World Class Wind Turbine Maintenance Operational excellence dankzij Condition Based Maintenance

World Class Wind Turbine Maintenance Operational excellence dankzij Condition Based Maintenance

Kamperland 27 Mei 2014 Twan Kohlen CEO Oliveira International Group

Data – Informatie - Kennis

Doelstelling

 Ontwikkelen en uittesten integrale control room voor het optimaliseren van O&M (technisch, organisatorisch en economisch)  Merkonafhankelijk, generiek meetsysteem, met een open, onafhankelijke architectuur  Low cost 10% - 20% huidige CBM systemen met veel meer functionaliteit

 Geen analyzers maar smartphone’s en tablets  Onafhankelijk windturbine cluster dat klanten kan ontzorgen op het gebied van O&M  Voorspelbaar maken van operationele performance, onderhoud, levensduur en opbrengsten  Windturbines met ‘bestaansrecht’, los van subsidies

Eindresultaat in modelvorm Dashboard Monitoring

Reports

Alerts

Analytical Predictive Model Workbench Modeling Management

Engineering Maintenance Intelligence

Business Rules & Processes SCADA data Maintenance Records Process & Asset Data (origin, options, etc.) Historian Data Error logs

Financial/Risk Data Reliability Reliabilit Information y Store

Informati on Store

Operator Data & Complaints Inspection Records Technician/Engineer Data Condition Monitoring data

Data combineren in ėėn systeem 12 verschilllende systemen en 200+ data ingangen Klimaat data Windsnelheid, luchtdruk, relatieve vochtigheid, windrichting, temperatuur etc

Proces data druk, temperatuur, vermogen, snelheid, belasting, bladhoekverstelling etc

OEM data tekeningen, onderhoudsadviezen, logistieke voorraad, gebruiksadviezen, modificatie standaard, etc

Onderhoud en storingen data aantal storingen, stilstand tijd, planning onderhoud, modificaties, MTBF etc

Manuele (gedigitaliseerde) data

CBM data

gehoor, visuele inspectie, lekkage, scheuren, verkleuring, etc

trillingen, ultrasound, infrarood, olie kwaliteit, resonanties, corrosie etc

Inventarisatie problemen en wensen 3G/4G – ethernet – satelliet

data

          

Ijsafzetting turbine bladen Tandwielkast schade Scheurvorming turbine bladen Towerbending Resonaties en structuur schade Gevolgen bliksem inslag Voorspellen van onderhoud Logistieke gevolgen van onderhoud Availability vs productivitie Turbine rendement absoluut Weer voorspelling vs productie …. Etc….

Data intelligentie

Stella Kapodistria

Stella Kapodistria

Condition Based Maintenance

Algorithms

CAMPI project Goal of the project: to develop service strategies that minimize cost whilst maximizing the availability and safety of assets.

Maintenance & logistics planning

WP I: Data Pooling, Eindhoven University of Technology

WP II: Maintenance Clustering, Groningen University

WP III: Integration and Implementation, Tilburg University Funding: Companies 26%, Universities 26%, Dinalog 48%

WP I: Data Pooling, Eindhoven University of Technology

Pooling of data means statistical analysis using multiple data sources related to multiple populations.

There are three fundamental objectives:  Aggregation in order to obtain more precise estimates.  Comparisons of trends and differences across populations and times.  Common interpretation of statistical information from multiple sources (similar or different populations). Taxonomy

Data Pooling: Taxonomy

Intra pooling (a.k.a. sensor fusion)

Different sensor readings on the same asset/component

Inter pooling

Similar sensor readings on different, but comparable, assets/components

Data Pooling

Dynamic mind map with literature links Literature overview paper

Projects with Oliveira

Projects with Oliveira

Data monitoring, fault detection and diagnostics

Out-of-control action plan (OPAC)

Estimation of the time to failure

Estimation of the initial state of a wind turbine

Industrial testbed for applying methods from the literature

Improve Oliveira procedures

Projects with Oliveira Current status:  data collection  data analysis (OMIS workstation)  control charts (X-bar charts)  identification of trends  diagnose a possible failure Data monitoring, fault detection and diagnostics

Intra pooling

Goal: Multivariate Statistical Process Control  joint control charts

 derivation of tighter, joint thresholds  business economic aspects of MSPC

Projects with Oliveira Current status:  Expertise  Monitoring (OMIS workstation)

Out-of-control action plan (OPAC)

Goal: creation of a flowchart for the out-of-control action plan  activators  checkpoints  terminators

Projects with Oliveira

Goal:  Identification of critical failures

Estimation of the time to failure

 Analysis of failure data  Analysis of propagated failures

 Derivation of failure time distribution

Projects with Oliveira Current status:

“Every wind turbine is unique”

Goal: Estimation of the initial state of a wind turbine

 Identification of comparable components  Creation of in-control indices  Creation of a flowchart  Structured algorithmic approach

Projects with Oliveira

Data monitoring, fault detection and diagnostics

Out-of-control action plan (OPAC)

Estimation of the time to failure

Estimation of the initial state of a wind turbine

Industrial testbed for applying methods from the literature

Improve Oliveira procedures

Stella Kapodistria

Condition Based Maintenance

Thank you very much! Questions?

Algorithms