Using machine learning to detect underperforming turbines

Using machine learning to detect underperforming turbines

It’s not always clear if a turbine is underperforming, but machine learning can assist in finding out.

VANCOUVER, BC: Clir Renewables has released a new product feature which automatically detects underperforming assets and highlights key actions to rectify underperformance.

A wind turbine will never generate its expected output one-hundred percent of the time, and its performance can and almost certainly will change over time. There are various reasons for this. Some are known but cannot be controlled or managed, such as fluctuating inflow conditions. However, occasionally the opposite is the case, the fault is unknown but can be controlled or rectified. With datasets full of noise from the known reasons, is it possible to extract data to identify the unknown causes? Yes, is the answer.

Using layered machine learning, built on an advanced data model, Clir Renewables has created an underperformance detector for its software solution. This detector works along with other algorithms in the software to analyze the data and classify them based on the reason for the underperformance. The detector creates a synthetic event when turbine power output is well below the historical mean for that wind speed. This invaluable piece of information helps identify ongoing issues at a turbine, not indicated by the SCADA data, inflow conditions under which the turbine does not perform well, and the duration and lost energy associated with the underperformance. It also highlights a hardware or software configuration change that reduces power performance. The detector removes the noise leaving a clean set of data from which the unknown causes can be deduced, and corrective actions created. Alternatively, if the cause is still unidentified, the owner can approach the manufacturer with the cleaned data looking for answers and solutions.

Selena Farris, Data Scientist at Clir Renewables, said “With noise filled datasets the uncertainty of any conclusions that can be drawn on causes of underperformance will increase significantly, and in a lot of cases, issues can be completely masked by the noise. Utilizing the advances in machine learning, a well-structured data model, and deep domain expertise Clir software provides a tool to reduce this uncertainty, generating actionable insights for owners to increase performance and protect their assets from faults and failures.”


About Selena Farris: Selena has over ten years of renewable energy industry experience with a masters degree specializing in wind energy. Working as a developer, consultant, and utility has given her experience and insight into all stages of the wind farm life cycle. She has managed multiple project portfolios and met campaigns, advising on the most financially responsible development path.

Meet the team: Matthew Lynn

Meet the team: Matthew Lynn

Who is Matthew Lynn?


Matthew is our Director of Product, turning our ideas, and those of our clients, into reality.


Having graduated from the University of Manchester (UK) with a degree in physics, Matthew’s career began as a software developer.  A change of tack saw him enjoy some time out travelling before studying renewable energy systems technology at Loughborough University.  From there, he joined UK wind energy consultancy Garrad Hassan (which later became GL Garrad Hassan and finally DNV GL), where he managed teams engaged in wind farm development across Europe and Canada and set up a company office in Vancouver BC.  His interest in tools development led him to found a new SaaS product managing wind and solar resource measurements, and a role managing the development of wind farm and solar PV resource assessment software with responsibility for the technical vision, business case, financial forecast and budgeting of software created.


Matthew joined Clir in 2018 and manages our development teams using Agile/Scrum methodology. With lead responsibility for Product, he ensures that the ideas and ambitions of our clients and the Clir team become a reality in a practical and effective fashion. As a talented and highly experienced renewables and software engineer Matthew is a crucial member of the team.

Interested in joining Matthew and the Clir Renewables team?

Clir makes noise in France

Clir makes noise in France

VANCOUVER, BC: Canadian optimization software company, Clir Renewables has secured its first project in France following the recent opening of its first office in Europe. The client, a GW scale owner-operator in the region has chosen Clir to support optimizing the performance of their current wind portfolio.

Having considered other providers in the market, Clir was chosen by the owner specifically due to the system’s ability to provide actionable items to increase production and hold the OEM to account.

France offers an excellent opportunity for Clir Renewables to expand its reach within Europe. At the end of 2017, France Energie Eolienne reported there was 13.5 GW of onshore wind installed. In 2018 the French Government released their Multiannual Energy Programme (programmation pluriannuelle de l’énergie, PPE) laying out their energy transition to a cleaner energy mix, including plans to triple installed onshore wind capacity by 2030. This commitment to onshore wind shows the market is strong and will grow steadily over the coming decade.

