Windlog: Preventing wind turbine failure with predictive analytics - Proxima Solutions
Preventing failures and damages through predictive analytics

Windlog: Preventing wind turbine failure with predictive analytics

Windlog, by Proxima Solutions GmbH, is a wind farm asset management software as a service solution (SaaS) designed to ease wind farms' monitoring and optimize energy production. Since the first version of Windlog was released in 2018, several relevant improvements have been made to help optimize wind turbine monitoring. Today Windlog counts with ten modules making it a powerful and holistic tool that allows portfolio and asset managers, technical-commercial managers, and O&M technicians to get in-depth information; helping them in their decision-making process and KPI’s achievement.

As part of Windlog improvements, we have developed artificial intelligence algorithms oriented to predict failures and outages. The wind turbine predictive maintenance service has served to early identify possible damages leading to savings of more than 1.1 million euros in one year and a half. Over 55 cases have been identified in a 500 MW portfolio. To make the user understand the findings better, we introduced 3 different severity classes:

  • Class 1 (green, watchlist)
  • Class 2 (yellow, check during next maintenance)
  • Class 3 (red, immediate action required)

The most frequent issues happened to be detected in the drive train - generators and the gearboxes), where 12.5% of all detected cases were grouped in severity class 3, accounting for over 80% of the realized savings.

The following real-life cases were estimated according to 2019 prices, and have been documented in detail in a pessimistic scenario, where core parts would have had to be replaced due to total damage. The estimated avoided expenses in each ofthe six following cases are the sum of:

  • Spare parts to be bought from the OEM or other suppliers.
  • Working hours to replace the component.
  • Expenses for a crane on-site and other tools and consumables.
  • Production loss due to the fact that the turbines had to be left idle until the day ofthe component replacement.

Showcase #1: Gearbox wheel

- In December 2019, Windlog detected an increasing trend in the gearbox oil and gearbox bearing temperature. An onsite check was programmed for February 2020.
- The maintenance provider analyzed the situation and confirmed damage on the gearwheel. An up-tower repair was programmed, and the turbine continued in limited operation until the day of the intervention
- Windlog notified the maintenance provider of the situation two months upfront allowing an early repair and avoiding expenses for a complete Gearbox replacementof €230’000.

Showcase #2: Generator bearing

- In May 2019, the temperature sensor of the generator bearing showed suspicious behavior. The digital twin showed a high-temperature trend for the turbine to be under observation.
- The generator bearing stood in observation for eight months. Turbine specialists suggested its replacement in January 2020, along with the maintenance of other assets.
- If the situation had remained undetected, severe damage would have led to a generator bearing replacement which could cost up to €220.000. The total spend of the intervention was only €6.000.

Showcase #3: Gearbox bearings

- During two months, the predictive analytics service showed an anomalous temperature in two gearbox bearings. The situation was identified in January 2020.
- The O&M service provider was informed of the issue, categorized with a severity class 3, and practiced an endoscopy.
- The decision taken was to replace both gearbox bearings before the entire gearbox would have been destroyed. The avoided expenses were approximately €190.000.

Showcase #4: Several temperature signals

- Temperature sensors identified peaks in different components like the main shaft bearing, generator bearings, and the nacelle.
- The O&M team evaluated the situation and expected the root cause to be situated either on the cooling or greasing system. The issue was categorized as a severity class 3.
- The root cause was the cooling conduit, which got detached. Unrecognized, the permanently too high temperatures could have caused severe damages estimated to be around €100.000.

Showcase #5: Gearbox cooling system

- In May 2019, an increasing trend in the oil and gearbox bearing temperatures was noticed.
- The FMEA categorization graph showed issues on the cooling system, specifically on the thermostatic valve, which was then replaced.
- If the issue persisted, the gearbox could have suffered major damages. The estimated avoided expenses were approximately €200.000.

Showcase #6: Generator bearing

- In May 2019, Windlog identified elevated temperatures in the generator bearings of one of the turbines.
- The situation did not require immediate action, so maintenance was planned for November 2019. During the repair, a blocked heat pipe was found and cleaned.
- Unrecognized, the high temperatures could have caused severe damage resulting in a generator replacement. The avoided expenses were between €200.000 and €250.000.

About us

Proxima Solutions GmbH is a German Company founded in 2018 aiming to digitize the energy transition. Combining artificial and human intelligence, data science, and renewable energy expertise, we offer a suite of software tools that enables asset owners and asset managers to increase energy production from their wind and hydro energy plants.

We can also support our customers with a set of services (plant supervision, predictive diagnostics, asset management), where we optimize asset performance and preserve asset lifetime by implementing the recommendations from our advanced analytics and AI predictive algorithms.