Avoid costly damage with predictive AI

20.08.2025

Source: E & M powernews

Artificial intelligence helps the renewables industry to detect faults in processes and systems at an early stage and prevent damage. The tool is even available free of charge.

What a promise: detecting damage to renewable energy systems before it occurs. The Fraunhofer Institute for Energy Economics and Energy System Technology IEE (Fraunhofer IEE) is now providing free software for the predictive maintenance of systems and processes.

The research institute and the Automated Energy Fault Detection and Identification (AEFDI) team have named their child "EnergyFaultDetector". It emerged from a series of research projects, according to a press release from the Kassel-based Fraunhofer IEE. The aim was to use algorithms to detect anomalies and create a flexible application that automatically detects faults at an early stage.

In wind farms, for example, damage and the resulting downtime can be costly. In another project, Fraunhofer IEE identified the technological innovations that wind farm operators most urgently need in order to prevent faults. The developed detector can detect conspicuous behavior in technical systems in real time, not only in wind turbines, but also in energy management processes or district heating systems.

The open source solution does not require any additional sensors or hardware and draws its conclusions from the existing operating data using an autoencoder approach. In addition to fault detection, the causes are analyzed, from which companies can derive maintenance measures. The detector can be "flexibly adapted to different systems", says Florian Rehwald, project manager at Fraunhofer IEE.

The institute provides the software as a Python library. Getting started is particularly quick because a module only needs one line of code to analyze operating data and pinpoint the source of the error. The Fraunhofer IEE also helps interested companies to integrate the software into their operational management systems. Free demo access to a wind monitoring platform is also available.

Author: Volker Stephan