Prediction of SO₂ Concentration Based on AR-LSTM Neural Network.

Area of Science:

• Environmental Science
• Data Science
• Chemical Engineering

Background:

• Sulphur dioxide (SO2) is a major air pollutant contributing to acid rain and respiratory illnesses.
• Real-time monitoring of industrial SO2 emissions is crucial for environmental and health protection.
• Accurate short-term forecasting of SO2 levels is needed for proactive control measures.

Purpose of the Study:

• To develop and evaluate an advanced analytical forecasting model for sulphur dioxide emissions.
• To improve the accuracy and timeliness of SO2 emission predictions.
• To provide a tool for better management and control of industrial air pollution.

Main Methods:

• Proposed an AR-LSTM analytical forecasting model integrating ARIMA and LSTM.
• Utilized sensor-based time series data for training and analysis.
• Conducted comparative experiments against other models using R², RMSE, and MAE metrics.

Main Results:

• The AR-LSTM model demonstrated significant effectiveness in SO2 emission forecasting.
• Comparative analysis confirmed the superiority of the proposed model over existing methods.
• The model successfully predicted SO2 emissions for the upcoming weeks.

Conclusions:

• The AR-LSTM model offers a robust solution for real-time SO2 emission monitoring and prediction.
• This approach can aid in mitigating the adverse environmental and health impacts of SO2 pollution.
• The model provides a valuable tool for industrial process control and air quality management.