By Gareth Brown, CEO, Clir Renewables
New research released in 2019 showed wind speeds had increased across North America, Europe and Asia since 2010 — and the trend is expected to continue. This study featured in the journal Nature Climate Change offered a stark contrast to the majority of previous research, which has demonstrated a long-term reduction in wind speed. Naturally, this has resulted in some excited assertions of a possible boom for wind power in the near future.
It’s certainly stating the obvious to suggest that a strong wind resource is fundamental to the success of wind power as an energy generating technology. However, the assumption that this potential change in wind speed will have a significant effect on the wind industry must be taken with a pinch of salt — particularly as the increase in wind speeds plateaus at 3 m/s, far from the 5 to 7 m/s that turbines require to generate electricity.
The suggestion of increasing wind speeds in the near future cannot be depended upon as a means to increase annual energy production (AEP). To maximize AEP across their portfolios, asset owners must focus instead on increasing their understanding of each asset and how they relate to the surrounding environment. Taking an AI-driven, contextual approach to asset optimization will maximize and secure project returns.
Wind vs. turbine
Wind power differs from oil, gas or coal power generation in countless ways, but a key contrast is in terms of consistency of resource intake. For coal, oil and gas, an operator has control over intake. This makes it far simpler to identify underperformance. When generating energy from wind, however, there’s no control over resource intake. Wind resource can change completely in a short period of time, making it much more difficult to understand the reason for underperformance: has there been a drop in wind speed, or is there a fault within the turbine?
The indicator most operators use to assess asset performance is “turbine availability.” Availability is a simple indicator of an asset’s reliability and potential for energy generation — after all, if the turbine is turned off it will not produce energy. However, “availability” tells owners nothing about whether the asset is performing as it should be with the current wind resource, or why it isn’t.
When availability is used to benchmark performance, longstanding underperformance issues are labeled as wind speed variation or missed entirely. This means that operators do not recognize low-level performance issues caused by faults within the turbine or misalignment, for example. A complete understanding of underperformance — and the extent to which it is due to resource versus technology — will empower asset owners and operators to increase AEP by up to 5%.
Context is key
Low wind speed is not the only environmental factor that can reduce energy production. There are a number of other contextual influences acting on a wind turbine that can impact performance:
- Ice buildup on turbines during colder months leads to increased loads on the blades, making it more difficult for the rotor to turn.
- Turbulence and blockage caused by nearby buildings, trees or other wind turbines can drastically alter the energy yield in comparison to initial forecasts.
- Imposed curtailment due to noise, the grid or local wildlife.
While SCADA data can indicate whether a turbine is underperforming, asset owners must understand how the environment is affecting each of their assets in order to find the root cause. To identify environmental issues such as blade icing or blockage as the source of any anomalies or patterns of minor underperformance, data collected from the turbine must be set in context.
Identifying and analyzing these environment-based issues requires digitizing the surrounding environment. This level of asset understanding is almost impossible to achieve through traditional data analysis methods. However, advanced, deep domain methods of data analysis using machine learning and AI can analyze and compare data streams from within the turbine and the surrounding environment and flag problems and quickly advise on their solutions. This allows operators to act on issues before they significantly curtail performance.
A fine line
Misalignment of pitch and yaw are common sources of turbine underperformance. This should make resolving misalignment a priority for asset owners looking to maximize portfolio performance — after all, angling the nacelle away from the wind by as little as 4º can reduce AEP by 1%.
Yaw misalignment is one of the simplest fixes to increase AEP as it is often due to sensor or controller error that will be rectified by replacement of the faulty part. However, identifying misalignment can be a complex, long-winded process, requiring operators to sift through a significant volume of data comparing the performance of individual turbines to others across the wind farm.
In contrast, using digital tools to compare power curve data between peer turbines to identify whether pitch or yaw has been misaligned massively reduces the time-cost involved in analyzing the data. As such, owners can identify and fix misalignment before it has a significant effect on AEP.
Sometimes wind farm underperformance is not due to the influence of environmental factors or issues within the turbine itself, but is the fault of an overly conservative derating strategy that curtails turbine output at a fleetwide level. While derating is a useful strategy to prevent blockage effects and increase asset lifetime, many derating strategies lower turbine performance to an excessive extent.
In order to find the optimal balance between preventing blockage effects and maximizing production, advanced analytic methods evaluate data from multiple sources within and around the turbine to provide asset-by-asset recommendations. By taking a tailored approach to their assets, owners can significantly improve annual energy production while maximizing lifespan.
Understanding how assets interact with the environment and finding potential opportunities for optimization will improve far more than turbine output. A complete understanding of assets will lead to greater certainty around performance, supporting increased risk management capabilities. This can lead to lower insurance premiums, the ability to borrow more capital and achieve greater financial returns. Greater understanding of asset performance allows for certainty around predicted output for investors, setting asset owners up to secure more favorable project financing.
In short, wind resource — whether it is increasing or decreasing — isn’t the be-all and end-all of energy production. Asset owners need to move on from the predictions and claims long-term wind speed studies generate and focus on optimizing their portfolio — now. By understanding their assets’ relationship with whatever wind speed each is afforded, owners will future-proof their portfolios and ensure they perform at their best, whichever way the wind blows.
Gareth Brown is CEO of Clir Renewables, a renewable energy AI software company. He is an entrepreneur and a chartered engineer with the IMechE. Gareth has over a decade of experience in the industry which spans the life cycle of renewable energy projects from identification, development, construction, through to financing and operation. Gareth set up Clir Renewables in 2017 alongside Jake Gray, offering an innovative, AI-driven software solution to help owners and operators better understand how their assets are performing and provide actionable insights on how to optimize their output. Headquartered in Vancouver, Canada, the company opened its European office in Glasgow, Scotland, in 2018, and now supports 6 GW of assets globally.