Characterization of the Concentrations of Volatile Organic Compounds in the Romanian Littoral using General Regression Neural Networks: A Case Study

ABSTRACT

In the last decade, preserving air quality in the northern Romanian Littoral has become a national priority. To monitor pollution, benzene, toluene, ortho-, meta-, and para-xylene, ethyl-benzene, and styrene were determined at six monitoring sites in three industrial and three urban areas. The sampling was done by an active method and the analysis by gas chromatography. The statistical analysis demonstrated that the concentrations of the volatile organic compounds were under the limits established by the European Union’s laws and that the pollutant distributions were skewed. Models were proposed for the concentrations of the volatile organic compounds, using the general regression neural network, with wind speed, air temperature, and humidity as regressors. The theoretical fit well with the experimental results, showing that the model is an effective alternative to conventional approaches.

KEYWORDS:

• Gas chromatography
• general regression neural network
• volatile organic compounds