{"id":8419,"date":"2026-09-11T05:32:12","date_gmt":"2026-09-11T03:32:12","guid":{"rendered":"https:\/\/www.scs.ch\/uncategorized-en-us\/detection-of-wheel-defects-with-smart-algorithms\/"},"modified":"2026-09-20T07:38:46","modified_gmt":"2026-09-20T05:38:46","slug":"detection-of-wheel-defects-with-smart-algorithms","status":"publish","type":"post","link":"https:\/\/www.scs.ch\/en\/case-studies\/detection-of-wheel-defects-with-smart-algorithms\/","title":{"rendered":"Detection of Wheel Defects with Smart Algorithms"},"content":{"rendered":"

Initial Situation<\/h2><\/div>

Wheel defects, such as flat spots or spalling, strain railway infrastructure and cause noise. The wheel load checkpoints of the train inspection system primarily measure the weight of passing vehicles. However, wheel defects alter the resulting signal curves. Thus, these same measurements also contain information about the condition of the wheels, which can be evaluated with suitable algorithms.<\/p>\n<\/div>

\"\"<\/span><\/div><\/div><\/div>

Project Overview<\/h3><\/div>

Client<\/strong>
SBB<\/p>\n<\/div>

<\/div><\/div>

Area<\/strong>
Public Transport<\/p>\n<\/div>

<\/div><\/div>

Solution<\/strong>
Detecting wheel damage from checkpoint measurement data.<\/p>\n<\/div>

<\/div><\/div>

System Environment <\/strong>
\nWheel Load Checkpoints (RLC) as part of the Train Inspection System (ZKE), additional sensors for wheel profiling<\/p>\n<\/div>

<\/div><\/div>

Data Basis <\/strong>
\nSignal curves of passing vehicles<\/p>\n<\/div><\/div><\/div><\/div><\/div>

SCS’s Contribution<\/h2><\/div>

SCS developed and tested three algorithmic approaches that detect and more precisely analyze various wheel defects from the measurement signals of the wheel load checkpoints.<\/p>\n<\/div><\/div><\/div>

<\/i><\/div>

Wheel Defects Detected via Characteristic Value<\/h4><\/div>
<\/div>
\n

A dynamic coefficient is calculated from the signal profile of each wheel. A low value indicates no wheel defect, while a high value points to a pronounced wheel defect.<\/p>\n<\/div><\/div><\/div>

<\/i><\/div>

Measuring Flat Spots<\/h4><\/div>
<\/div>
\n

An algorithm developed by SCS recognizes the typical signal profile of a flat spot. Characteristic points in the measurement curve then allow for an estimation of its length.<\/p>\n<\/div><\/div><\/div>

<\/i><\/div>

Wheel Surface Profiled<\/h4><\/div>
<\/div>
\n

Another algorithm maps the condition of the tread across the entire wheel circumference, making individual damaged areas visible. Since existing RLC sensors do not capture the full circumference, additional sensors are used for this purpose.<\/p>\n<\/div><\/div><\/div>

<\/i><\/div>

Measurement Data Becomes Actionable Recommendation<\/h4><\/div>
<\/div>
\n

SCS processes the signals from individual sensors and combines them for an overall evaluation. This allows for the derivation of recommendations on which vehicles should be reported for maintenance.<\/p>\n<\/div><\/div><\/div>

<\/div><\/div><\/div><\/div>
\"\"<\/span><\/div><\/div><\/div><\/div><\/div>
<\/div>

Result<\/h2><\/div>

The three approaches allow for the detection and more precise analysis of existing wheel defects. For certain types of defects, their severity can also be estimated. Affected vehicles can be reported to the railway company and specifically repaired. This evaluation supports condition-based maintenance and helps reduce strain on infrastructure and the environment.<\/p>\n<\/div><\/div><\/div>

3 Approaches<\/h3><\/div>

for detection, flat spot analysis, and wheel profiling.<\/p>\n<\/div><\/div><\/div>

Length Estimated<\/h3><\/div>

Flat spots measured based on characteristic points in the measurement signal.<\/p>\n<\/div><\/div><\/div>

Entire Wheel Circumference<\/h3><\/div>

Condition of the tread displayed as a profile.<\/p>\n<\/div><\/div><\/div><\/div><\/div>

Services Provided<\/h2><\/div><\/div><\/div>

Algorithms, ML & AI<\/h3><\/div><\/div><\/div><\/div><\/div><\/div>

Public Transport<\/h3><\/div><\/div><\/div><\/div><\/div><\/div><\/div><\/div>
\"The<\/div><\/div>

Your Contact Person<\/h2><\/div>
Pascal Iselin<\/div><\/div>

Algorithms, Analytics & AI
\n+41 43 456 16 00<\/a><\/p>\n<\/div>

Speak with the responsible specialist<\/span>Speak with the responsible specialist<\/span><\/div><\/a><\/div>

For questions regarding algorithms, signal evaluation, and the analysis of complex measurement data.<\/p>\n<\/div><\/div><\/div><\/div><\/div><\/div><\/div><\/div><\/div>

Further Case Studies<\/h2><\/div>