Railroad Switch Point Machine Diagnosis Using Component-Mapped Event Signals
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing methods for diagnosing railroad switch point machines are often time-consuming, non-specific, or require large amounts of training data, making them inefficient for precise and timely detection of malfunctions.
Innovation Solution
A method that compares first and second time series of sensor signals from a point machine, allocates event points to specific components using simulation modeling, and quantifies mismatches or similarities to identify component-specific faults without requiring pre-trained data-driven models.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If data driven machine learning methods are used for diagnosis, then prediction reliability can be improved, but huge amounts of training data are required
Solution Approach 1:
The patent applies preliminary action by pre-defining event points in the time series data that correspond to specific component operations or states. Instead of requiring large amounts of training data to learn these patterns, the method proactively identifies and marks critical events based on domain knowledge and simulation modeling, enabling reliable diagnosis with minimal data preparation
Solution Approach 2:
The patent introduces simulation models as an intermediary between raw sensor data and diagnosis. These simulation models generate expected time series data for different component states, serving as a bridge that reduces the need for extensive real-world training data while maintaining prediction reliability
2Measurement precision
If commercial diagnosis systems are used to evaluate sensor signals, then diagnostic support can be improved, but the process becomes time consuming
Solution Approach 1:
The patent segments the diagnostic process by dividing the time series data into distinct phases based on event points. Each phase corresponds to a specific component operation, allowing targeted analysis of relevant data segments rather than processing entire data sets, thus improving diagnostic precision while reducing time consumption
Solution Approach 2:
The patent applies partial action by focusing analysis only on time periods surrounding identified event points rather than continuously monitoring all data. This selective approach maintains high diagnostic precision for critical events while significantly reducing the overall time required for evaluation
3Measurement precision
If human experts perform inspections to infer health conditions, then diagnostic accuracy can be improved, but the process is time consuming and not automated
Solution Approach 1:
The patent enables self-service diagnosis by automating the expert knowledge embedded in event point identification and allocation algorithms. The system automatically processes sensor data, identifies critical events, and diagnoses component issues without requiring human expert intervention, thereby maintaining diagnostic accuracy while dramatically improving efficiency and productivity
4Reliability
If general sensor signal processing is used to detect deviations, then abnormal behavior can be detected, but the diagnosis is not specific to individual components
Solution Approach 1:
The patent applies local quality by allocating different event points to different components based on their specific operational characteristics. Each component has its own set of relevant event points and analysis parameters, enabling the system to detect malfunctions reliably while maintaining component-specific diagnostic information rather than providing generic alerts
Data Source
Figure 1~2
Figure 3~4
Figure 5
AI summary
For diagnosing a railroad switch with a point machine, a first and a second time series (TS1, TS2) of a sensor signal (SS) of the point machine are received. Moreover, changes (CH) in the first and the second time series (TS1, TS2) are detected indicating changes of operational conditions of the point machine (PM). Furthermore, an event point (E1-E4) of a respective change (CH) in the first and in the second time series (TS1, TS2) is allocated to a respective component of the railroad switch (SW) or of the point machine based on a simulation modelling the respective component. Then for a respective component: - event points (E1-E4) allocated to that respective component (C1) are identified, - the sensor signal (SS) at a first identified event point (E1-E4) in the first time series (TS1) is compared with the sensor signal (SS) at a second identified event point (E1-E4) in the second time series (TS2), and - depending on the comparison a component-specific fault information and an identification of the respective component are output.