Optimisation of Wind Energy O&M Decision Making Under Uncertainty
Lead Research Organisation:
University of Strathclyde
Department Name: Electronic and Electrical Engineering
Abstract
Wind energy will play a full part in decarbonisation of the future energy mix - if the costs can be reduced. This project develops a technological concept that helps achieve that cost reduction, by utilising data in a way which directly supports quick and reliable decision making in the everyday operation of a wind farm, either on- or offshore. The volume of data available from wind turbine assets is staggering - from component temperature traces, to weather forecasts, to sea conditions. But ultimately that data needs to be used by a control room engineer to change a decision in order to be useful. This innovative project develops a decision-making system that combines advanced visualisation methods and component health systems developed by UK SMEs with decision-theory from academia, and brings this together in a way that a wind farm operator can utilise to drive down the cost of operating a wind farm.
Planned Impact
see main InnovateUK energy catalyst grant proposal for detail
Publications
Dawid R
(2018)
Decision Support Tool for Offshore Wind Farm Vessel Routing under Uncertainty
in Energies
Gilbert C
(2018)
A Hierarchical Approach to Probabilistic Wind Power Forecasting
Gilbert C
(2021)
Probabilistic access forecasting for improved offshore operations
in International Journal of Forecasting
Gilbert C
(2019)
A Data-driven Vessel Motion Model for Offshore Access Forecasting
Rubert T
(2019)
Wind turbine lifetime extension decision-making based on structural health monitoring
in Renewable Energy
| Description | We have developed a decision making methodology which takes uncertainty and temporal dependency & structure into account. Our illustrative example is a wind turbine blade replacement decision - windy sites in the UK & elsewhere rarely drop below the crane lifting limit for blades (winds of 7 metres per second or below are needed to lift a blade). The key uncertainty here is the 1 week weather forecast, which the decision is based around. Currently there is no method to directly translate weather forecast uncertainty into the key decision (what day to hire crane and perform lift) - this decision would today be made deterministically (that is to say, not taking uncertainty into account). Our developed decision making methodology has shown that by taking uncertainty into account, it is possible to reduce cost and risk when making key decisions such as this. This key finding is being disseminated throughout the wind industry and will form the basis of future higher TRL work. The ideas have been built on in the follow on project ORACLES (https://oracles.eee.strath.ac.uk/). THis has developed some of the concepts tested in this grant and applied them to a new application (offshore wind access forecasting). |
| Exploitation Route | The methodology for structuring similar decisions under uncertainty will be published in an upcoming journal paper (IEEE transactions on engineering management). In principle this model applies to any decision under uncertainty. The next logical step is application to operational decisions when considering wind turbine condition monitoring data feeds and uncertainty regarding remaining useful life, accuracy of prognostic algorithms. We are in the process of making our code publically available on a platform that will support all future forecasting and decision support models developed by our research team. |
| Sectors | Energy,Environment |
| URL | https://pure.strath.ac.uk/ws/portalfiles/portal/79626897/McMillan_Browell_2017_Optimisation_of_wind_energy_O_and_M_decision_making_under_uncertainty_report.pdf |
| Description | This methodology was built upon in the ORACLES supergen wind project (2019) which was then fpollowed by a live demo project at Teesside (2021-22) funded by EdF (https://www.edfenergy.com/for-home/energywise/research-development-safe-vessels-renewables). Edf estimate the annual cost savings could sum to £1m per annum per wind farm. |
| First Year Of Impact | 2021 |
| Sector | Energy |
| Impact Types | Economic |
| Description | ORACLES - Offshore Renewables Accessibility for Crew transfer, Loss Estimation & Safety. Supergen Wind Flexible Fund R3 |
| Amount | £120,000 (GBP) |
| Funding ID | ORACLES (Supergen Wind Flexible Funding) |
| Organisation | Engineering and Physical Sciences Research Council (EPSRC) |
| Sector | Public |
| Country | United Kingdom |
| Start | 02/2018 |
| End | 03/2019 |
| Description | KTP with Onyx Insight Ltd |
| Organisation | ONYX InSight |
| Country | United Kingdom |
| Sector | Private |
| PI Contribution | Intellectual leadership to KTP focusing on improved used of data analytics in the decision support/ OPEX control of operational wind turbine assets. |
| Collaborator Contribution | Provided domain expertise, problem specification, and large amounts of operational data. |
| Impact | none yet reported |
| Start Year | 2018 |
| Description | Onyx Advanced Work Order Data Analytic Tool |
| Organisation | ONYX InSight |
| Country | United Kingdom |
| Sector | Private |
| PI Contribution | Building on the relationship with Romax/ Onyx, my team was directly commissioned by Onyx to develop and implement an automated system for the analysis of wind turbine work order data. |
| Collaborator Contribution | Funder, user and vendor of developed system |
| Impact | The main output was a protoype work order analysis tool which was used by Onyx as part of their suite of data analysis tools. |
| Start Year | 2019 |
| Description | Optimisation of Wind Energy O&M Decision Making Under Uncertainty |
| Organisation | Romax Technology |
| Country | United Kingdom |
| Sector | Private |
| PI Contribution | Strathclyde team developed intellectual core of decision model & underpinning methods based on Real Options. |
| Collaborator Contribution | Expertise: Romax brought deep knowledge of wind turbine failure modes especially drivetrain. Data: Romax provided extensive data sets for testing decision models. |
| Impact | Outputs: 1) Prototype blade lift decision support tool. 2) Prototype main bearing life management decision support tool. Both were multi disciplinary and included statistics (strathclyde), forecasting (strathclyde), decision modelling (strathclyde), drivetrain expertise (Romax), operational planning (SPR). |
| Start Year | 2016 |
| Title | Onyx Insight Work Order Analytic Tool |
| Description | A tool which automates the recovery of usable and taxonomised data from unstructured hitoric maintenance records. |
| Type Of Technology | New/Improved Technique/Technology |
| Year Produced | 2020 |
| Impact | Model was presented at RenwableUK onshore 2020 |
| Description | Wind Energy Science Conference 2017 |
| Form Of Engagement Activity | A talk or presentation |
| Part Of Official Scheme? | No |
| Geographic Reach | International |
| Primary Audience | Professional Practitioners |
| Results and Impact | Wind Energy Science Conference 2017 |
| Year(s) Of Engagement Activity | 2017 |
| Description | A Heavy Lift Decision Support Tool |
| Form Of Engagement Activity | A talk or presentation |
| Part Of Official Scheme? | No |
| Geographic Reach | Local |
| Primary Audience | Other audiences |
| Results and Impact | A Heavy Lift Decision Support Tool |
| Year(s) Of Engagement Activity | 2017 |
| Description | Optimisation of Wind Energy O&M Decisions Making Under Uncertainty A Heavy Lift Decision Support Tool |
| Form Of Engagement Activity | A talk or presentation |
| Part Of Official Scheme? | No |
| Geographic Reach | Local |
| Primary Audience | Other audiences |
| Results and Impact | Onshore Wind 3 - Challenges and Opportunities |
| Year(s) Of Engagement Activity | 2017 |
| Description | Royal Statistical Society Conference 2017 |
| Form Of Engagement Activity | A talk or presentation |
| Part Of Official Scheme? | No |
| Geographic Reach | Local |
| Primary Audience | Other audiences |
| Results and Impact | Royal Statistical Society Conference 2017 |
| Year(s) Of Engagement Activity | 2017 |