Wind/solar farm predictive maintenance
As part of its wind and solar farm operation, production and maintenance activities, our client wants to implement a set of health check-up tools for the different mechanical and electric systems for monitoring and conditional and predictive maintenance purposes.
Our consultants work on over a dozen use cases relating to the conditional and predictive maintenance of the wind and photovoltaic systems with the operations, diagnosis and maintenance teams. We used different approaches (statistics, mathematical modelling, machine learning) to design and develop the following analytical tools and algorithms:
• Anomaly detection: identification and characterisation of abnormal asset operating and production data patterns (for example, detection of ice on wind turbine blades, production system under-performance)
• Failure/breakage prediction models (for example, anticipation of wind turbine blade breakage risks, prediction of rotor and generator bearing block damage)
• System and sub-system digital twins (for example generator behaviour modelling)
• Key component remaining useful life estimation (for example, start/stop cycle studies to anticipate bearing wear)
Some algorithms were integrated into the PI infrastructure (Osisoft) that was already present at our client, and others were integrated as micro-services using the PI software REST interfaces.
use cases in production
of systems and components covered by conditional and predictive maintenance
of checking actions and maintenance operations