VIB2SE: Development of an Intelligent Tool for Vibration Analysis and Prediction of Large Synchronous Electric Machines

At a glance

• The generator is a limiting factor in increasing the power output of wind turbines. A key technical challenge here is structural dynamics caused by electromagnetically excited vibrations.
• The VIB2SE research project is developing an experimentally validated computational model to predict the vibration behavior of large electric machines induced by electromagnetic forces with low uncertainty as early as the design phase.
• The project partners are developing a physics-informed condition monitoring system for generators based on dynamic characteristics derived from field measurements to continuously monitor the generator’s condition and thereby increase the machine’s reliability.
• In addition to coordinating the project, Fraunhofer IWES is responsible, among other things, for simulating the drive train, conducting vibration tests on the test bench, and AI-based condition monitoring.

The challenge

The power output, reliability, and technical availability of a wind turbine are key factors influencing the levelized cost of electricity. However, while the increasing power output of generators generally leads to higher electricity production by the turbines, technical vulnerabilities also rise with size.

One of the most challenging aspects of increasing the power output of low- and medium-speed wind turbines is structural dynamics, particularly due to electromagnetically excited vibrations. Predicting this complex multiphysical problem is currently only possible to a limited extent during the generator’s design phase, meaning this aspect can hardly be taken into account in the design.

In a wind turbine, additional vibration-related phenomena occur during operation, such as interactions between the generator and other drive components or grid events that can compromise the system’s reliability. However, every generator failure during operation results in high costs due to repairs and lost revenue.

The solution

This is where the VIB2SE research project comes in. The scientists are developing an experimentally validated tool for predicting vibrations in large electrical machines as early as the design phase, as well as an intelligent, physics-informed condition monitoring system based on dynamic characteristics derived from field measurements.

To this end, a simulation chain is being developed that integrates structural and electromagnetic models to enable vibration predictions for generators in the drive train of wind turbines. Experimental tests and measurements on the Fraunhofer IWES 30-MW test bench are helping to develop a method for operational modal analysis tailored to high-power machines. The AI-supported prototype of a condition monitoring system significantly increases the reliability of the drive train.

The added value

VIB2SE aims to reduce vibrations as early as the design phase and to monitor them during operation. Thus, the results not only improve the reliability and longevity of wind turbines but also make a significant contribution to the development of high-power generators in the 15-MW class.