to their complex geometry and exposure to dynamic loads, they account for around 20% of infrastructure-related delays and contribute significantly to maintenance spending. Traditional approaches, such as visual inspections and geometry measurements, provide valuable snapshots of surface condition but they often miss the dynamic behaviour of a crossing (aka frog) when it’s under load. This is where the earliest signs of degradation typically appear.

KONUX’s Wheel Trajectory Reconstruction (WTR) provides a new layer of insight. WTR is a fully remote, physics-based method for understanding how wheels interact with crossings in near real-time. Rather than relying on visible defects or static indicators, it reconstructs the actual movement of wheels across the crossing uncovering the forces and motion that drive degradation and failure over time.

Why Wheel Trajectory Reconstruction Matters

Most failures in S&C are caused not only by wear, but also by how the asset performs under loads. A crossing may appear visually sound, yet still be at risk of rapid deterioration if it experiences excessive impact forces, loss of contact, or structural instability during wheel passage. These failure modes are difficult (if not impossible) to detect without tracking the dynamic response of the crossing.

WTR fills this gap by offering insight into what’s happening at the wheel-rail interface:

• Detect early signs of wear before visible defects appear

• Distinguish between different failure modes, guiding the right maintenance action

• Maintain proactively by replacing reactive interventions with condition-based strategies

By continuously monitoring wheel-rail interaction, WTR helps operators focus efforts where they’re needed most avoiding unnecessary work while preventing costly failures.

How Wheel Trajectory Reconstruction Works

KONUX WTR operates in three key steps:

1. Calibration

KONUX IoT devices, installed on the sleepers adjacent to the crossing, capture vertical acceleration data from every passing train. These readings are used to calibrate the trackbed conditions on a daily basis, reflecting changes in ballast stiffness, rail support, or traffic mix.

2. Digital Twin

A physics-based simulation, built using real field measurements and rail geometry, models the structural response of the crossing under train loads. The Digital Twin captures how the wheel transitions through the crossing, and how the rail reacts including displacement, lift-off, or deformation.

3. Outputs

From these simulations, KONUX reconstructs:

• The wheel trajectory, showing the vertical path of the wheel through the crossing
• The impact force, indicating stress concentrations and areas of concern quantifying the forces at the wheel-rail interface
• The dip angle, a well-recognised metric for monitoring defects at rail discontinuities

This output is automatically processed and displayed in the KONUX UI, with clear indicators of change over time and alerts when intervention thresholds are reached.

From Days to Seconds: WTR at Scale

Until now, this level of physical modelling was limited to research environments, due to its high computational cost. Reconstructing wheel-rail trajectories using multi-body simulations could take days for a single passage.

KONUX has changed that. By using an inverse problem-solving approach known as the Green’s Kernel Function Method (GKFM), the system pre-computes the structural response and solves for unknowns in real time. This reduces processing time to under 0.3 seconds per passage, making it feasible to run WTR across thousands of assets every day, with no manual input required.

This breakthrough unlocks the potential of physics-based monitoring at operational scale not only for detailed case studies, but also for daily decision-making.

Proven in the Field

WTR is now active on over 1,500 monitored crossings in Germany and the UK, running autonomously and delivering insights from over 100 train passages per asset per day. The system has already identified multiple cases of poor contact, unexpected impact forces, and progressive degradation trends many of which were not visible through standard inspection techniques.

Validation has been carried out by comparing reconstructed trajectories with 3D rail geometry scans and aligning WTR outputs with known defect cases. In several instances, WTR has revealed patterns of force unloading, asymmetric impacts, or premature contact loss that were not visible using conventional inspection methods.

Combined with visual tools such as One Big Circle’s AIVR imagery, WTR offers a more complete understanding of both the condition and the behaviour of each asset. Essentially providing a fuller picture of crossing condition helping engineers connect the dots between structural behaviour and physical appearance.

A Step Toward Condition-Based Maintenance

WTR provides a foundation for moving from interval-based inspections to evidence-based, condition based maintenance. It supports decisions not only around when to intervene, but also how, enabling targeted actions that are aligned with the root cause of degradation.

• By revealing the dynamic behaviour of crossings under load, WTR enables infrastructure managers to:
• Prioritise earlier interventions with greater accuracy
• Understand how maintenance actions influence performance over time
• Detect degradation trends before failures occur
• Align capital planning with asset performance data

As the demands on rail infrastructure continue to grow, budgets continue to shrink and performance expectations are at an all time high, features like WTR will be critical in improving performance, reducing risk, and managing resources more effectively.