Craig McCall, Director of Europe “In recent years we have seen a shift in owners focusing more on their operational assets in France. A significant contributor to this shift is due to the recently imposed regulatory requirements surrounding noise impact in 2011. The change allows the undertaking of a new noise impact assessment at any point either as a result of a residential complaint or ICPE inspection. This leaves many owners exposed as non-compliance of a wind farm can be revealed after several years of operation.

It’s great to see a large owner-operator in the region having the confidence to bring on Clir despite having their own extremely competent team of engineers. Our view as a business has always been to provide clients with the tools to conduct analysis quickly and at scale to optimize their assets.

Using Clir allows owners to bypass time-consuming analysis by clearly quantifying the impact of all potential areas of underperformance. This allows their technical teams to focus on implementing the performance improvement measures that have the highest return on investment, as opposed to getting caught up trying to make sense of the poorly structured data provided by the OEM’s.”


About Craig McCall: Craig has over six years’ experience in the renewable energy industry and most recently led a world-leading renewable energy technical consultants Optimisation and Control service line. To date, Craig has been involved in more than 50 wind farm optimization projects across Europe, Asia, and the Americas.



Simplify your GADS reporting

Simplify your GADS reporting

The compulsory GADS-W reporting can be a time-consuming and cumbersome task, however, advances in software capability can drastically reduce this to produce a one-click report.

VANCOUVER, BC: Clir Renewables software’s suite of capabilities includes functions that automate GADS reporting, reducing the time and labor required to prepare Generating Availability Data System, Wind (GADS-W) reports.

In 2017, North American Electric Reliability Corporation (NERC) introduced GADS-W reporting. It was only in 2018 that the NERC made the GADS reporting mandatory for wind farms with 200 MW or greater total installed capacity. This year there are new requirements. As of January 1, 2019 operators of wind farms with 100 MW or greater total installed capacity are required to compile reports on a monthly basis and submit those reports to NERC each quarter. And sites of greater than 75 MW will be required to report from January 1, 2020.

With less than a month to submit the Q4 2018 GADS report, analysts across North America are scrambling to find time to prepare reports to meet the February 15 deadline.

Operators are required to submit two reports. First, the Performance Data Report, which summarizes overall wind turbine operation in a particular month. Second, the Component Outage Report which identifies the area or reason for lost production as reported in the Performance Data Report. NERC uses the information in these reports to calculate wind farm performance, reliability, and availability statistics. The requirements in producing these reports are challenging with complex rules and conditions that must be considered when allocating downtime to appropriate GADS-W reporting categories.

For many wind farms the data streams available to operators to produce these reports, e.g. OEM SCADA, are not set up or designed to comply with GADS-W reporting directly. Also, operators generally don’t have the tools to process their available data streams. These structural challenges leave operators faced with the periodic, time-consuming and complicated task of manually processing and editing available data streams to try and produce GADS-W reports. To address these specific challenges, Clir Renewables has developed a GADS-W Reporting tool which makes generating and submitting reports to NERC a straightforward task.

Leveraging Clir’s industry-leading proprietary data model, the software processes turbine SCADA data streams to seamlessly and automatically produce GADS-W reports, ready for submission to NERC, at the click of a button. It is a collection of functions working together that produce the one-click report. How does it all work? Firstly, SCADA data is processed by Clir to identify and categorize wind turbine downtime and lost energy accurately. Clients can then interact with the software to edit default allocations, a functionality we call ‘error code reconciliation’, to ensure downtime is accurately categorized. In addition, Clir has developed software to ingest documents like service reports, to ensure the reasons for downtime and lost production are accurately described on the platform. Clir’s software also identifies sequences of status codes automatically grouping these into outages enabling easier editing of event allocations and maps status codes to the appropriate default GADS-W reporting allocation. All of these functions drastically reduce the amount of time required in preparing GADS-W reports.

