Nils Twietmeyer

Predictive Analytics for Wind Farms – Project Experience with Intelligent IT Solutions for Optimization of Plant Efficiency and Availability

Abstract

As a wind generator’s performance is highly dependent on the current wind speed, inefficiencies are not always visible from SCADA data. Often component degradation or failure remain undetected and negatively affect electricity generation or plant availability. As maintenance is complex and costly due to the nacelle’s altitude and the distance between wind park and operation center, reliable information at an early stage about component condition and performance would be very beneficial. To identify inefficiencies, derogation or impending defects, SCADA measurements should be continuously analyzed and compared with a theoretical calculated value that takes into account the current environmental conditions.
For thermal power plants numerous different IT tools are available, which are successfully used since many years to detect such situations. Their potential for use in the wind power sector have been investigated, tried and tested. It was found that statistical calculation methods for process performance monitoring and early damage-detection fit best to the requirements of the wind sector. These methods and their application to wind data will be demonstrated within this paper.

Biographical Sketch

German diploma in process engineering and master of science in renewable energies.
Working as sales engineer since 2011 for different companies for renewable and conventional energies as well as environmental technologies.