Estimation and Short Term Prediction of Wave Elevation Using Artificial Neural Networks

Research toward the development of point absorber ocean wave energy converters (WEC) is migrating from energy absorption of ocean wave heave only, to the absorption of heave and surge. Due to the additional dimension of surge, the mathematic relationship between the ocean waves and the WEC has shifted from a linear model to a nonlinear model. This change has made back driven wave height from device motion extremely difficult. Even with linearization, the accuracy of the wave height estimation is not precise enough to predict future incoming waves, which is important for many advanced control approaches. In this paper, an artificial neural network is trained to report wave height from real-time dynamic data acquired from the WEC. Additionally, two short term forecasting methods are used to predict future wave elevations. Both predictions of current wave elevation and forecasting of future wave elevations show promise.