Related Experiment Videos
Prediction of SO2 levels using neural networks
Belén M Fernández de Castro1, José Manuel Prada Sánchez, Wenceslao González Manteiga
1Department of Statistics and Operations Research, Faculty of Mathematics, University of Santiago de Compostela, Santiago, Spain. fdcastro@usc.es
Summary
This study adapted an air pollution control system for a power plant using neural networks and semiparametric models. The system provides 30-minute air quality predictions to prevent pollution episodes, aiding compliance with European directives.
Area of Science:
- Environmental Science
- Computer Science
Background:
- Power plants significantly impact local air quality.
- Compliance with European Council Directive 1999/30/CE is crucial for industrial operations.
- Existing air quality monitoring systems require timely predictive capabilities.
Purpose of the Study:
- To adapt an existing air pollution control help system for Endesa Generación S.A.'s power plant in As Pontes, Spain.
- To enhance the system's predictive accuracy for air quality episodes.
- To ensure alignment with European Council Directive 1999/30/CE standards.
Main Methods:
- Implementation of neural network models for air quality prediction.
- Development of a semiparametric model for comparative analysis.
- Integration of a 30-minute پیش بینی (prediction) lead time into the control system.
Main Results:
- The adapted system provides statistical air quality predictions 30 minutes in advance.
- Neural network predictions were compared against those from a semiparametric model.
- The system assists power plant staff in proactively managing air quality.
Conclusions:
- The adapted air pollution control system effectively aids in preventing air quality episodes.
- The use of neural network models offers a viable approach for short-term air quality forecasting.
- The system facilitates compliance with European air quality directives.