Introduction
Train wheels are one of the most critical components of rail systems, and even minor defects can lead to safety risks, increased maintenance costs, and service delays. Modern Train Wheel Defect Detection (TWDD) systems use high-speed cameras and sensors mounted along the tracks to capture wheel images and vibration data as trains pass by. These systems require real-time processing at the edge to detect cracks, flat spots, and wear before they become operational hazards.
To support these demanding workloads, railway operators need an industrial computer that can process large image datasets and sensor inputs directly at trackside—often in outdoor or vibration-prone environments. This case study shows how Premio partnered with a leading railway technology company to power its Train Wheel Defect Detection solution with a rugged, PoE-enabled edge computer.
The Company
The customer is a trusted technology provider for India’s railway sector, specializing in Signaling & Telecom, Rolling Stock, and Trackside Equipment. They have a long history of delivering innovative solutions to Indian Railways, Metro Rail, and Rapid Transit Systems. With a focus on safety and operational efficiency, the company integrates AI-driven analytics and modern sensing technologies to advance rail infrastructure performance.
The Challenges
In earlier deployments, the customer’s Train Wheel Defect Detection system faced several limitations that impacted performance and maintainability:
Extra Hardware Complexity
External PoE switches increased installation time, consumed more space inside trackside cabinets, and required additional power and cooling considerations.
Bandwidth Limitations
The external PoE solution struggled with high-speed camera data transmission, occasionally resulting in dropped frames and incomplete defect analysis.
Higher Maintenance & Downtime
More components meant more potential points of failure. Maintaining and troubleshooting external PoE devices increased operational costs and risk of unplanned downtime.
Demanding AI Processing Capability
As train wheel inspection technology evolved to include AI-driven analytics, the previous hardware struggled to handle advanced image recognition and predictive maintenance tasks. The customer needed an IPC capable of supporting higher AI workloads for future algorithm enhancements.
Scalability Concerns
Adding more cameras or sensors required redesigning the I/O layout, limiting the system’s flexibility to scale with future inspection requirements.
The Solution
The customer needed a computing platform capable of handling high-speed image processing from multiple trackside cameras while being simple to install and maintain in harsh rail environments. After evaluation, they selected Premio’s ACO-6000-CML fanless, railway-certified computer, which offers native support for eight PoE LAN ports via two EDGEBoost I/O modules.
Why Direct PoE Matters
Previous solution, camera-based inspection systems require a separate PoE switch to power cameras and transmit data to the computer. This adds wiring complexity, requires additional power distribution, and creates another potential failure point.
With the ACO-6000-CML:
- Each camera connects directly to the computer using a single Ethernet cable.
- Power and data are delivered together, eliminating the need for separate power adapters or an external PoE switch.
- Simplified cabinet design reduces installation time, frees up space, and improves overall system reliability.
AI Processing Capability for Advanced Wheel Analysis
Train Wheel Defect Detection systems depend on AI-driven image and data analysis to ensure railway safety and operational efficiency. With this vision AI workload remaining quite minimal in processing requirements, the 10th Gen Intel Core TE processor can streamline this workload of intaking multiple video data streams and utilizing AI inferencing to detect abnormalities in the pretrained model.
With this processor, the system enables:
- Detection of Wheel Flaws for Safety Enhancement: Identifies cracks, flat spots, and other defects early to prevent failures.
- Monitoring of Wheel Wear and Tear: Tracks gradual changes in wheel profiles to schedule timely maintenance.
- Alerts on Abnormal Wheel Behavior: Detects irregular vibration or noise patterns for immediate operator attention.
- Predictive Maintenance Assistance: Uses data-driven models to forecast potential issues, reducing downtime and repair costs.
- Ensuring Optimal Wheel-Rail Interaction: Analyzes wheel geometry and alignment to improve ride quality and reduce track wear.
Railway-Ready Performance & Certifications
The ACO-6000-CML is designed for harsh rail environments with:
- EN50155 & EN50121-3-2 compliance, ensuring safe and reliable operation near railway tracks.
- 10th Gen Intel® Comet Lake-S processors (up to 65W/35W TDP) and up to 64GB DDR4 RAM, delivering the processing power needed for real-time defect detection algorithms.
- Flexible storage with one internal and two hot-swappable 2.5” SATA drives for fast data handling and easy field maintenance.
- Rugged design supporting -25°C to 70°C operation, 50G shock, and 5Grms vibration.
By integrating high-speed PoE cameras directly into the ACO-6000-CML, the customer achieved a single-box solution for both data capture and edge analytics, eliminating extra hardware while improving inspection accuracy and reducing downtime.
The Benefit
- Simplified Architecture: Removed external PoE switches, reducing wiring, space requirements, and potential points of failure.
- Improved Reliability: Fewer components translated to fewer maintenance calls and higher uptime of the defect detection system.
- Enhanced Real-Time Processing: High-performance CPU and high-speed interfaces supported multiple camera streams for fast, accurate wheel analysis.
- Scalable & Future-Ready: Built-in PoE and flexible I/O allow easy addition of sensors or advanced analytics software without redesign.
- Compliance & Safety: Certified for railway environments, ensuring safe, long-term deployment.
Conclusion
By integrating Premio’s ACO-6000-CML computer with native 8x PoE support, the customer streamlined their Train Wheel Defect Detection system and improved operational reliability. This solution not only simplified installation and maintenance but also provided a powerful, rugged, and scalable platform for future rail safety innovations.