Posters

Abstract

The use of diagnostic systems is on the rise in the wind industry on its quest to maximize the efficiency of the operation of the turbines. Considering the wide range and availability of diagnostic systems in the market, several of these systems are often installed in a single turbine. Most diagnostic systems are designed to monitor a single component and do not communicate with other prognostic systems or the programmable logic controller of the turbine. This approach usually yields the desired result, i.e., notifying the operator/owner of the turbine when the turbine is not operating optimally, or a component is about to fail. It is, however, nowhere near utilizing the recorded data efficiently, i.e., providing less insights than could be extracted from the available data. This requires an improved assessment of their use and combination, to yield the optimal insights from the data collected. This work will focus on a diagnostic model framework that supports an in-depth analysis of the various diagnostic systems installed in a turbine, their performance, and their short comings. This allows us to organize them effectively and identify areas where their output overlaps and/or complement each other. Finally, it makes it possible to improve the entire system comprising of the various prognostic processes to optimize the utilization of the available data. This yields an optimal set of information for decision makers, so that they not only receive a comprehensive but also the understanding of its short comings and potential steps to improve it.