Rail Fault Report Analysis for Failure-Based Maintenance Planning
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing rail system maintenance planning methods struggle to accurately identify the cause of malfunctions and prioritize maintenance effectively due to the complexity of rail systems and the lack of clear operational consequences of device failures.
Innovation Solution
A method and device for maintenance planning that utilize natural language processing to analyze fault reports, map them to specific failure mode types, determine associated failure effects, and calculate the time or distance until maintenance is required, using a failure-mode-and-effects model and statistical topic models.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If regular inspections and maintenance are performed based on fault reports alone, then maintenance activities are conducted, but it is difficult to ascertain the correct manner of proceeding and prioritize maintenance effectively due to the complexity of rail systems
Solution Approach 1:
The patent segments the complex rail system into distinct failure mode types (e.g., component failure, system failure, operational failure) and associates each with specific failure effects. This segmentation allows maintenance personnel to categorize fault reports into manageable failure mode categories, making it easier to determine appropriate maintenance actions despite the overall system complexity.
Solution Approach 2:
The patent introduces a failure-mode-and-effects model as an intermediary between fault reports and maintenance decisions. This model acts as a mediator that translates raw fault report data into structured failure mode classifications and associated operational consequences, enabling systematic maintenance prioritization without requiring direct expert analysis of every complex system interaction.
2Reliability
If maintenance is planned based on fault reports without analyzing operational consequences, then maintenance scheduling is simplified, but it is unclear what effect the malfunctioning part has on the safe and reliable execution of rail services
Solution Approach 1:
The patent performs preliminary analysis by pre-establishing the failure-mode-and-effects model that links failure modes to their operational consequences before actual maintenance planning occurs. This preliminary action includes identifying which failure modes have safety-critical effects and which have minor operational impacts, so that when fault reports are received, the system can quickly retrieve pre-analyzed consequences rather than performing time-consuming analysis during maintenance planning.
Solution Approach 2:
The patent implements a feedback mechanism where fault reports are analyzed against the failure-mode-and-effects model to generate maintenance recommendations that incorporate operational consequence information. This feedback loop provides maintenance personnel with actionable insights about the safety and operational implications of detected faults, enabling informed prioritization decisions without requiring extensive manual analysis time.
3Measurement precision
If detailed analysis of fault reports is performed to identify underlying causes, then maintenance accuracy is improved, but the complexity of the analysis process increases
Solution Approach 1:
The patent changes the parameter of analysis from examining raw fault report text directly to analyzing structured failure mode classifications derived from the failure-mode-and-effects model. By transforming the analysis parameter from unstructured text to categorized failure modes with associated effects, the system achieves precise failure cause identification while keeping the analysis process manageable through standardized categorization schemes.
Data Source
Figure 1~2
Figure 3
Figure 4~5
AI summary
A method and device for maintenance planning of a rail system device, in particular a rolling stock device or a railway infrastructure device, are disclosed, relating to the steps of receiving (S1 ) a fault report message relating to a malfunctioning rail system device; analyzing (S2) the fault report message using natural language processing; mapping (S3) the analyzed fault report message to a particular failure mode type, the failure mode type identifying at least one probable cause of malfunction of the rail system device related to the fault report message; determining (S4), using the failure mode type, a failure effect associated with the failure mode type, wherein the failure effect identifies one or more operational consequences of the particular failure mode type; and determining (S5), one or more of: a time or a distance until maintenance of the rail system device, using the failure effect.