Dr. Shane Butler, Data Scientist at Clir, stated, “Generating the monthly reports on a quarterly basis is a time-consuming and onerous task for an operator to produce. Mistakes due to human error and misunderstandings in allocating downtime can be expected when dealing with such large amounts of data, especially at larger wind farms.

“Using Clir’s software to process the data required for GADS-W reports, not only reduces the time required but also ensures greater accuracy,” Butler continued.

In 2018 Clir Renewables’ clients with wind farms over 200 MW of installed capacity utilized the Clir GADS-W Reporting tool to produce their reports, easily meeting each deadline and will do so again for the February 15 deadline for the final quarter of the 2018 production year. Feedback on the tool was all positive, with clients telling of their delight at the ease with which the reports are produced and the amount of time saved against manual production of the reports.

Deadlines for the 2019 production year are May 15 for the first quarter, August 15 for the second quarter, November 15 for the third quarter and February 15, 2020, for the fourth quarter, mark them in the calendar now.


About Dr. Shane Butler: Shane is a recognized industry expert in the application of data science skills to real-world problems to enable and support informed decision making. For the past five years, Shane has been applying his big data analytics skill set within the domain of wind farm operations, performance analytics, and predictive maintenance. Expert user of Matlab, Python, Tableau, and Cloud-based computing platforms.


Image: Adobe Stock

Have you validated your turbine upgrade plan?

Have you validated your turbine upgrade plan?

As wind turbine design advances and existing fleets age, there is a case to be made for implementing system upgrades, but how do you know if it is worth the initial investment?

VANCOUVER, BC: Clir Renewables has released a new software feature to support the validation of turbine upgrades. Whether it is a software or hardware upgrade validation needn’t only be theoretical.

As the design of wind turbines advances reviewing the performance of the existing fleet can highlight areas where improvements can be made by upgrading the current hardware or software. Original Equipment Manufacturers (OEMs) and third-party companies are regularly offering owners ways to improve the performance of current turbines. These upgrades can be physical changes to the turbine like aerodynamic enhancements such as vortex generators, gurney flap or leading edge tape. Software-based changes could include new blade pitching algorithms, improved yaw control or increased cut-out and re cut-in wind speeds. Loads on turbine components may be increased, depending on the nature of the upgrade. The benefit of increased Annual Energy Production (AEP) should be carefully weighed against the implications of increased loading.

As owners know, no two sites are the same so how can they be sure investing in these upgrades will increase AEP? By validating the impact of the upgrade against the claimed gain by the manufacturer. However, independent validation studies on specific upgrade options are scarce, or not applicable to every site. Without independent validation, owners may never know the true benefit of making the upgrade.

Independent validation is possible by implementing the upgrade on a small percentage of turbines on site, assessing the impact on AEP and evaluating it against the investment required to install the upgrade across the whole farm. Clir software has the most advanced data model in the industry which makes this validation easier to complete than via traditional methods. By integrating SCADA, CMS, meteorological, geospatial data and much more to build predictive models of turbine behaviour, Clir software can identify changes in a turbine’s performance and benchmark it against numerous models. An assessment of the cost-benefit of the potential upgrade is then performed before full implementation.

Additionally, Clir Renewables has extensive knowledge of upgrades offered by OEMs. The data science team has conducted in-depth reviews of documentation provided by OEMs as well as contributed to conversations about upgrade implementation at many wind farm sites. Clir Renewables provides independent and impartial information based on site-specific data and OEM documents.

Andrew Brunskill, Data Scientist at Clir Renewables, said “There are few simple answers to turbine upgrades as comparing two wind farm sites to each other is like comparing an apple to an orange.

“The impact of each upgrade on energy production and turbine loads depends on the local wind climate and the specific turbine model in use at the site, among other factors, that’s why independent validation is important,” Brunskill continued.

By using Clir software to validate upgrades owners avoid expensive consultancy fees and have the information to negotiate confidently with the OEM or third-party company on the cost.


