Machine Learning in Renewable Energy Asset Management - Proxima Solutions
Predictive maintenance of wind turbines, using machine learning as the technological foundation.

Machine Learning in Renewable Energy Asset Management

A strategic combination for more efficient operations

Throughout the last three years, Proxima Solutions GmbH - and its team of data scientists, engineers, IT developers, and UI/UX designers - has enabled various players in the renewable energy industry to go beyond monitoring and analyzing their portfolio's activity in real-time. Using Windlog, the wind farm asset management software, the Asset managers, technical-commercial managers, and O&M technicians find interactive and user-friendly solutions to gain ownership of their energy projects.

Our passion and dedication for energy optimization and the energy transition are implemented through the combination of various technologies. Our wind farm asset management software is currently connected to Assets in Germany, France, Italy, Norway, Sweden, Spain, Greece, and Brazil. The focus, for now, is mainly on the wind energy sector, but pilots for solar assets are being already set-up. We owe a large part of our growth to partners who have trusted us since the beginning in a process that continues to benefit both sides. As a company that is constantly developing its portfolio, we look forward to consolidating strong alliances that allow us to share knowledge and thus renew and improve our products and services.

In 2018, aiming to create value in the renewable energy value chain, we entered a strategic alliance with CSEM, a Swiss research, and technology organization whose goal is to contribute to innovation processes by preparing industry and society for the future. CSEM started its AI research in renewable energies with photovoltaics as part of the PV Center R&D area. The partnership with the BKW Group and Proxima Solutions was crucial for CSEM’s breakthrough in other renewable energies.

“The collaboration with BKW and Proxima Solutions was our first foray into wind energy. We learned a lot about this renewable energy technology by working with the Proxima Solutions team. We also enjoyed moving ideas and algorithms from research to production in very short cycles. It’s very satisfying to see how our solutions contribute to new products and services from Proxima Solutions.” - Pierre-Jean Alet, Sector Head Digital Energy Solutions at CSEM.

The primary aim of this collaboration was to develop a software that enables wind turbine predictive maintenance, using machine learning as the technological foundation. One of the primary functions of this solution is to improve the process of predictive maintenance, with the aim to reduce operational losses that are associated with failures of the turbine's main components.

The journey of knowledge sharing, development, and innovation of this alliance started in 2018 with a pilot project that allowed us to realize the potential of predictive analytics in wind farm asset management. Given the excellent results, the project was resumed in September 2019 to professionalize what we developed during the pilot. We consider the ability to detect failures in wind turbines through data collection and analysis to be one of the key milestones of this project.

“For Proxima Solutions, the projects with CSEM have allowed us to complement our knowledge. The experience of our partner in machine learning and the expertise of our IT team facilitated the development of an algorithm for the early detection of anomalies in the operation of wind turbines. In this way, it is possible to increase the availability and energy production of wind farms.” - Giuseppe Madia, CEO of Proxima Solutions.

The integration of knowledge at a technical level on both sides was the main challenge, especially the integration with our server to customize the characteristics of the elements. “Effective and consistent communication between the parties is key. Thanks to the proximity with CSEM, our team could learn new data mining techniques and develop new algorithms for a correct and profitable analysis” - Alejandro Sampedro Senén, Data Scientist at Proxima Solutions.

The development of the project included three phases to improve the platform and introduce new functionalities. A first phase focused on reviewing the current state of the platform, identifying potential improvement opportunities to be designed and implemented. The second phase consisted of the creation of new modules, allowing the integration of the new functionalities to improve the platform. Finally, the third phase brought the main result of the project, the development of a complex dynamic algorithm to improve the alarm sensitivity and add new alarms.

“When the project started, alarms were only triggered by comparing readings from turbines within individual wind farms. There was no diagnosis or prioritization. Now we can diagnose problems with high accuracy by learning the normal behavior of individual turbines and coding expert knowledge. For example, the algorithms that evaluate vibration data have 93% accuracy over human experts.” - Pierre-Jean Alet, Sector Head Digital Energy Solutions at CSEM.

Partnership projects confirm the importance of creating solutions that combine human knowledge with technologies that facilitate the work of users in each sector. The energy sector today faces a major challenge in achieving carbon neutrality by 2050. To achieve the agreed targets, policies, and technologies, investments are needed above all, but knowledge sharing is also becoming an essential part of the process.

“The partnership between the BKW Group, Proxima Solutions and CSEM has allowed us to absorb and share more knowledge than we expected. This journey has gone through three main phases: learning, developing, and improving. The last phase has allowed us to consider new challenges and ensure further collaboration for the future.”- Dirk Oehlmann, Head of Sales at Proxima Solutions.

The mission of Proxima Solutions GmbH is to provide the market with tools that facilitate and enhance renewable energy asset management. Alliances will continue to be part of our business strategy to expand our portfolio and deal with industry challenges to ensure we are on the right path. We thank CSEM and the entire team for their support and interest in bringing this project to life.


The author would like to thank Pierre-Jean Alet and Alejandro Sampedro Senen for providing information to communicate the experiences and findings of this project.

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.