Fleet Maintenance Prioritization via Burst Detection

Resolve Bottlenecks,
Find Innovative SolutionsGenerate Solutions

Solution Overview

Problem

Current maintenance systems for railway vehicles face challenges in efficiently analyzing the large volume of generated events to prioritize maintenance needs across a fleet, requiring significant operator time and cost, and often rely on empirical decisions due to the complexity of establishing correlations between events.

Innovation Solution

A method utilizing burst detection algorithms, specifically logarithmic maximum likelihood, to analyze time series data from monitoring systems, automatically identifying trains with urgent maintenance needs by detecting abnormal states and scheduling them based on detected bursts, allowing for efficient prioritization.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Measurement precision

If manual analysis of monitoring events is performed by maintenance operators, then detailed diagnosis and maintenance planning can be conducted, but the work time and costs increase significantly due to the large volume of events generated

Engineering Contradiction:

Improvemaintenance diagnosis accuracyVSAvoidoperator time for event analysis

Core Design Contradiction:

Measurement precisionVSLoss of time

Solution Approach 1:

The system enables automatic self-diagnosis by using burst detection algorithms to automatically identify abnormal patterns in monitoring events, eliminating the need for manual operator analysis while maintaining diagnostic accuracy

Inventive Principle:

Principle #25Self-service

Solution Approach 2:

The patent replaces manual mechanical analysis with automated computational algorithms that process monitoring events through burst detection to identify maintenance needs, substituting human cognitive processing with automated systems

Inventive Principle:

Principle #28Mechanics substitution (Replace mechanical system)

2Reliability

If manual correlation analysis of events is performed to establish meaningful patterns, then accurate maintenance decisions can be made, but the complexity and time required for analysis increase significantly

Engineering Contradiction:

Improvemaintenance decision accuracyVSAvoidanalysis process complexity

Core Design Contradiction:

ReliabilityVSDevice complexity

Solution Approach 1:

The system transforms complex event data into simplified temporal patterns by detecting bursts - concentrated sequences of events within specific time windows. This parameter transformation reduces complex multi-dimensional event correlations to manageable temporal patterns that automatically indicate maintenance needs

Inventive Principle:

Principle #35Parameter changes

3Ease of operation

If empirical decisions are made by operators to prioritize vehicle maintenance, then operational flexibility is maintained, but the ability to objectively compare vehicles and prioritize urgent maintenance needs is compromised

Engineering Contradiction:

Improvemaintenance scheduling flexibilityVSAvoidmaintenance priority assessment accuracy

Core Design Contradiction:

Ease of operationVSMeasurement precision

Solution Approach 1:

The system provides objective feedback by automatically calculating and comparing burst metrics across all vehicles in the fleet, enabling data-driven priority assessment that feedbacks maintenance decisions with quantitative measures of actual vehicle condition and urgency

Inventive Principle:

Principle #23Feedback

Data Source

PatentEP4361920B1Method for ordering vehicles of a fleet of vehicles according to a maintenance requirement; associated computer program and computer system

Publication Date: 2025.01.15 ALSTOM HOLDINGS SA

Figure 1
Figure 2

AI summary

This process (100) includes the steps of: for each vehicle in the fleet, determining (110) a time series having, for each time step, an instantaneous value of at least one quantity of interest obtained from monitoring events acquired by means of a fleet vehicle monitoring system; analyzing (120), over a predetermined time interval, the time series, taking into account the time series determined for all vehicles in the fleet, considering that at each time step, the state of a vehicle is either a "normal" state or an "abnormal" state, to obtain an optimal sequence of states over the predetermined time interval; and, detecting (130) the possible presence of one or more bursts in the optimal sequence of states, then ordering (140) a list of the vehicles in the fleet according to the properties of the bursts detected for each of the vehicles.