Macro Forum: The Street View

Macro Article

Actis in practice: Using artificial intelligence and machine learning

Hernan Arrigone

by Hernan Arrigone
Energy Infrastructure

Actis in practice: Using artificial intelligence and machine learning Image

Echoenergia Operations Control Centre (OCC)

In our current environment, investments in the renewable energy sector only work under high-performance assumptions. A wind asset must operate with an availability rate of above 97% to be considered efficient. A solar PV park is even higher, requiring availability rates of 98% and above. As a result, owners must look for every opportunity to improve operations. One great ally is the application of artificial intelligence and machine learning (AI and ML). These tools can read big sets of data acquired directly from the assets (wind turbines, substations, inverters, etc.) and based on trends and patterns, predict component failures well in advance of the events. This information, in turn, allows the asset manager to plan repairs or component exchanges without suffering from unexpected downtimes.

Echoenergia (“Echo”), an Actis Energy Fund 4 investee company, has been at the forefront of applying AI and ML tools for a number of years. Developing these tools in-house, Echo has created algorithms capable of identifying failures in wind turbines gearboxes weeks before they would materialise. As a result of this technology, Echo recognised and mitigated several events, which if they were to occur, would have cost the company close to half a million dollars in energy losses and emergency turbine repairs. Being able to partner with original equipment manufacturers (“OEMs”) and service providers - by supplying them data points to investigate – allowed Echo to avoid these losses.

Such technology can be scaled up and integrated into other components by using the same platform. This produces savings without significant investment and allows the company to expand the tools into other assets.

Echoenergia is one of the pioneers in Brazil’s renewable industry, and as such, will continue to develop and refine tools and algorithms that predict failures and improve the operation of its wind farms. Artificial intelligence and machine learning is one of the most promising applications of digital solutions for the renewable energy sector and Echo, and other Actis’ platforms, are in the vanguard of that trend. Given the small size of the investment – in the range of a few thousand dollars – and the big payback that we have seen, we know that the value added here is enormous and still untapped by most operators.