Remote sensing and machine learning-based monitoring of seawater temperature variations and their effects on desalination efficiency

Due to their high energy consumption, desalination plants have historically been configured to accommodate erratic weather patterns. In order to evaluate desalination for the needs of industrial and community water supplies, the Kalpakkam Atomic Power Station in Tamil Nadu, India, has been added. To evaluate and forecast the effects of SST variability on desalination efficiency and energy demand, two Random Forest (RF)-based regression models were created by combining satellite-derived SST (MODIS) data obtained through remote sensing with the operational performance parameters of the desalination plant. The findings clearly demonstrate the opposite relationship: energy consumption changed from 3.5 to 3.9 kWh/m³ in the opposite directions, while efficiency decreased from 96% in February at 27.1°C to 92% in June at 29.6°C. High prospecting ability was demonstrated by the RF model, which scored R-square values with 0.99 proximity and RMSE of only 0.11%. This model effectively represented complex non-linear correlations between SST and plant performance. This paves the way for the eventual attainment of water and energy security in climate-sensitive coastal locations by creating excellent opportunities for the integration of remote sensing and machine learning on desalination management applications. Additionally, it establishes the framework for resilient reactions that are scaled up in response to environmental variability.