Data-driven Model Development and Validation for Laboratory Scale OC-OTEC Plant

Abstract

The growth of Renewable Energy (RE) resources is increasing rapidly to meet the energy and freshwater demand of the remote islands. Ocean Thermal Energy Conversion (OTEC) is identified as one of the promising renewable energy resources for tropical islands. Major challenges associated with OTEC are to improve the efficiency and reduce the production cost. Data-driven approach is one of the efficient techniques to develop models from Real time data to improve the efficiency of industrial processes and develop advanced control schemes. In this paper, linear data-driven models are developed for the Open Cycle (OC-OTEC) process by considering it as Single -Input Single Output (SISO) and Multi-Input Multi-Output (MIMO). Data is collected from OTEC Laboratory scale experimental setup established at National Institute of Ocean Technology (NIOT) and data driven models are developed using system identification techniques. To identify the proper inputs and outputs, sensitivity analysis is carried out based on the collected data. Then, models for OC-OTEC are developed to generate power and freshwater. The manipulated variables for model development are considered as sea surface warm water flow rate and deep sea cold water flow rate and disturbance as sea surface temperature variation due to climate changes. Finally, the developed models are validated by evaluating their performance metrics.