Summary

In a collaborative effort with a major Class 1 Railroad, ITG successfully completed Phase 1 of the Predictive Analytics project with the SORBA Advanced Analytics Software. The initial phase, focused on analyzing historical data related to the electrical currents used in switch motors over several months of operations. Utilizing SORBA's cutting-edge data analytics and machine learning capabilities, the objective was to identify and classify past failures.

The completion of Phase 1 laid the groundwork for Phase 2, which aimed to further expand the analytics scope and deploy machine learning agents for real-time failure prediction. The ultimate goal was to offer actionable insights for preventing incipient issues like wear, rust, debris, or vandalism. The pilot justified its existence by estimating a 30% to 40% reduction in maintenance costs, along with other operational benefits.

Benefits Achieved

The project provided the customer with several advantages, including:

• Early warning signs for machine failure.

• Improved efficiency for railway maintenance staff.
• Targeted replacement of faulty components.
• Minimized operational disruption.
• Controlled maintenance costs.

Solution and Deliverables

For the second phase, the customer identified specific rail switch locations for analysis. Once SORBA's machine learning agents, also known as SORBOTS, were developed, they were integrated into the real-time prediction engine. These SORBOTS were designed for both anomaly detection and specific failure mode predictions and could easily be replicated across other similar switch types, subject to additional licensing.

Scope:

1. Reviewed protocol and repair documentation.

2. Imported switch and failure data for the specific rail switches as identified by customer.

3. Performed advanced machine learning on this data.

4. Compiled and reported findings.

5. Developed a data import integrator for the SORBA-SDC platform.

6. Created and deployed SORBOTS to the SORBA-SDC runtime engine.

7. Configured SORBOTS to send email alerts for specific anomalies.

Additional Services

The next steps were intended to align both customer and ITG towards broader production adoption of the SORBA solutions. Additional activities included:

1. Advanced machine learning analyses on rail switch data.

2. On-site collaboration sessions to establish expectations, roles, and future roadmaps.

3. Technical deep-dives within Intellectual Property constraints.

Services Provided:

1. Discovery Services: Comprehensive review of documentation and system integration.

2. Engineering & Data Science: Detailed data analysis, including data cleansing and model analysis.

3. Development & Data Science: Building predictive models and integrating them into the rail switch data systems.

4. Meeting and Collaboration: Iterative review of results and necessary adaptations.

5. On-site Commissioning: Implementation and testing of the fully integrated system.

For more information about this completed project or to discuss future projects, please

ITG is committed to innovation in advanced data analytics and machine learning solutions. Your railway’s efficiency is our priority.

‍