Predictive maintenance necessitates continuous and systematic monitoring of relevant parameters such as vibration levels to reduce OPEX (operational expenditures) and prevent premature termination of products’ normal lifespan.

Ensuring uninterrupted power generation is a major requirement in electricity power grids. As wind turbines operate in challenging conditions with highly fluctuating loads, continuous monitoring of key components such as the drive train is a necessary measure towards identifying changes which are indicative of an impending fault or failure, as a much more effective alternative to periodical assessment by operators.

Based on this, Condition Monitoring Systems (CMS) are implemented on a range of applicable equipment and products in real time, which not only provide a preventive foresight tool but also deliver an inclusive overview regarding the current and historical state of the product. CMS incorporates a range of devices, from vibration and thermal sensors and diagnosis, to torque measurements, through supervisory control and data acquisition (SCADA) and online condition monitoring systems to anticipate performance over time and potential degradation.

Possible undesired phenomena in different areas of turbine operation and their effect on other parameters are studied and subsequently various hypotheses are proposed and tested. The cause of an irregular occurrence  (such as fault, damage, etc.) can be classified into two main categories: hardware and software. On the hardware side, all components in the drive train and generator readings and the signal processing chain are tested. On the software side, signal conditioning in broadband features and narrowband features, signal sampling, signal feature extraction, and data storage means are implemented to store and analyze operational data across time and trigger notification mechanisms when necessary.