Rail Diagnostic System Merging On-Board and Off-Board Sensor Data
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current diagnostic systems for rail systems fail to fully integrate data from both rail infrastructure and vehicles, leading to inefficient monitoring and inability to identify previously unknown failure signatures, particularly in infrastructure components.
Innovation Solution
A diagnostic system that combines on-board and off-board data acquisition from rail vehicles and infrastructure, merging sensor data to generate categorized event data for comparative analysis, allowing for the identification of unusual patterns and failure predictions without relying on accurate prediction tools.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If data from only on-board sensors is used for monitoring, then the monitoring system is simpler to implement, but the ability to identify infrastructure faults and unknown failure signatures is limited
Solution Approach 1:
The patent merges on-board data acquisition systems with off-board infrastructure monitoring systems into a unified diagnostic platform. This integration combines sensor data from both mobile rail vehicles and stationary infrastructure components, enabling comprehensive fault identification while distributing system complexity across multiple coordinated units rather than concentrating all functionality in a single complex system
2Reliability
If separate feature detectors are used for monitoring specific aspects, then the monitoring coverage is comprehensive, but the system cannot effectively integrate mobile and stationary data sources
Solution Approach 1:
The diagnostic system implements a universal data processing platform that handles multiple data types from both mobile and stationary sources through a common architecture. The system uses standardized data fusion techniques that can process vibration data, acoustic data, operational parameters, and environmental data from various sources uniformly, enabling the system to adapt to different sensor types and data formats without requiring separate processing paths
3Productivity
If maintenance is performed based on mileage or timescale, then the maintenance schedule is simple to manage, but trains are taken out of operation for unnecessary servicing or unforeseen repairs
Solution Approach 1:
The system implements continuous feedback loops where sensor data from both on-board and off-board sources are constantly monitored and analyzed to assess the actual condition of rail vehicles and infrastructure. This condition-based feedback replaces fixed mileage or time-based maintenance schedules, allowing maintenance to be performed only when actual degradation thresholds are reached, thereby optimizing train availability while preventing unforeseen failures
4Measurement precision
If historical data comparison is used for fault identification, then the system can predict failure times, but the system cannot identify previously unknown failure signatures
Solution Approach 1:
The system performs preliminary analysis of data patterns from multiple sources before formal fault diagnosis is required. By continuously pre-processing and correlating data from on-board sensors, off-board infrastructure sensors, and historical records, the system builds a baseline understanding of normal and abnormal conditions. This preliminary action enables the system to detect previously unknown failure signatures by comparing real-time multi-source data against established patterns, allowing adaptation to new failure modes without requiring complete retraining of prediction models
Data Source
Figure 1
Figure 2
AI summary
The vehicles (12) of at least one fleet of rail vehicles are provided with on-board sensors (22) and rail vehicle positioning means (23). The rail infrastructure on which the rail vehicles circulate is provided with fixed rail infrastructure sensors (18). The rail infrastructure-related sensor data is merged with the rail vehicle-related sensor data, with location data representative of the location of the rail infrastructure-related sensors and with the rail vehicle position data for generating series of categorized event data representative of the occurrence of categorized events at a given location on the rail infrastructure over time and/or on a given rail vehicle of the fleet over time. The series of categorized events data representative of at least one category of events can be compared over any predetermined period of time to identify any location of the rail infrastructure and/or any rail vehicle which exhibits a series of events data that is significantly different from the other locations of the rail infrastructure and/or rail vehicles of the fleet over said predetermined period of time.