How adopting data-driven decisions on trackbed maintenance can save 66% of costs

Case Study, 15.06.2022

Infrabel, the Belgian public rail infrastructure service provider, embarked on a pilot with KONUX in 2019. The main goal of the pilot was to demonstrate the capability of KONUX Switch in trackbed health monitoring, trackbed health prediction, and maintenance validation.

The Challenge

Optimisation of trackbed maintenance has been an old challenge for Infrabel. While not having the transparency of the problem due to a lack of insights, Infrabel was aware they could greatly improve their tamping practices and suboptimal maintenance planning.

One of the main problems was repetitive tamping. As the quality of the tamping operations could not be timely and adequately evaluated, the switches were tamped repetitively without a clear picture of their actual conditions, the root causes of the degradation, and the right methodology.

Another major challenge was that, in some regions, tamping operation planning was largely inefficient as the decisions were not based on actual switch conditions. The problematic assets were tamped on a fixed time-based approach and partly based on the convenience of simply being in the area. Occasionally, these ‘just-in-case’ tamping practices heightened the vertical displacement and worsened the stability of the trackbed (as demonstrated in Example 1 below).

Acknowledging that the current approach is suboptimal, Infrabel aimed to understand their current practices better, gain insights into how the tamping activities affect trackbed health, and collect data to improve maintenance methodology. Specifically, the central switch department aimed to:

  • Monitor the number of tamping operations per switch and prevent actions that are not supported by clear reasons
  • Gather insights around the percentage of the inefficient or unnecessary operations
  • Gather the best practices on what works and what doesn’t

The Solution

The pilot began by equipping ten switches with 20 IIoT devices (two devices per switch). At the pilot’s later stage, two additional switches (one device per switch) were added. The devices were installed in the frog and point machine areas and both on wood and concrete sleepers. The installation was easy, safe, and speedy (performed within 10 minutes per device), and during the entire pilot, no device presented any potential safety hazard.


Image 1: Installation of the device on a point machine area

Throughout the pilot, KONUX Switch gathered and processed a large amount of train traces from each of the monitored switches. For each use case, KONUX Switch provided the following solutions:

Trackbed health monitoring:

  • Providing continuous monitoring of trackbed stability and capturing the accurate trend of the switch health, unlike manual measurements which give only a snapshot of the switch condition
  • Helping avoid in-person inspections and promoting efficiency and safety

Trackbed health prediction:

  • Helping determine if a tamping operation is needed on each monitored switch in the next three months
  • Helping prevent unnecessary tamping actions by providing future-forwarding insight
  • Helping improve maintenance planning to be more efficient

Maintenance validation:

  • Providing an overview on the maintenance effectiveness both at a switch and network level
  • Helping define the best practices for tamping (e.g. understand which machine works best for given conditions)
  • Helping validate the new maintenance methodology (e.g. new pad material)

The Results

The pilot successfully proved KONUX Switch to be an optimal tool for Infrabel to support its decision making process for scheduling tamping operations and evaluating the effectiveness of such interventions. It was concluded that the costs currently associated with such interventions would be substantially reduced with the help of KONUX Switch. The pilot revealed a potential 66% savings in tamping operations.

Example: Trackbed monitoring & prediction

The system identified an asset where an unnecessary tamping was done although its 3-month forecast of the vertical displacement was stable. KONUX Switch revealed that this tamping operation caused degradation of the trackbed, which otherwise would have been kept sound.


Example 1 – Before tamping: 3-month forecast at the beginning of July, showing a stable development under the predefined threshold (3 mm)


Example 1 – After tamping: the progression of the vertical displacement after the tamping done in July

In terms of the accuracy of the insights, a comparison test confirmed the accuracy and reliability of KONUX Switch, showing the deviation margin within +/- 0.25 mm in comparison to the standard measurements.

Example: Maintenance validation

For maintenance validation, 34 tamping actions were detected from 12 switches, of which 33 were evaluated. KONUX Switch revealed that 40% of the tamping was unnecessary or untimely (the vertical displacement was not above or immediately approaching the threshold) and only 15% of the actions were deemed effective (the vertical displacement after the tamping is kept at least 15% below the initial value for the whole duration of 60 days). Also, it was found that three switches were tamped twice or more in a year.


Example 2 – Effective tamping (before: 3.36 mm → after 60 days: 1.64 mm)


Example 3 – Ineffective tamping (before 3.31 mm → 17 days: 2.63 mm → 60 days: 2.90 mm)

Additionally, KONUX Switch allowed validation of the new tamping machine before further deployment by automatically detecting the operation done by the new machine and evaluating its effectiveness.


Example 4 – Validation of new tamping machine: 3-month prediction at the 60-day point after the tamping operation

Furthermore, in January 2022, Infrabel also tested the effectiveness of the new pad material as part of its initiative to build the best practice guidelines with the help of KONUX Switch. In this test, a significant improvement in the vertical displacement was observed after replacing the material, allowing Infrabel to validate the effectiveness of the new methodology.


Would you like to learn more about the case? Get in touch with our team here.