J. Duhaney, T. Koshgoftaar, and A. Napolitano – 11th International Conference on Machine Learning and Applications (ICMLA), December, 2012
Class imbalance is prevalent in many real world datasets. It occurs when there are significantly fewer examples in one or more classes in a dataset compared to the number of instances in the remaining classes. When trained on highly imbalanced datasets, traditional machine learning techniques can often simply ignore the minority class(es) and label all instances as being of the majority class to maximize accuracy. This problem has been studied in many domains but there is little or no research related to the effect of class imbalance in fault data for condition monitoring of an ocean turbine. This study makes the first efforts in bridging that gap by providing insight into how class imbalance in vibration data can impact a learner’s ability to reliably identify changes in the ocean turbine’s operational state. To do so, we empirically evaluate the performances of three popular, but very different, machine learning algorithms when trained on four datasets with varying class distributions (one balanced and three imbalanced) to distinguish between a normal and an abnormal state. All data used in this study were collected from the testbed for an ocean turbine and were under sampled to simulate the different levels of imbalance. We find here, as in other domains, that the three learners seemed to suffer overall when trained on data with a highly skewed class distribution (with 0.1% examples in a faulty/abnormal state while the remaining 99.9% were captured in a normal operational state). It was noted, however, that the Logistic Regression and Decision Tree classifiers performed better when only 5% of the total number of examples were representative of an abnormal state (the remaining 95% therefore indicating normal operation) than they did when there was no imbalance present.
G Galanis, G Zodiatis, D Hayes, A Nikolaidis, G Georgiou, S Stylianou, G Kallos, C Kalogeri, PC Chu and A Charalambous – Advances in Meteorology, 2013
In a rapidly evolving operational and research framework concerning the global energy resources, new frontiers have been set for the scientific community working on environmental and renewable energy issues. In particular, new numerical techniques supporting the accurate estimation of renewable energy sources are highly emphasized. In this framework, wave energy – the energy that can be captured from sea waves – provides an alternative option with critical advantages. In the present paper, recent advances and some preliminary results obtained in two European projects will be discussed: Marina Platform and E-wave projects are focusing on the estimation of the wave energy potential in North Atlantic coastline of Europe and in Eastern Mediterranean Sea, respectively. Special emphasis is given to the utilization of numerical atmospheric and wave modeling systems able to accurately monitor the atmospheric and sea conditions in the area of interest. On the other hand, advanced statistical techniques are utilized for the local adaptation of the results and the estimation of the spatial and temporal distribution of the wave energy potential.
S Behrens, J Hayward, M Hemer, P Osman- Renewable Energy, July 2012
► The performance of three different wave energy converter (WEC) technologies was evaluated for Australian coastal regions. ► One of these WECs was found to operate with a capacity factor greater than 54.3%. ► The levelised cost of electricity (LCOE) for these WECs is in the order of $78/MWh.
J Hardisty – Energy and Power Engineering, May 2012
It is important to understand the relationship between the ambient ebb and flood currents and the electricity generated by tidal stream power generators to minimize investment risk and to optimize power generation for distribution purposes. Such analyses no longer rely on average descriptions of the flow field or on single values for the device efficiency. In the present Paper, we demonstrate a new method involving the integration of synthesized longterm flow vectors with logistic descriptions of the device power curves. New experiments are then described with the Neptune Proteus vertical axis tidal stream power generator involving to tests at speeds to 1.5 m/s in William Wright Dock on the Humber. The results are used to derive appropriate coefficients in the logistic curve and to estimate the device’s annual electrical output