A predictive model for oscillating water column wave energy converters based on machine learning

Abstract

Research on digital twin technology for efficient operation in various indus-trial and manufacturing sites is being actively conducted currently in South Korea. The gradual depletion of fossil fuels and environmental pollution issues require new renewable and eco-friendly power generation methods such as wave power plants. However, in wave power generation, which generates electricity by the energy of waves, it is criti-cal to understand and predict the amount of power generation and operational efficiency factors such as breakdown because these are closely related to highly variable wave ener-gy. Therefore, firstly, it is necessary to derive a meaningful correlation between highly volatile data such as wave height and sensor data in oscillating water column (OWC) chamber. Secondly, methodological study that can predict desired information should be conducted by learning the prediction situation with the extracted data based on the derived correlation. In this study, we design a workflow-based training model using a machine learning framework to predict the pressure of the OWC and verify the validity of pressure prediction analysis through the verification and evaluation dataset, using Internet of Things sensor data to enable smart operation and maintenance with the digital twin of the wave generation system. ？ 2021 ICIC International.