Autonomous Rail Defect Classification for High-Speed Inspection

Resolve Bottlenecks,
Find Innovative Solutions
Generate Solutions

Solution Overview

Problem

Current rail inspection methods rely on manned vehicles traveling at low speeds, which are inefficient and do not provide real-time defect classification and prioritization, potentially leading to safety risks and delays in maintenance.

Innovation Solution

An autonomous rail inspection system (ARIS) that uses non-destructive testing techniques and position determination to detect and classify rail defects, ranking them for maintenance teams in real-time, optimizing repair efficiency and safety.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Measurement precision

If manned guided vehicles are used for rail inspection, then an operator can detect and locate rail defects, but the inspection speed is limited to low speeds and real-time classification and prioritization are not provided

Engineering Contradiction:
Improvedefect detection and classification accuracyVSAvoidinspection speed
Core Design Contradiction:
Measurement precisionVSSpeed

Solution Approach 1:

The patent replaces the manual inspection process with an automated inspection system that uses sensors, processors, and communication devices to detect, classify, and prioritize rail defects automatically, eliminating the need for operators to manually inspect at low speeds while maintaining high detection accuracy

Inventive Principle:
Principle #28Mechanics substitution (Replace mechanical system)

Solution Approach 2:

The inspection system performs self-classification and self-prioritization of defects using onboard processing capabilities, automatically ranking defects by severity and generating maintenance priorities without requiring external human analysis, thereby enabling high-speed inspection with accurate defect characterization

Inventive Principle:
Principle #25Self-service

2Loss of information

If manual defect verification and classification is performed by maintenance teams, then defects can be categorized by type, but time is lost in hand verification and repair prioritization is not optimized

Engineering Contradiction:
Improvedefect classification informationVSAvoidtime for defect verification and prioritization
Core Design Contradiction:
Loss of informationVSLoss of time

Solution Approach 1:

The system performs preliminary classification and prioritization of defects during the inspection process itself, using onboard processing to categorize defects by type and severity before maintenance teams arrive, so that when defects are verified, their priority and classification are already established, eliminating redundant time-consuming steps

Inventive Principle:
Principle #10Preliminary action

Solution Approach 2:

The system provides real-time feedback to maintenance teams through communication devices, transmitting classified and prioritized defect information immediately after detection, enabling maintenance teams to prepare appropriate resources and respond efficiently without waiting for manual verification delays

Inventive Principle:
Principle #23Feedback

3Productivity

If defects are detected without automatic classification and ranking, then all defects are identified, but maintenance efficiency is reduced due to lack of repair prioritization

Engineering Contradiction:
Improvemaintenance efficiencyVSAvoidinspection system complexity
Core Design Contradiction:
ProductivityVSDevice complexity

Solution Approach 1:

The system segments defects into different priority categories and types using automated classification algorithms, dividing the maintenance task into prioritized groups that can be addressed systematically, thereby improving maintenance efficiency by focusing on critical defects first while managing system complexity through structured classification

Inventive Principle:
Principle #1Segmentation

Data Source

PatentEP3333043B1Rail inspection system and method
Publication Date: 2021.04.28 SIEMENS RAIL AUTOMATION
  • EP3333043B1 patent drawingFigure 1

AI summary

The present invention concerns an automatic rail inspection system and method for automatically inspecting a rail (21) of a railway system, detecting, locating and classifying rail defects, the automatic rail inspection system (1) comprising: - a rail defect detection system (10) for automatically detecting a rail defect, said rail defect detection system (10) outputting measurement data related to its inspection of the rail (21); - a processing unit (11) configured for analyzing the measurement data in order to automatically determine for each detected defect a defect characteristic; - a defect classification module (12) configured for classifying the detected defects in a list in function of the determined defect characteristics, wherein the classification of the detected defects automatically defines a repair order for the detected defects.