For advanced failure detection and maintenance of wind turbines.
Repeated failures of the Wind Turbines were causing downtime leading to significant costs as well as operational issues. The Client wanted to identify and predict failures based on patterns of usages and operational conditions like wind velocity, temperature etc.
Usage data was collected and streamed to a Big Data Platform. A Supervised model was built to identify the failure points and predict failures in future. Advanced quantitative approaches like Classification trees, Logistic regression, Neural Networks etc, were used to solve the problem.
The operational cost decreased by 15% due to the accurate prediction of turbine failure leading to a huge reduction in the downtime of Wind Turbines.