Tidal Energy Forecasting and Power Generation in Sri Lanka using ARIMA Fourier Models: A Case Study of Trincomalee.

Abstract

Sri Lanka is an island in the Indian Ocean and generates electricity from thermal power, hydropower, and other renewable energy sources such as solar and wind power. The country is expected to achieve 70% electricity demand through renewable energy by 2025. Although the country's geographical location offers the feasibility of tidal energy generation, the country has not expanded renewable energy enough. Tidal energy emerges as a promising new frontier in the power generation process with the initiation of lagoons instead of high-cost tidal barrages. In this study, Trincomalee Bay is selected as the study area, and a comparative analysis of time series models was conducted to forecast tidal ranges. A further novel power generation model was introduced to calculate the potential power by incorporating turbine efficiency, gravitational acceleration, and seawater density. Daily tidal range data for the period 2021 to 2023 from the National Aquatic Resources Research and Development Agency website was selected and validated through the cubic spline interpolation technique. Among the SARIMA (Seasonal Autoregressive Integrated Moving Average), Hybrid model (ARIMA with Fourier transform), LSTM (Long Short-Term Memory), and Prophet forecasting model, the LSTM model outperforms with a mean squared error (MSE) of 0.0040 and a 0.52 MW and 3.43 MW power prediction per day. Results indicate that tidal energy presents a promising avenue for Sri Lanka’s energy generation for the national grid and encourages policymakers and engineers to further investigate tidal power generation, incorporating environmental factors.