About Andrew Brunskill: Andrew is a PEng from a major global consultancy and has taken on a wide variety of roles during his career including project engineer, energy analyst, modeling specialist, and project manager. In recent times he has undertaken performance analytics across Vestas, Siemens, GE, Suzlon, and Siemens wind turbine technology.


Image: Adobe Stock

Is your wind turbine performance lost in the forest?

Is your wind turbine performance lost in the forest?

Turbulent wind inflows increase fatigue loads on wind turbines, lowering life expectancy, impacting performance and reducing annual energy production (AEP).

VANCOUVER, BC: Clir Renewables has developed renewable optimization and reporting software to quantify the impacts of forestry on wind farms. The impact of forestry on inflow conditions is often poorly accounted for when developing wind farms, and once a project is operating, there can be a significant physical effect on the life expectancy of the turbines and an adverse effect on performance.

For wind farm owners, wind turbine blades effortlessly gliding through the air when met by smooth, undisturbed inflow conditions is easy on the equipment. However, projects situated in or near forested terrain typically present additional challenges. As the wind passes over forestry canopies, wind flow is disturbed, generating more complex conditions such as turbulence and shear, creating slower uneven loads on the blades. This turbulent air stresses the blades in differing proportions increasing the level of wear and tear across wind turbine, including but not limited to accelerated blade cracking and gearbox and generator failures, reducing life expectancy.

Until now owners have not had the tools to quickly and easily understand the impact of forestry at a wind farm, often relying on often prohibitive, expensive studies to assess the impact. Clir software is built on a domain-specific data model that takes account of geospatial considerations, such as forestry, that drive turbine inflow conditions. By utilizing the geospatial information and atmospheric stability, performance issues at the turbine itself, and in the inflow conditions, can be more accurately parsed out to enable owners, operators and other wind farm stakeholders to understand project underperformance and performance risk for specific items such as forestry.

“Many owners are likely experiencing generation losses due to forestry in the area of their wind farms; the degree to which they may be unaware. Additionally, the disturbed wind inflow may be increasing turbine loading, reducing their operational life. Additional operational costs associated with premature component failures and reduced turbine service life can surprise owners who may not have accounted for this in their financial model,” said Andrew Cameron, Data Scientist at Clir Renewables.

“By assessing this early in the wind farm’s operational life many of the costs and losses can be better understood, accounted for and potentially reduced.” Andrew continued.

Once an owner is aware of the forestry impact on AEP and loading conditions, a cost-benefit analysis can be formulated to identify the appropriate next steps. When implemented the optimization recommendations can produce increases of around 5% AEP for the wind farm and extend asset life by up to five years. In situations where a wind farm’s forestry is poorly managed gains of almost 10% AEP, and 5+ years of life extension are possible.

Clir CEO Gareth Brown says: “The wind industry is unique. It’s the only power source we don’t know what the inflow conditions are. The industry to date has tried to build digital and physics-based models, but they are limited in scope and capabilities as they fail to take into account the inflow conditions. By digitizing beyond an individual turbine and what drives their inflow conditions enables Clir to more accurately build effective advanced analytics and AI tools to identify where performance can be improved.”

Brown continues, “The challenges defining performance issues due to no inflow measurement are compounded by, poorly structured SCADA data, inadequate industry performance tools and metrics that focus on the question is the turbine spinning rather than performing, and conflicted service providers or OEMs who are focused on protecting their internal know-how rather than building products to enable effective performance management. The Clir platform circumvents this by building a domain-specific data model that gives owners the tools to maximize their understanding of the asset, where lost energy is occurring and get clear actions to mitigate the lost energy for items like forestry and many others.”

About Andrew Cameron: Andrew has a masters degree in mechanical engineering and over ten years’ experience in the wind energy industry. His expertise includes both pre-construction analysis as well as operational analysis for a wide variety of wind regimes, turbine technologies, data collection techniques, and analysis goals.


Image: Adobe Stock