WindSmart: Predictive Maintenance for Offshore Wind Turbines
The energy transition from traditional oil and gas energy production to other sources of energy like wind and solar is critical in meeting climate initiative targets. However, renewable sources of energy like wind are notoriously more challenging to make profits due to lower margins. That means every dollar saved in operating costs is encouragement for other investments into the space and therefore, good for the environment.
For every turbine, the expected cost is about 45K USD per megawatt hour for the first 10 years after a turbine is built. There is a market expectation in the US for growth of up to 30 gigawatt hours of investment in turbines, up from only 42 megawatts today. That is significant growth and potential for optimizing operational costs across the board. Being able to predict and plan maintenance has an opportunity to save power production downtime, planning of logistics teams to service the maintenance, and plan for expected inventory of parts. Our target users are offshore wind engineers and technologists that need readily available predictions for which turbines need scheduled maintenance.
Our product, WindSmart provides anomaly predictions and flags events for offshore engineers and offers the major contributing variables to the total anomaly score. Through subject matter expert interviews, we found that being able to identify issues and what is driving the behavior allows for experts to identify the root cause. One challenge faced by predictive maintenance models today is the time horizon at which events are identified. It is critical to identify maintenance with at least 2 weeks, preferable a month in advance so that inventory can be checked, items ordered if not available, and scheduling and logistics of shipping and crew to the necessary site. WindSmart provides this service.