Using railway vibrations to support better maintenance decisions
Passing trains provide a source of information about the infrastructure beneath them. Distributed fiber-optic measurements can capture how a track and its surroundings respond to repeated loading. The engineering opportunity is to move beyond observing that a train has passed and investigate whether the measured response contains useful evidence about infrastructure condition. However, that transition requires a suitable installation and a clear understanding of what the sensor is physically measuring.
A communications fiber and an instrumented track are not the same measurement system.
A 2025 field study on an operational railway investigated a mechanically coupled sensing fiber installed along 50 metres of rail, including a 7.5-metre section in the trackbed. The researchers demonstrated quantitative strain measurements and compared them with established measurement methods and simulations. This matters because it shows the role of deliberate sensor coupling: a fiber bonded to an asset can provide a different type of evidence from a telecommunications cable located somewhere beside the railway.
For an operator, the implication is practical. Before selecting an analytics package, establish whether the available installation supports corridor-level event awareness, asset-specific condition measurement or both. These objectives should not be combined into a single promise of “complete track monitoring.”
Repeated observations become valuable when they are comparable.
A useful maintenance approach would compare measurements under sufficiently similar conditions and investigate persistent changes. Train type, speed, loading and the relationship between the fiber and the structure should be considered when designing those comparisons. Otherwise, a change in operating conditions may be mistaken for a change in infrastructure condition.
The recommended initial role is advisory: directing inspections, supporting engineering review and building an evidence history. Treating a new sensing installation as a replacement for signalling or another safety-related function would require a separate safety justification and validation programme; train-detection capability alone does not establish that suitability.
Research is expanding from event detection towards trackbed characterisation.
An EGU 2026 conference contribution by researchers working with Swiss Federal Railways described measurements on two approximately 40-kilometre railway fiber routes over three months. One objective was to assess whether existing telecommunications fibers could support characterisation of the shallow subsurface beneath railway corridors. This is a promising research direction, but it was presented as a feasibility investigation not as a universally deployable, certified track-diagnostic product.
A practical pilot should therefore pair sensing observations with independent inspections and maintenance records. The strongest outcome is not an impressive vibration plot; it is a repeatable relationship between a measured change, an engineering finding and a better maintenance decision.
