Wind Power Performance Onwards and Upwards – exploiting SCADA data

In this article, Matthew Zhang, Specialist – Energy Resource Services at Lloyd’s Register discusses the potential of utilising wind farm SCADA data to improve the performance of wind power generation.

Over the past 10 years, we have seen an exponential growth of wind power with global installed capacity tipping over 600 Gigawatts. Fleets of wind turbines are now pumping green electricity into the grid, powering up economies for a sustainable future. Coming with it is a colossal amount of SCADA (Supervisory Control and Data Acquisition) data sitting mostly dormant in owner’s secret storages which are rich gold mines ripe to be exploited. SCADA data is a record of everything from power produced to temperatures of individual components and everything in between. From SCADA data it is possible to build an accurate picture of many aspects of the performance or failure of a wind turbine. Digging into those gold mines of SCADA data, therefore, will help us realise the full potential of the existing capacity while creating a positive feedback loop to inform future design of wind turbines and wind farms. Here a few of the benefits of being able to take full advantage of the operational SCADA data are discussed.

Pre-construction prediction precision – reduced uncertainty

The investment decision of a wind power development is broadly-speaking driven by the energy yield prediction made pre-construction. Many assumptions have to be made in terms of losses and uncertainty which rely heavily on an individual analyst’s own professional judgement. The feedback from reality to date, however, is sketchy, to say the least. Take power degradation as an example, we have seen older wind farms whose power performance degrades by close to 1% per year, that is a ~10 % production loss over a 20-year period! Whilst the degradation could be reduced if sub-optimal performance were diagnosed early and thus rectified, it is still a dramatic risk that could keep any shareholders up during the night. The fleet SCADA data provides an invaluable window to what is really out there, so that the bias in pre-construction prediction, if any, can be corrected, reducing the risk and uncertainty in future investments.

Performance assessment – increased revenue

Every stakeholder of a wind farm will have an interest in knowing how the wind farm has been performing and if anything can be done to improve the performance. It is speculated that on average at least 5 percent more production can be gained if sub-optimal performance issues are rectified. The SCADA data recorded during the operational period provides a wealth of information that can help with this diagnosis. After removing the elements of windiness and production loss assumptions from the recorded performance, the production losses due to sub-optimal performance are exposed to scrutiny. The cause of the sub-optimal performance can be found through correlation to various parameters within or resulting from the SCADA data, such as pitch schedule and yaw behaviours. Any step changes and degradation of the trends and correlations of various signals can also lead to the areas where the performance of the turbine can be improved. 

Smart maintenance– reduced cost

The wind as a fuel is abundant and better, free of charge. Once construction is completed, the only operational cost of running a wind farm is the maintenance cost. Doing it smart requires informed decisions to be made. The SCADA data can be used to detect and predict failure of components, informing spare part inventories and maintenance schedules. It can prevent minor problems from escalating into major faults. This can be achieved through creating a risk profile of different turbines and components in addition to an understanding of Failure Modes. Besides lowering the maintenance cost, the increased reliability of the turbines brought about by such smart maintenance is of paramount importance for an offshore wind farm where accessibility is greatly limited by tides and waves.

Big data and machine learning – getting increasingly smarter

With the rapid development of the Internet of Things (IoT), data science and cloud-based computation we are now able to dig deep into the big data and use machine learning and deep learning algorithms to recognise patterns and phenomena and carry out diagnoses in ways we could not have imagined a few years ago. These gold mines can now be exploited at an industrial scale. The principal obstacle seems to be the unwillingness to share the data which is essential to train these algorithms. Saying that I am confident that the industry will come together one way or another in this matter driven by the competition of the market and the benefits of doing so.

It is truly an exciting time for the wind energy industry with tremendous potential to be realised both in terms of new installations and the performance of existing ones. If you are a stakeholder of an operational wind farm and have SCADA data sleeping in storage, it is time to do something about it.

Please click here to download a case study, and see how we have helped the Client understand an underperforming asset.