Power and freshwater prediction against seasonal variation in OC-OTEC plant at Lakshadweep using DNN

Ocean Thermal Energy Conversion(OTEC) is one of the most important renewable energy resources for islanders to meet the power and freshwater demand. It uses the temperature gradient between warm surface sea water and cold deep sea water. Challenges associated with the OC-OTEC are efficiency and economy. The efficiency of OTEC is mainly influenced by Sea Surface Temperature(SST) variations. The prediction of sea surface temperature is also a challenging task in a region with high SST variability. Hence an accurate SST prediction plays a major role in estimating the power and freshwater generation. Nowadays, Deep Neural Network(DNN) based prediction models are used for accurate SST prediction In this study, energy and freshwater production of the Open Cycle (OC) - OTEC plant at Lakshadweep is assessed based on the predicted SST variations of Lakshadweep and experimental data collected from the OTEC plant at NIOT, Chennai. First, the SST variation of Lakshadweep is predicted by developing prediction models using Recurrent Neural Network (RNN) based Long Short-Term Memory (LSTM) Deep Neural networks (DNN) using ten-year ECMWF satellite image monthly data. These predictive models are used to forecast OTEC power and freshwater. Finally, the net power and freshwater generations over a complete year have been evaluated for monthly as well as seasonal variations. The proposed methodology can be extensively used to optimize the usage of other renewable resources to satisfy the power and freshwater demand of islanders.