Acciona Energía

Customer Reference



 

An AI approach to control and monitor wind farm assets using satellite imagery

Acciona Energia, a subsidiary of Acciona based in Madrid, is a Spanish company developing renewable energy projects, including small hydro, biomass, solar energy and thermal energy and biofuels. Acciona also has assets in the field of co-generation and wind turbine manufacturing.

Acciona has 222 wind farms throughout the planet, some of them are in remote and isolated regions, in Mexico, Chile or Australia for example. These wind farms reach an availability factor of 97%, thanks to the usage of cutting-edge technology.

Acciona Energia client References
Acciona Energia client References
Download WindEurope's 2019
project poster

The Challenge: Acciona’s difficulties to monitor wind farm assets

Wind energy has become the strongest and fastest growing renewable energy technology worldwide over the last 30 years, thanks largely to recent technological advances and commercial growth. It now plays a central role in the immediate and longer-term energy strategies of many countries.

Reducing the cost of energy from wind to economically sustainable levels is the most important challenge facing the industry today. Operation and maintenance costs constitute up to 30 % of the total cost of energy from wind in large farms. The industry must overcome the challenges associated with improving component reliability and the development and adoption by operators of appropriate condition monitoring systems and maintenance strategies, in order to reduce costs to sustainable levels.

The control and monitoring of assets such as access roads, electrical substations, or wind turbines typically relies on in-situ human supervision and impacts the overall cost of Energy.

In addition to being located in remote areas, extreme weather can cause these parks to be completely inaccessible: access road cut as a result of a snowfall, a high risk of fire during a summer drought. The monitoring of these areas results increasingly expensive and complex.

On the one hand, the area surrounding the wind turbines is usually quite controlled, since it is the area most frequented by maintenance workers. On the other hand, the rest of assets, such as access roads, sub-stations and especially evacuation lines (high voltage), are less monitored and more exposed to external risks. In particular, the proximity of vegetation to these assets implies a high risk of fire, in addition to not complying with local fire regulations.

For all this, Acciona was seeking a remote and automatic control of its assets in the wind farms, which allows reducing costs and increasing the recurrence of monitoring.

The Solution: High precision automatic remote monitoring of assets using satellite imagery and Artificial Intelligence

Acciona Energía Innovation Department in collaboration with Pervasive Technologies, developed an automatic approach based on deep-learning that aims to measure several asset control indicators from raw satellite images

The contribution of this project is two-fold. Firstly, to classify every single pixel of a satellite image into a predefined set of categories (asset, road access, tree, shrub layer, etc.), using semantic segmentation. Secondly, to deliver analytics that can be used to monitor, with few human resources, wind farm assets.

For this, Pervasive Technologies has implemented a solution based on satellite images and Deep Learning that allows to carry out this monitoring in a precise and effective way.

To achieve this objective, the following pipeline was defined:

  • Acquire periodic satellite images of the area of ​​interest. This can be a couple of times a year, monthly, weekly or even with greater recurrence. It really depends on the needs of the wind farm. However, there is a compromise between this temporal resolution and the cost of the images. The same way, it happens with spatial resolution, more pixels more expensive. For this project we captured images once a week with a spatial resolution of 0.8m/px.
  • Apply semantic segmentation using Deep Learning with predefined categories (assets, roads ...) pixel by pixel.
  • Calculate analytics to estimate the risk and changes in the ROI of these assets. For this we will apply GIS processing.

This methodology has been validated on two wind farms with very different characteristics: Barásoain and Carballeira. While Barásoain is an arid area, mainly composed of fields and very low vegetation, Carballeira is a very green area with a large shrub and herbaceous layer.

The system can generate alarms when an asset (wind turbine, sub-station, access route or evacuation line) is too close to the vegetation. In addition, we come to distinguish brush trees, which is of vital importance, since the safety distance reflected in the legislation is different for trees and bushes.

The predictive module for this project also provides additional information about the wind farm, such as the width of the access roads (which allows us to detect damage with timeseries analysis), or the characterization of the land (how much vegetation, buildings, fields, etc. has the area of ​​interest).

 

Outcomes and Applications

  • Wind farm asset surveillance
  • Pruning planning based on vegetation growth analysis and prediction
  • Pruning audit and certification
  • Wind farm wildfire risk estimation and scoring
  • New wind farms prospection based on accurate territory analysis (land tipologies, infrastructures, etc.)
  • Asset control & surveillance in hydroelectric power plants (pipes and swamps)
  • Immediate asset affectation analysis in natural disasters or other risk situations

Advantages:

  • Fully Automated system
  • Remote monitoring
  • Increase monitoring frequency
  • Reduce operational costs
  • Additional outputs

Results

The automation of Operations and Maintenance services for wind farms using satellite imagery and deep learning has disruptive potential for the energy industry. Moreover, the results suggest that operational costs can be drastically reduced by using the proposed highly automated system.

 

Listo para comenzar?

Contacte con nosotros para agendar una demostración o para conocer más sobre nuestros servicios y soluciones, o simplemente para saber cómo podemos ayudarle.

Contactar