Forecasting Air Pollutant Emissions in Thermal Power Plants: A Study of Prophet and ARIMA Models at Satpura Thermal Power Plant

The proposed study investigates the application of Prophet and ARIMA forecasting models for predicting air pollutant emissions at the Satpura Thermal Power Plant in Sarni, Madhya Pradesh, India. The plant’s substantial role in regional electricity generation necessitates accurate emission forecasts for effective environmental management. The study reviews the plant’s environmental impact and examines various forecasting methods, with a focus on Prophet and ARIMA. Data from 2019 to 2023 were meticulously preprocessed. The Prophet model achieved a Mean Absolute Error (MAE) of 7.8, providing smooth forecasts and capturing long-term trends. The ARIMA model, with an MAE of 8.67, was effective for short-term fluctuations but less smooth. The findings suggest Prophet’s superior trend forecasting capability, while ARIMA handles short-term variations. Future research should explore hybrid models and additional variables to enhance forecasting accuracy. AI-driven systems are recommended for better operational efficiency and environmental compliance.