Turbine Design and Reliability

Smart MHK Generation System Modeling and Monitoring

Led by Yufei Tang, Ph.D.
Affiliated Home Campus: Boca Raton
Affiliated Department: Computer & Electrical Engineering and Computer Science


Marine Hydrokinetic Turbines and other similar ocean energy devices have shown an increasing potential for their ability to produce clean renewable energy. Industrial scale deployment of such devices would be ideal to meet the growing demand for renewable energy, however the inertly harsh and isolated operating conditions they are exposed to makes this a difficult endeavor. Innovative designs of MHK devices have been proposed, but operation and maintenance costs drive the efficiency of these energy production methods down. Monitoring the condition of such devices can provide information on developing faults and, ideally, once a fault begins to develop, it can be repaired before machine failure occurs driving operation and maintenance costs down and making these devices a viable renewable energy production method. Towards this end, we use a graph-based machine condition monitoring algorithm that represents time series data as adjacency matrices and applies four metrics to adjacent graphs based on graph entropy, distance between matrix elements, angle between graphs in inner product space, and matrix eigenvalues to measure the similarity of portions of adjacent sensor output. Based on these metrics, a machine learning approach is used to develop an improved metric which is used to detect changes in sensor output caused by a change in the condition of a machine.

Click here to watch the student presentation.

HBOI ocean fish