Share this page on:

Conference programme 

Back to the programme printer.gif Print

Poster session

Lead Session Chair:
Stephan Barth, Managing Director, ForWind - Center for Wind Energy Research, Germany
Frank O'Connor ServusNet Informatics, Ireland
Co-authors:
Frank O'Connor (1) F P Paul Leahy (2) Richard Kavanagh (2) Des Farren (1)
(1) ServusNet Informatics, Cork, Ireland (2) University College Cork (UCC), Cork, Ireland

Printer friendly version: printer.gif Print

Poster
Download poster(0.75 MB)

Presenter's biography

Biographies are supplied directly by presenters at OFFSHORE 2015 and are published here unedited

Frank O'Connor is co-founder and director of ServusNet Informatics, an Irish software company developing Operations & Maintenance (O&M) and Operational Intelligence (OI) solutions for the renewable energy industry. The ServusNet solution is deployed at multiple sites across North America and has data on close to 1 GW of assets. Frank was recently awarded a prestigious scholarship from the Irish Research Council to pursue PhD work developing new technologies for evaluating and reporting wind turbine availability. Frank is a contributing member to the IEA Task 33 Reliability Data: Standardization of Data Collection for Wind Turbine Reliability and Maintenance Analyses

Abstract

Making sense of wind-farm data - a novel semantic framework for interrogating heterogeneous wind-farm data -

Introduction

In the first six months of 2014, Europe fully grid connected 224 offshore wind turbines. Each turbine will produce data at regular intervals describing various aspects of turbine behaviour (energy production, operational state, etc.) In addition to the data produced by the turbine a vast array of operational data, e.g. maintenance data, environmental data, etc. will be generated. Storing and more importantly organising this data in a way that can be used to discover cost savings patterns is a considerable challenge. This paper presents a novel wind-farm semantic model designed to effectively link, structure, organise and interrogate wind-farm data.

Approach

The proposed wind-farm semantic model was developed using a modified version of the “METHONTOLOGY” methodology. The modification involved using Unified Modelling Language (UML) notation to model the domain and resulted in a generalised intermediate model. This Model was developed with inputs from wind-farm experts and includes comprehensive definitions for not just wind-farm data, but also for wind-farm activities and key wind-farm “actors”. The intermediate UML model was then used to create a domain-specific semantic model. Suitable existing ontologies were identified and reused, and when new domain-specific ontologies were created they referenced existing standards and published best practices.

Main body of abstract

A large wind-farm, like any large generation plant, produces significant quantities of data. Storing and more importantly organising all of this data in a consistent manner is a considerable challenge. In order to complete sophisticated data analytics, which is a key driver for lower wind-farm costs, it is not only necessary to have knowledge of turbine measurements and events, but it is also critical to have some knowledge of the context in which these signals were generated. For example, where the asset is located, what is the asset’s maintenance record, what is the terrain like, etc.? This paper provides a framework for answering these questions using semantic data. Semantic data consists of collections of concepts and relationships where a concept is any idea or topic that has meaning in a particular domain. A semantic model consists of terminological data, which is data that defines classes, properties, and relationships, as well as assertion data, which represents the individual class instances. This separation of assertion and instance data facilitates the creation of a detailed generic wind-farm model. This model not only provides a way to enforce the consistency of the instance data, but also allows for the querying of data in a more structured and meaningful manner. The wind-farm semantic model proposed in this paper has been created with inputs from wind experts and developed using state-of-the-art methodologies and tools. Specific model extensions have been developed which has enabled key industry standards and best practices to be formally codified into the model.

Conclusion

This paper addresses the need for better information management solutions to help make sense of growing quantities of wind-farm data. It will be shown that a modified ontology development methodology can be used to create a comprehensive domain specific wind-farm semantic model. Additionally, it will be shown that this model reused relevant ontologies from other domains and also how new model extensions have been developed specifically for wind power generation facilities. These extensions formalise existing industry standards and best practices and facilitate the efficient and effective organising, linking and interrogation of separate but related data in a semantically rich fashion.


Learning objectives
The reader will be introduced to the key concepts of semantic knowledge engineering and how these concepts can be applied to the wind energy domain. The reader will see how a formal methodology can be used to create wind-farm ontologies and obtain a working knowledge of how semantic technologies can be used to solve practical problems, such as how to efficiently perform sophisticated, domain-aware queries on wind-farm data.