Big data for predictive maintenance
Lead Research Organisation:
Cranfield University
Department Name: Sch of Aerospace, Transport & Manufact
Abstract
Provide the state of art regarding tools and techniques used for big data and predictive maintenance.
Implement and validate models for feature extraction and fault detection and isolation or diagnostics.
Implement and validate models for fault prognostics.
Design, implement and validate the framework
Stage 1: Establish base lines and literature review. Examine common practices, and discoveries in the field related to techniques and process, and common tools and infrastructures used. Risk assessment.
Stage 2: Identify research gap, and review aims. Data collection, establish requirements for the framework requirement. Data cleaning and preparation, data management plan.
Stage 3: Using a condition based maintenance process, begin data analysis. Feature extraction; fault detection and isolation; diagnosis and prognostics using novel "big data" tools and knowledge discovery techniques.
Stage 4: Results visualisation; framework implemented and validated.
The work aims at a novel approach to time series analysis for system and machine faults, integrating physics, engineering know-how and discovered data patterns. The work has been trialled on train engine performance behaviour, door response, and wheel slip, seeking early and robust detection, diagnosis and prognosis.
Implement and validate models for feature extraction and fault detection and isolation or diagnostics.
Implement and validate models for fault prognostics.
Design, implement and validate the framework
Stage 1: Establish base lines and literature review. Examine common practices, and discoveries in the field related to techniques and process, and common tools and infrastructures used. Risk assessment.
Stage 2: Identify research gap, and review aims. Data collection, establish requirements for the framework requirement. Data cleaning and preparation, data management plan.
Stage 3: Using a condition based maintenance process, begin data analysis. Feature extraction; fault detection and isolation; diagnosis and prognostics using novel "big data" tools and knowledge discovery techniques.
Stage 4: Results visualisation; framework implemented and validated.
The work aims at a novel approach to time series analysis for system and machine faults, integrating physics, engineering know-how and discovered data patterns. The work has been trialled on train engine performance behaviour, door response, and wheel slip, seeking early and robust detection, diagnosis and prognosis.
Publications
Bernadin NAMOANO
(2020)
FAULT DIAGNOSIS IN TIME SERIES DATA WITH APPLICATION TO RAILWAY ASSETS
Namoano B
(2020)
Change detection in streaming data analytics: A comparison of Bayesian online and martingale approaches
in IFAC-PapersOnLine
Namoano B
(2019)
Online change detection techniques in time series: An overview
Namoano B
(2022)
Data-Driven Wheel Slip Diagnostics for Improved Railway Operations
in IFAC-PapersOnLine
SHIMIZU MINORU
(2022)
Real-Time Techniques for Fault Detection on Railway Door Systems
Studentship Projects
| Project Reference | Relationship | Related To | Start | End | Student Name |
|---|---|---|---|---|---|
| EP/R512515/1 | 30/09/2017 | 29/09/2021 | |||
| 2203091 | Studentship | EP/R512515/1 | 20/09/2017 | 20/03/2021 | Bernadin Namoano |
| Description | -Models have been build for predictive maintenance with application to railways train equipment. -Fault types have been found in the data using the designed models |
| Exploitation Route | - Models built may be useful for detecting incipient fault or anomalies in streaming datasets. Therefore, it can be applied to many cross disciplinary dealing with anomalies, and surprising pattern detection |
| Sectors | Aerospace, Defence and Marine,Digital/Communication/Information Technologies (including Software),Financial Services, and Management Consultancy,Healthcare,Transport,Other |
| Description | The current findings are used to detect earlier degradation in trains assets such as passengers doors, train engines and wheel set to improve the passengers journey and reduce unplanned maintenance costs. |
| Sector | Transport |
| Description | EPSRC doctoral prize |
| Amount | £1,584,860 (GBP) |
| Funding ID | EP/T518104/1 |
| Organisation | Cranfield University |
| Sector | Academic/University |
| Country | United Kingdom |
| Start | 03/2021 |
| End | 03/2023 |
| Title | Doors datasets |
| Description | Data represents urban train condition monitoring raw data. Data contains, current and door motions during opening and closing periods. |
| Type Of Material | Database/Collection of data |
| Year Produced | 2020 |
| Provided To Others? | Yes |
| Impact | Doors datasets present current and and speed profiles for train doors. Doors present normal and faulty conditions which can be served as basis for other researchers to device methods for fault detection. |
| URL | https://cord.cranfield.ac.uk/articles/Doors_datasets/7327148/1 |
| Title | Engines datasets |
| Description | Data represents urban DMU trains engines condition monitoring data. |
| Type Of Material | Database/Collection of data |
| Year Produced | 2020 |
| Provided To Others? | Yes |
| Impact | Data present normal and abnormal engine operations |
| URL | https://cord.cranfield.ac.uk/articles/Engines_datasets/12378053/1 |
| Description | Partnership with Instrumentel |
| Organisation | Instrumentel |
| Country | United Kingdom |
| Sector | Private |
| PI Contribution | Data provided by the company helped for the demonstration cases studies in the thesis. |
| Collaborator Contribution | The company helps understanding data gathering context, signal description and several uncertainties. |
| Impact | Outcome of the the collaboration includes: - Train door fault diagnosis - DMU engines fault diagnosis - Wheel slide recognition |
| Start Year | 2017 |
| Description | IFAC Workshop on Advanced Maintenance Engineering, Services and Technology |
| Form Of Engagement Activity | Participation in an activity, workshop or similar |
| Part Of Official Scheme? | No |
| Geographic Reach | International |
| Primary Audience | Professional Practitioners |
| Results and Impact | More than 100 hundred postgraduate and professional practitioner attended to the AMEST workshop, which sparked questions and discussion afterwards, the workshop reported high interest in engineering application of machine learning techniques as well as data framework architectures and data management tools |
| Year(s) Of Engagement Activity | 2020 |
| URL | https://www.amest2020.eng.cam.ac.uk/ |
| Description | UkRRIN conference |
| Form Of Engagement Activity | Participation in an activity, workshop or similar |
| Part Of Official Scheme? | No |
| Geographic Reach | National |
| Primary Audience | Industry/Business |
| Results and Impact | Poster presentation was shared with delegates. |
| Year(s) Of Engagement Activity | 2019 |
| URL | https://www.ukrrin.org.uk/building-uk-rail-capabilities-ukrrin-annual-conference-2019/ |